1 2013-11-14 Andrew MacLeod <amacleod@redhat.com>
3 * gimplify-me.h: New file. Add prototypes.
4 * gimplify.h: Don't include gimple.h.
5 (struct gimplify_hasher, struct gimplify_ctx, is_gimple_sizepos):
6 Relocate from gimple.h.
7 * gimple.h (struct gimplify_hasher, struct gimplify_ctx,
8 is_gimple_sizepos): Move to gimplify.h.
9 (gimplify_hasher::hash, gimplify_hasher::equal): Move to gimplify.c.
10 (enum gsi_iterator_update): Move to gimple-iterator.h.
11 * gimple-iterator.h (enum gsi_iterator_update): Relocate from gimple.h.
12 * gimplify-me.c: New File.
13 (force_gimple_operand_1, force_gimple_operand,
14 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
15 gimple_regimplify_operands): Relocate from gimplify.c.
16 * gimplify.c (force_gimple_operand_1, force_gimple_operand,
17 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
18 gimple_regimplify_operands): Move to gimplify-me.c.
19 (gimplify_hasher::hash, gimplify_hasher::equal): Relocate from gimple.h.
20 * Makefile.in (OBJS): Add gimplify-me.o
21 * asan.c: Include only gimplify.h, gimplify-me.h, and/or gimple.h as
23 * cfgloopmanip.c: Likewise.
24 * cgraphunit.c: Likewise.
25 * cilk-common.c: Likewise.
26 * fold-const.c: Likewise.
27 * function.c: Likewise.
28 * gimple-expr.c: Likewise.
29 * gimple-fold.c: Likewise.
30 * gimple-ssa-strength-reduction.c: Likewise.
32 * graphite-clast-to-gimple.c: Likewise.
33 * graphite-sese-to-poly.c: Likewise.
34 * ipa-prop.c: Likewise.
35 * ipa-split.c: Likewise.
37 * langhooks.c: Likewise.
38 * omp-low.c: Likewise.
40 * stor-layout.c: Likewise.
41 * targhooks.c: Likewise.
42 * trans-mem.c: Likewise.
43 * tree-affine.c: Likewise.
44 * tree-cfg.c: Likewise.
45 * tree-cfgcleanup.c: Likewise.
46 * tree-complex.c: Likewise.
47 * tree-if-conv.c: Likewise.
48 * tree-inline.c: Likewise.
49 * tree-loop-distribution.c: Likewise.
50 * tree-nested.c: Likewise.
51 * tree-parloops.c: Likewise.
52 * tree-predcom.c: Likewise.
53 * tree-profile.c: Likewise.
54 * tree-scalar-evolution.c: Likewise.
55 * tree-sra.c: Likewise.
56 * tree-ssa-address.c: Likewise.
57 * tree-ssa-ccp.c: Likewise.
58 * tree-ssa-dce.c: Likewise.
59 * tree-ssa-forwprop.c: Likewise.
60 * tree-ssa-ifcombine.c: Likewise.
61 * tree-ssa-loop-im.c: Likewise.
62 * tree-ssa-loop-ivopts.c: Likewise.
63 * tree-ssa-loop-manip.c: Likewise.
64 * tree-ssa-loop-niter.c: Likewise.
65 * tree-ssa-loop-prefetch.c: Likewise.
66 * tree-ssa-loop-unswitch.c: Likewise.
67 * tree-ssa-math-opts.c: Likewise.
68 * tree-ssa-phiopt.c: Likewise.
69 * tree-ssa-phiprop.c: Likewise.
70 * tree-ssa-pre.c: Likewise.
71 * tree-ssa-propagate.c: Likewise.
72 * tree-ssa-reassoc.c: Likewise.
73 * tree-ssa-sccvn.c: Likewise.
74 * tree-ssa-strlen.c: Likewise.
75 * tree-ssa.c: Likewise.
76 * tree-switch-conversion.c: Likewise.
77 * tree-tailcall.c: Likewise.
78 * tree-vect-data-refs.c: Likewise.
79 * tree-vect-generic.c: Likewise.
80 * tree-vect-loop-manip.c: Likewise.
81 * tree-vect-loop.c: Likewise.
82 * tree-vect-patterns.c: Likewise.
83 * tree-vect-stmts.c: Likewise.
86 * value-prof.c: Likewise.
87 * config/aarch64/aarch64.c: Likewise.
88 * config/alpha/alpha.c: Likewise.
89 * config/darwin.c: Likewise.
90 * config/i386/i386.c: Likewise.
91 * config/ia64/ia64.c: Likewise.
92 * config/mep/mep.c: Likewise.
93 * config/mips/mips.c: Likewise.
94 * config/rs6000/rs6000.c: Likewise.
95 * config/s390/s390.c: Likewise.
96 * config/sh/sh.c: Likewise.
97 * config/sparc/sparc.c: Likewise.
98 * config/spu/spu.c: Likewise.
99 * config/stormy16/stormy16.c: Likewise.
100 * config/tilegx/tilegx.c: Likewise.
101 * config/tilepro/tilepro.c: Likewise.
102 * config/xtensa/xtensa.c: Likewise.
104 2013-11-14 Joern Rennecke <joern.rennecke@embecosm.com>
106 * config/arc/arc.md (doloop_begin_i): Remove extra alignment;
109 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
111 * config/rs6000/sysv4le.h (LINUX64_DEFAULT_ABI_ELFv2): Define.
113 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
114 Alan Modra <amodra@gmail.com>
116 * config/rs6000/rs6000.h (RS6000_SAVE_AREA): Handle ABI_ELFv2.
117 (RS6000_SAVE_TOC): Remove.
118 (RS6000_TOC_SAVE_SLOT): New macro.
119 * config/rs6000/rs6000.c (rs6000_parm_offset): New function.
120 (rs6000_parm_start): Use it.
121 (rs6000_function_arg_advance_1): Likewise.
122 (rs6000_emit_prologue): Use RS6000_TOC_SAVE_SLOT.
123 (rs6000_emit_epilogue): Likewise.
124 (rs6000_call_aix): Likewise.
125 (rs6000_output_function_prologue): Do not save/restore r11
126 around calling _mcount for ABI_ELFv2.
128 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
129 Alan Modra <amodra@gmail.com>
131 * config/rs6000/rs6000-protos.h (rs6000_reg_parm_stack_space):
133 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Remove.
134 (REG_PARM_STACK_SPACE): Call rs6000_reg_parm_stack_space.
135 * config/rs6000/rs6000.c (rs6000_parm_needs_stack): New function.
136 (rs6000_function_parms_need_stack): Likewise.
137 (rs6000_reg_parm_stack_space): Likewise.
138 (rs6000_function_arg): Do not replace BLKmode by Pmode when
139 returning a register argument.
141 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
142 Michael Gschwind <mkg@us.ibm.com>
144 * config/rs6000/rs6000.h (FP_ARG_MAX_RETURN): New macro.
145 (ALTIVEC_ARG_MAX_RETURN): Likewise.
146 (FUNCTION_VALUE_REGNO_P): Use them.
147 * config/rs6000/rs6000.c (TARGET_RETURN_IN_MSB): Define.
148 (rs6000_return_in_msb): New function.
149 (rs6000_return_in_memory): Handle ELFv2 homogeneous aggregates.
150 Handle aggregates of up to 16 bytes for ELFv2.
151 (rs6000_function_value): Handle ELFv2 homogeneous aggregates.
153 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
154 Michael Gschwind <mkg@us.ibm.com>
156 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
157 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
158 (rs6000_discover_homogeneous_aggregate): Likewise.
159 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
160 (rs6000_function_arg_advance_1): Likewise.
161 (rs6000_function_arg): Likewise.
162 (rs6000_arg_partial_bytes): Likewise.
163 (rs6000_psave_function_arg): Handle BLKmode arguments.
165 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
166 Michael Gschwind <mkg@us.ibm.com>
168 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
169 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
170 (rs6000_discover_homogeneous_aggregate): Likewise.
171 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
172 (rs6000_function_arg_advance_1): Likewise.
173 (rs6000_function_arg): Likewise.
174 (rs6000_arg_partial_bytes): Likewise.
175 (rs6000_psave_function_arg): Handle BLKmode arguments.
177 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
179 * config/rs6000/rs6000.c (machine_function): New member
181 (rs6000_emit_prologue): Set r2_setup_needed if necessary.
182 (rs6000_output_mi_thunk): Set r2_setup_needed.
183 (rs6000_output_function_prologue): Output global entry point
184 prologue and local entry point marker if needed for ABI_ELFv2.
185 Output -mprofile-kernel code here.
186 (output_function_profiler): Do not output -mprofile-kernel
187 code here; moved to rs6000_output_function_prologue.
188 (rs6000_file_start): Output ".abiversion 2" for ABI_ELFv2.
190 (rs6000_emit_move): Do not handle dot symbols for ABI_ELFv2.
191 (rs6000_output_function_entry): Likewise.
192 (rs6000_assemble_integer): Likewise.
193 (rs6000_elf_encode_section_info): Likewise.
194 (rs6000_elf_declare_function_name): Do not create dot symbols
195 or .opd section for ABI_ELFv2.
197 (rs6000_trampoline_size): Update for ABI_ELFv2 trampolines.
198 (rs6000_trampoline_init): Likewise.
199 (rs6000_elf_file_end): Call file_end_indicate_exec_stack
202 (rs6000_call_aix): Handle ELFv2 indirect calls. Do not check
203 for function descriptors in ABI_ELFv2.
205 * config/rs6000/rs6000.md ("*call_indirect_aix<mode>"): Support
206 on ABI_AIX only, not ABI_ELFv2.
207 ("*call_value_indirect_aix<mode>"): Likewise.
208 ("*call_indirect_elfv2<mode>"): New pattern.
209 ("*call_value_indirect_elfv2<mode>"): Likewise.
211 * config/rs6000/predicates.md ("symbol_ref_operand"): Do not
212 check for function descriptors in ABI_ELFv2.
213 ("current_file_function_operand"): Likewise.
215 * config/rs6000/ppc-asm.h [__powerpc64__ && _CALL_ELF == 2]:
217 (FUNC_NAME): Define ELFv2 variant.
218 (JUMP_TARGET): Likewise.
219 (FUNC_START): Likewise.
220 (HIDDEN_FUNC): Likewise.
221 (FUNC_END): Likeiwse.
223 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
225 * config.gcc [powerpc*-*-* | rs6000-*-*]: Support --with-abi=elfv1
226 and --with-abi=elfv2.
227 * config/rs6000/option-defaults.h (OPTION_DEFAULT_SPECS): Add "abi".
228 * config/rs6000/rs6000.opt (mabi=elfv1): New option.
229 (mabi=elfv2): Likewise.
230 * config/rs6000/rs6000-opts.h (enum rs6000_abi): Add ABI_ELFv2.
231 * config/rs6000/linux64.h (DEFAULT_ABI): Do not hard-code to AIX_ABI
233 (ELFv2_ABI_CHECK): New macro.
234 (SUBSUBTARGET_OVERRIDE_OPTIONS): Use it to decide whether to set
235 rs6000_current_abi to ABI_AIX or ABI_ELFv2.
236 (GLIBC_DYNAMIC_LINKER64): Support ELFv2 ld.so version.
237 * config/rs6000/rs6000-c.c (rs6000_cpu_cpp_builtins): Predefine
238 _CALL_ELF and __STRUCT_PARM_ALIGN__ if appropriate.
240 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Handle ABI_ELFv2.
241 (debug_stack_info): Likewise.
242 (rs6000_file_start): Treat ABI_ELFv2 the same as ABI_AIX.
243 (rs6000_legitimize_tls_address): Likewise.
244 (rs6000_conditional_register_usage): Likewise.
245 (rs6000_emit_move): Likewise.
246 (init_cumulative_args): Likewise.
247 (rs6000_function_arg_advance_1): Likewise.
248 (rs6000_function_arg): Likewise.
249 (rs6000_arg_partial_bytes): Likewise.
250 (rs6000_output_function_entry): Likewise.
251 (rs6000_assemble_integer): Likewise.
252 (rs6000_savres_strategy): Likewise.
253 (rs6000_stack_info): Likewise.
254 (rs6000_function_ok_for_sibcall): Likewise.
255 (rs6000_emit_load_toc_table): Likewise.
256 (rs6000_savres_routine_name): Likewise.
257 (ptr_regno_for_savres): Likewise.
258 (rs6000_emit_prologue): Likewise.
259 (rs6000_emit_epilogue): Likewise.
260 (rs6000_output_function_epilogue): Likewise.
261 (output_profile_hook): Likewise.
262 (output_function_profiler): Likewise.
263 (rs6000_trampoline_size): Likewise.
264 (rs6000_trampoline_init): Likewise.
265 (rs6000_elf_output_toc_section_asm_op): Likewise.
266 (rs6000_elf_encode_section_info): Likewise.
267 (rs6000_elf_reloc_rw_mask): Likewise.
268 (rs6000_elf_declare_function_name): Likewise.
269 (rs6000_function_arg_boundary): Treat ABI_ELFv2 the same as ABI_AIX,
270 except that rs6000_compat_align_parm is always assumed false.
271 (rs6000_gimplify_va_arg): Likewise.
272 (rs6000_call_aix): Update comment.
273 (rs6000_sibcall_aix): Likewise.
274 * config/rs6000/rs6000.md ("tls_gd_aix<TLSmode:tls_abi_suffix>"):
275 Treat ABI_ELFv2 the same as ABI_AIX.
276 ("*tls_gd_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
277 ("tls_ld_aix<TLSmode:tls_abi_suffix>"): Likewise.
278 ("*tls_ld_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
279 ("load_toc_aix_si"): Likewise.
280 ("load_toc_aix_di"): Likewise.
282 ("call_value"): Likewise.
283 ("*call_local_aix<mode>"): Likewise.
284 ("*call_value_local_aix<mode>"): Likewise.
285 ("*call_nonlocal_aix<mode>"): Likewise.
286 ("*call_value_nonlocal_aix<mode>"): Likewise.
287 ("*call_indirect_aix<mode>"): Likewise.
288 ("*call_value_indirect_aix<mode>"): Likewise.
289 ("sibcall"): Likewise.
290 ("sibcall_value"): Likewise.
291 ("*sibcall_aix<mode>"): Likewise.
292 ("*sibcall_value_aix<mode>"): Likewise.
293 * config/rs6000/predicates.md ("symbol_ref_operand"): Likewise.
294 ("current_file_function_operand"): Likewise.
296 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
298 * config/rs6000/rs6000.c (rs6000_arg_partial_bytes): Simplify logic
299 by making use of the fact that for vector / floating point arguments
300 passed both in VRs/FPRs and in the fixed parameter area, the partial
301 bytes mechanism is in fact not used.
303 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
305 * config/rs6000/rs6000.c (rs6000_psave_function_arg): New function.
306 (rs6000_finish_function_arg): Likewise.
307 (rs6000_function_arg): Use rs6000_psave_function_arg and
308 rs6000_finish_function_arg to handle both vector and floating
309 point arguments that are also passed in GPRs / the stack.
311 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
313 * config/rs6000/rs6000.c (USE_FP_FOR_ARG_P): Remove TYPE argument.
314 (USE_ALTIVEC_FOR_ARG_P): Likewise.
315 (rs6000_darwin64_record_arg_advance_recurse): Update uses.
316 (rs6000_function_arg_advance_1):Likewise.
317 (rs6000_darwin64_record_arg_recurse): Likewise.
318 (rs6000_function_arg): Likewise.
319 (rs6000_arg_partial_bytes): Likewise.
321 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
323 * config/rs6000/rs6000.c (rs6000_option_override_internal): Replace
324 "DEFAULT_ABI != ABI_AIX" test by testing for ABI_V4 or ABI_DARWIN.
325 (rs6000_savres_strategy): Likewise.
326 (rs6000_return_addr): Likewise.
327 (rs6000_emit_load_toc_table): Replace "DEFAULT_ABI != ABI_AIX" by
328 testing for ABI_V4 (since ABI_DARWIN is impossible here).
329 (rs6000_emit_prologue): Likewise.
330 (legitimate_lo_sum_address_p): Simplify DEFAULT_ABI test.
331 (rs6000_elf_declare_function_name): Remove duplicated test.
332 * config/rs6000/rs6000.md ("load_toc_v4_PIC_1"): Explicitly test
333 for ABI_V4 (instead of "DEFAULT_ABI != ABI_AIX" test).
334 ("load_toc_v4_PIC_1_normal"): Likewise.
335 ("load_toc_v4_PIC_1_476"): Likewise.
336 ("load_toc_v4_PIC_1b"): Likewise.
337 ("load_toc_v4_PIC_1b_normal"): Likewise.
338 ("load_toc_v4_PIC_1b_476"): Likewise.
339 ("load_toc_v4_PIC_2"): Likewise.
340 ("load_toc_v4_PIC_3b"): Likewise.
341 ("load_toc_v4_PIC_3c"): Likewise.
342 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Simplify DEFAULT_ABI test.
343 (RS6000_SAVE_AREA): Likewise.
344 (FP_ARG_MAX_REG): Likewise.
345 (RETURN_ADDRESS_OFFSET): Likewise.
346 * config/rs6000/sysv.h (TARGET_TOC): Test for ABI_V4 instead
348 (SUBTARGET_OVERRIDE_OPTIONS): Likewise.
349 (MINIMAL_TOC_SECTION_ASM_OP): Likewise.
351 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
353 * config/rs6000/rs6000.c (rs6000_call_indirect_aix): Rename to ...
354 (rs6000_call_aix): ... this. Handle both direct and indirect calls.
355 Create call insn directly instead of via various gen_... routines.
356 Mention special registers used by the call in CALL_INSN_FUNCTION_USAGE.
357 (rs6000_sibcall_aix): New function.
358 * config/rs6000/rs6000.md (TOC_SAVE_OFFSET_32BIT): Remove.
359 (TOC_SAVE_OFFSET_64BIT): Likewise.
360 (AIX_FUNC_DESC_TOC_32BIT): Likewise.
361 (AIX_FUNC_DESC_TOC_64BIT): Likewise.
362 (AIX_FUNC_DESC_SC_32BIT): Likewise.
363 (AIX_FUNC_DESC_SC_64BIT): Likewise.
364 ("call" expander): Call rs6000_call_aix.
365 ("call_value" expander): Likewise.
366 ("call_indirect_aix<ptrsize>"): Replace this pattern ...
367 ("call_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
368 ("*call_indirect_aix<mode>"): ... by this insn pattern.
369 ("call_value_indirect_aix<ptrsize>"): Replace this pattern ...
370 ("call_value_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
371 ("*call_value_indirect_aix<mode>"): ... by this insn pattern.
372 ("*call_nonlocal_aix32", "*call_nonlocal_aix64"): Replace by ...
373 ("*call_nonlocal_aix<mode>"): ... this pattern.
374 ("*call_value_nonlocal_aix32", "*call_value_nonlocal_aix64"): Replace
375 ("*call_value_nonlocal_aix<mode>"): ... by this pattern.
376 ("*call_local_aix<mode>"): New insn pattern.
377 ("*call_value_local_aix<mode>"): Likewise.
378 ("sibcall" expander): Call rs6000_sibcall_aix.
379 ("sibcall_value" expander): Likewise. Move earlier in file.
380 ("*sibcall_nonlocal_aix<mode>"): Replace by ...
381 ("*sibcall_aix<mode>"): ... this pattern.
382 ("*sibcall_value_nonlocal_aix<mode>"): Replace by ...
383 ("*sibcall_value_aix<mode>"): ... this pattern.
384 * config/rs6000/rs6000-protos.h (rs6000_call_indirect_aix): Remove.
385 (rs6000_call_aix): Add prototype.
386 (rs6000_sibcall_aix): Likewise.
388 2013-11-14 Jakub Jelinek <jakub@redhat.com>
391 * asan.c (asan_emit_stack_protection): Ensure -fsection-anchors
392 isn't confused by the artificial decl.
394 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
396 * config/rs6000/rs6000.c (rs6000_emit_prologue): Do not place a
397 RTX_FRAME_RELATED_P marker on the UNSPEC_MOVESI_FROM_CR insn.
398 Instead, add USEs of all modified call-saved CR fields to the
399 insn storing the result to the stack slot, and provide an
400 appropriate REG_FRAME_RELATED_EXPR for that insn.
401 * config/rs6000/rs6000.md ("*crsave"): New insn pattern.
402 * config/rs6000/predicates.md ("crsave_operation"): New predicate.
404 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
405 Alan Modra <amodra@gmail.com>
407 * function.c (assign_parms): Use all.reg_parm_stack_space instead
408 of re-evaluating REG_PARM_STACK_SPACE target macro.
409 (locate_and_pad_parm): New parameter REG_PARM_STACK_SPACE. Use it
410 instead of evaluating target macro REG_PARM_STACK_SPACE every time.
411 (assign_parm_find_entry_rtl): Update call.
412 * calls.c (initialize_argument_information): Update call.
413 (emit_library_call_value_1): Likewise.
414 * expr.h (locate_and_pad_parm): Update prototype.
416 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
418 * calls.c (store_unaligned_arguments_into_pseudos): Skip PARALLEL
421 2013-11-14 DJ Delorie <dj@redhat.com>
423 * config/rx/rx.c (rx_mode_dependent_address_p): Allow offsets up
426 2013-11-14 Jeff Law <law@redhat.com>
428 * tree-ssa-threadedge.c (thread_through_normal_block): Only push
429 the EDGE_START_JUMP_THREAD marker if the jump threading path is
432 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
434 * doc/invoke.texi: Update documentation for AArch64's -mcpu
437 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
439 * config/aarch64/aarch64-cores.def (example-1): Remove.
440 (example-2): Likewise.
441 * config/aarch64/aarch64-tune.md: Regenerate.
442 * config/aarch64/aarch64.md: Do not include "large.md" or "small.md".
443 (generic_sched): Remove "large", "small".
444 * config/aarch64/large.md: Delete.
445 * config/aarch64/small.md: Delete.
447 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
449 * config/aarch64/aarch64-cores.def (cortex-a57): Tune for cortexa15.
450 * config/aarch64/aarch64-tune.md: Regenerate.
451 * config/aarch64/aarch64.md: Include cortex-a15 pipeline model.
452 (generic_sched): "no" if we are tuning for cortexa15.
453 * config/arm/cortex-a15.md: Include cortex-a15-neon.md by
456 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
458 * config/aarch64/aarch64-arches.def (armv8-a): Tune for cortex-a53.
459 * config/aarch64/aarch64.md: Do not include aarch64-generic.md.
460 * config/aarch64/aarch64.c (aarch64_tune): Initialize to cortexa53.
461 (all_cores): Use cortexa53 when tuning for "generic".
462 (aarch64_override_options): Fix comment.
463 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Set to cortexa53.
464 * config/aarch64/aarch64-generic.md: Delete.
466 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
468 * config/aarch64/aarch64.c (all_architectures): Remove "generic".
470 2013-11-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
472 * config/aarch64/aarch64.c: Include aarch-cost-tables.h.
473 (generic_rtx_cost_table): Remove.
474 (aarch64_rtx_costs): Use fields from cpu_cost_table.
475 * config/aarch64/aarch64-protos.h (tune_params): Use cpu_cost_table for
477 (cpu_rtx_cost_table): Remove.
479 2013-11-14 Julian Brown <julian@codesourcery.com>
480 Joey Ye <joey.ye@arm.com>
482 * config/arm/arm.c (arm_cortex_m_branch_cost): New.
484 (arm_slowmul_tune, arm_fastmul_tune,
485 arm_strongarm_tune, arm_9e_tune, arm_v6t2_tune,
486 arm_cortex_tune, arm_cortex_a15_tune,
487 arm_cortex_a5_tune, arm_v6m_tune): Add comments
489 * config/arm/arm-cores.def (cortex-m4, cortex-m3):
492 2013-11-14 Kirill Yukhin <kirill.yukhin@intel.com>
495 * config/ia64/ia64.c (ia64_split_tmode_move): Relax `dead'
498 2013-11-14 Jakub Jelinek <jakub@redhat.com>
499 Uros Bizjak <ubizjak@gmail.com>
502 * config/i386/i386.md (*anddi_2): Only allow CCZmode if
503 operands[2] satisfies_constraint_Z that might have bit 31 set.
505 2013-11-13 Jeff Law <law@redhat.com>
507 PR tree-optimization/59102
508 * gimple-ssa-isolate-paths.c
509 (insert_trap_and_remove_trailing_statments): Ensure STMT is a
510 gimple assignment before looking at gimple_assign_lhs.
512 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
514 * ira.c: Add comment about threads at the top of file.
516 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
518 * ira-color.c (coalesce_allocnos): Don't allocate and free
521 2013-11-14 Tom de Vries <tom@codesourcery.com>
523 * tree-ssa-tail-merge.c (gimple_equal_p): Add test for structural
524 equality for GIMPLE_ASSIGN.
526 2013-11-14 Tom de Vries <tom@codesourcery.com>
528 * tree-ssa-tail-merge.c (gimple_operand_equal_value_p): Factor new
530 (gimple_equal_p): ... here.
532 2013-11-14 Tom de Vries <tom@codesourcery.com>
534 * trans-mem.c (is_tm_ending): New function.
535 * gimple.h (is_tm_ending): Declare.
536 * tree-ssa-tail-merge.c (gimple_equal_p): Remove test on
538 (find_duplicate): Use is_tm_ending instead of is_tm_ending_fndecl.
540 2013-11-14 Tom de Vries <tom@codesourcery.com>
542 * tree-ssa-tail-merge.c (gimple_equal_p): Remove equal variable.
544 2013-11-13 Andrew MacLeod <amacleod@redhat,com>
546 * gimple-walk.h: New File. Relocate prototypes from gimple.h.
547 (struct walk_stmt_info): Relocate here from gimple.h.
548 * gimple-iterator.h: New File. Relocate prototypes from gimple.h.
549 (struct gimple_stmt_iterator_d): Relocate here from gimple.h.
550 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
551 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
552 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
553 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
554 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Relocate here from gimple.h.
555 * gimple.h (struct gimple_stmt_iterator_d): Move to gimple-iterator.h.
556 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
557 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
558 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
559 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
560 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Move to gimple-iterator.h.
561 (struct walk_stmt_info): Move to gimple-walk.h.
562 (gimple_seq_set_location): Move to gimple.c
563 * gimple-walk.c: New File.
564 (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm, walk_gimple_op,
565 walk_gimple_stmt, get_base_loadstore, walk_stmt_load_store_addr_ops,
566 walk_stmt_load_store_ops): Relocate here from gimple.c.
567 * gimple-iterator.c: Include gimple-iterator.h.
568 * gimple.c (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm,
569 walk_gimple_op, walk_gimple_stmt, get_base_loadstore,
570 walk_stmt_load_store_addr_ops, walk_stmt_load_store_ops): Move to
572 (gimple_seq_set_location): Relocate from gimple.h.
573 * tree-phinodes.h (set_phi_nodes): Move to tree-phinodes.c.
574 * tree-phinodes.c (set_phi_nodes): Relocate from tree-phinodes.h.
575 * gengtype.c (open_base_files): Add gimple-iterator.h to include list.
576 * Makefile.in (OBJS): Add gimple-walk.o
577 * asan.c: Update Include list as required for gimple-iterator.h and
579 * cfgexpand.c: Likewise.
580 * cfgloop.c: Likewise.
581 * cfgloopmanip.c: Likewise.
582 * cgraph.c: Likewise.
583 * cgraphbuild.c: Likewise.
584 * cgraphunit.c: Likewise.
585 * gimple-fold.c: Likewise.
586 * gimple-low.c: Likewise.
587 * gimple-pretty-print.c: Likewise.
588 * gimple-ssa-isolate-paths.c: Likewise.
589 * gimple-ssa-strength-reduction.c: Likewise.
590 * gimple-streamer-in.c: Likewise.
591 * gimple-streamer-out.c: Likewise.
592 * gimplify.c: Likewise.
593 * graphite-blocking.c: Likewise.
594 * graphite-clast-to-gimple.c: Likewise.
595 * graphite-dependences.c: Likewise.
596 * graphite-interchange.c: Likewise.
597 * graphite-optimize-isl.c: Likewise.
598 * graphite-poly.c: Likewise.
599 * graphite-scop-detection.c: Likewise.
600 * graphite-sese-to-poly.c: Likewise.
601 * graphite.c: Likewise.
602 * ipa-inline-analysis.c: Likewise.
603 * ipa-profile.c: Likewise.
604 * ipa-prop.c: Likewise.
605 * ipa-pure-const.c: Likewise.
606 * ipa-split.c: Likewise.
607 * lto-streamer-in.c: Likewise.
608 * lto-streamer-out.c: Likewise.
609 * omp-low.c: Likewise.
610 * predict.c: Likewise.
611 * profile.c: Likewise.
613 * tracer.c: Likewise.
614 * trans-mem.c: Likewise.
615 * tree-call-cdce.c: Likewise.
616 * tree-cfg.c: Likewise.
617 * tree-cfgcleanup.c: Likewise.
618 * tree-complex.c: Likewise.
619 * tree-data-ref.c: Likewise.
620 * tree-dfa.c: Likewise.
621 * tree-eh.c: Likewise.
622 * tree-emutls.c: Likewise.
623 * tree-if-conv.c: Likewise.
624 * tree-inline.c: Likewise.
625 * tree-into-ssa.c: Likewise.
626 * tree-loop-distribution.c: Likewise.
627 * tree-nested.c: Likewise.
628 * tree-nrv.c: Likewise.
629 * tree-object-size.c: Likewise.
630 * tree-outof-ssa.c: Likewise.
631 * tree-parloops.c: Likewise.
632 * tree-predcom.c: Likewise.
633 * tree-profile.c: Likewise.
634 * tree-scalar-evolution.c: Likewise.
635 * tree-sra.c: Likewise.
636 * tree-ssa-ccp.c: Likewise.
637 * tree-ssa-coalesce.c: Likewise.
638 * tree-ssa-copy.c: Likewise.
639 * tree-ssa-copyrename.c: Likewise.
640 * tree-ssa-dce.c: Likewise.
641 * tree-ssa-dom.c: Likewise.
642 * tree-ssa-dse.c: Likewise.
643 * tree-ssa-forwprop.c: Likewise.
644 * tree-ssa-ifcombine.c: Likewise.
645 * tree-ssa-live.c: Likewise.
646 * tree-ssa-loop-ch.c: Likewise.
647 * tree-ssa-loop-im.c: Likewise.
648 * tree-ssa-loop-ivcanon.c: Likewise.
649 * tree-ssa-loop-ivopts.c: Likewise.
650 * tree-ssa-loop-manip.c: Likewise.
651 * tree-ssa-loop-niter.c: Likewise.
652 * tree-ssa-loop-prefetch.c: Likewise.
653 * tree-ssa-loop.c: Likewise.
654 * tree-ssa-math-opts.c: Likewise.
655 * tree-ssa-phiopt.c: Likewise.
656 * tree-ssa-phiprop.c: Likewise.
657 * tree-ssa-pre.c: Likewise.
658 * tree-ssa-propagate.c: Likewise.
659 * tree-ssa-reassoc.c: Likewise.
660 * tree-ssa-sink.c: Likewise.
661 * tree-ssa-strlen.c: Likewise.
662 * tree-ssa-structalias.c: Likewise.
663 * tree-ssa-tail-merge.c: Likewise.
664 * tree-ssa-ter.c: Likewise.
665 * tree-ssa-threadedge.c: Likewise.
666 * tree-ssa-threadupdate.c: Likewise.
667 * tree-ssa-uncprop.c: Likewise.
668 * tree-ssa-uninit.c: Likewise.
669 * tree-ssa.c: Likewise.
670 * tree-stdarg.c: Likewise.
671 * tree-switch-conversion.c: Likewise.
672 * tree-tailcall.c: Likewise.
673 * tree-vect-data-refs.c: Likewise.
674 * tree-vect-generic.c: Likewise.
675 * tree-vect-loop-manip.c: Likewise.
676 * tree-vect-loop.c: Likewise.
677 * tree-vect-patterns.c: Likewise.
678 * tree-vect-slp.c: Likewise.
679 * tree-vect-stmts.c: Likewise.
680 * tree-vectorizer.c: Likewise.
681 * tree-vrp.c: Likewise.
684 * value-prof.c: Likewise.
685 * vtable-verify.c: Likewise.
687 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
689 * gimple-ssa-isolate-paths.c (pass_isolate_erroneous_paths): Comment
692 2013-11-13 Jeff Law <law@redhat.com>
694 * PR middle-end/59119
695 * gimple-ssa-isolate-paths.c (find_implicit_erroneous_behaviour): New
696 function, extracted from gimple_ssa_isolate_erroneous_paths.
697 (find_explicit_erroneous_behaviour): Similarly.
698 (insert_trap_and_remove_trailing_statements): Remove statements
701 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
703 * cfgrtl.c (can_fallthru): Reorder code to move tablejump check up.
704 Make that check explicit. BB_HEAD cannot be NULL, remove check for it.
705 * haifa-sched.c (ready_remove_first_dispatch): Check INSN_P before
706 looking at INSN_CODE.
707 * reload1.c (delete_dead_insn) Do not expect JUMP_TABLE_DATA to be an
708 active_insn_p object, respect basic block boundaries.
709 * reorg.c (follow_jumps): Use invariant that JUMP_TABLE_DATA always
710 follows immediately after the jump table data label.
711 * config/nds32/nds32.c (nds32_output_casesi_pc_relative): Likewise.
712 * config/sh/sh.c (barrier_align): Likewise. Rearrange code such
713 that JUMP_TABLE_DATA is not expected to be an active_insn_p object.
715 2013-11-13 Teresa Johnson <tejohnson@google.com>
718 * predict.c (drop_profile): Error is currently too strict.
719 (handle_missing_profiles): Pass call_count to drop_profile.
721 2013-11-13 Teresa Johnson <tejohnson@google.com>
724 * ipa-inline.c (edge_badness): Fix overflow.
726 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
728 PR rtl-optimization/59036
729 * ira-color.c (struct allocno_color_data): Add new members
730 first_thread_allocno, next_thread_allocno, thread_freq.
731 (sorted_copies): New static var.
732 (allocnos_conflict_by_live_ranges_p, copy_freq_compare_func): Move
734 (allocno_thread_conflict_p, merge_threads)
735 (form_threads_from_copies, form_threads_from_bucket)
736 (form_threads_from_colorable_allocno, init_allocno_threads): New
738 (bucket_allocno_compare_func): Add comparison by thread frequency
740 (add_allocno_to_ordered_bucket): Rename to
741 add_allocno_to_ordered_colorable_bucket. Remove parameter.
742 (push_only_colorable): Call form_threads_from_bucket.
743 (color_pass): Call init_allocno_threads. Use
744 consideration_allocno_bitmap instead of coloring_allocno_bitmap
745 for nuillify allocno color data.
746 (ira_initiate_assign, ira_finish_assign): Allocate/free
748 (coalesce_allocnos): Use static sorted copies.
750 2013-11-13 Jakub Jelinek <jakub@redhat.com>
752 * passes.c (execute_todo): Don't call do_per_function if
754 (execute_one_ipa_transform_pass, execute_one_pass): Don't call
755 execute_function_dump if dump_file is NULL.
757 2013-11-13 Martin Jambor <mjambor@suse.cz>
759 * cgraph.c (cgraph_get_create_node): Do what
760 cgraph_get_create_real_symbol_node used to do.
761 (cgraph_get_create_real_symbol_node): Removed. Changed all users to
762 call cgraph_get_create_node.
763 * cgraph.h (cgraph_get_create_real_symbol_node): Removed.
764 * lto-streamer-in.c (input_function): Call cgraph_get_node instead of
765 cgraph_get_create_node. Assert we get a node.
767 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
769 * config/aarch64/aarch64-simd.md (vec_extract): New.
771 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
773 * config/aarch64/aarch64-simd.md (vec_set<mode>): Add w -> w option to
776 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
778 * cfgexpand.c (expand_used_vars): Allocate space for partitions based
779 on PARM_DECLs or RESULT_DECLs only if they are ignored for debug info
780 or if optimization is enabled.
781 * tree-ssa-coalesce.c (coalesce_ssa_name): If optimization is disabled,
782 require that all the names based on a PARM_DECL or a RESULT_DECL that
783 isn't ignored for debug info be coalesced.
785 2013-11-13 Jan-Benedict Glaw <jbglaw@lug-owl.de>
787 * config/c6x/c6x.c: Include "gimple-expr.h".
789 2013-11-13 Richard Biener <rguenther@suse.de>
791 * gimple-streamer-out.c (output_gimple_stmt): Also wrap
792 decls in ADDR_EXPR operands inside a MEM_REF and optimize that.
793 * gimple-streamer-in.c (input_gimple_stmt): Remove now dead
794 code dealing with type mismatches inside component reference
797 2013-11-13 Marc Glisse <marc.glisse@inria.fr>
799 PR tree-optimization/59077
800 * ipa-pure-const.c (better_state): Update *state.
802 2013-11-13 Christophe Lyon <christophe.lyon@linaro.org>
804 * config/aarch64/aarch64.h (FRAME_GROWS_DOWNWARD): Define to 1.
805 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
806 Update offset calculations.
808 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
811 * dwarf2out.c (add_byte_size_attribute): Also use int_size_in_bytes
812 for fields. Do not add the attribute if the size is negative.
814 2013-11-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
816 * config/arm/arm.c: Include aarch-cost-tables.h.
817 (generic_extra_costs): Move from here...
818 * config/arm/aarch-cost-tables.h: ... To here. New file.
820 2013-11-13 Alexander Ivchenko <alexander.ivchenko@intel.com>
821 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
822 Sergey Lega <sergey.s.lega@intel.com>
823 Anna Tikhonova <anna.tikhonova@intel.com>
824 Ilya Tocar <ilya.tocar@intel.com>
825 Andrey Turetskiy <andrey.turetskiy@intel.com>
826 Ilya Verbin <ilya.verbin@intel.com>
827 Kirill Yukhin <kirill.yukhin@intel.com>
828 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
830 * config/i386/i386.c (ix86_print_operand): Support z-masking
831 * config/i386/predicate.md (const_0_to_4_operand): New.
832 (const_0_to_5_operand): Ditto.
833 * config/i386/sse.md (UNSPEC_COMPRESS): New.
834 (UNSPEC_COMPRESS_STORE): Ditto.
835 (UNSPEC_EXPAND): Ditto.
836 (UNSPEC_EMBEDDED_ROUNDING): Ditto.
837 (define_mode_attr ssescalarsize): Ditto.
838 (avx512f_load<mode>_mask): Ditto.
839 (avx512f_store<mode>_mask): Ditto.
840 (avx512f_storedqu<mode>_mask): Ditto.
841 (avx512f_vmcmp<mode>3_mask): Ditto.
842 (avx512f_fmadd_<mode>_mask): Ditto.
843 (avx512f_fmadd_<mode>_mask3): Ditto.
844 (avx512f_fmsub_<mode>_mask): Ditto.
845 (avx512f_fmsub_<mode>_mask3): Ditto.
846 (avx512f_fnmadd_<mode>_mask): Ditto.
847 (avx512f_fnmadd_<mode>_mask3): Ditto.
848 (avx512f_fnmsub_<mode>_mask): Ditto.
849 (avx512f_fnmsub_<mode>_mask3): Ditto.
850 (avx512f_fmaddsub_<mode>_mask): Ditto.
851 (avx512f_fmaddsub_<mode>_mask3): Ditto.
852 (avx512f_fmsubadd_<mode>_mask): Ditto.
853 (avx512f_fmsubadd_<mode>_mask3): Ditto.
854 (vec_unpacku_float_lo_v16si): Ditto.
855 (avx512f_vextract<shuffletype>32x4_mask): Ditto.
856 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
857 (avx512f_vextract<shuffletype>64x4_mask): Ditto.
858 (vec_extract_lo_<mode>_maskm): Ditto.
859 (vec_extract_hi_<mode>_maskm): Ditto.
860 (avx512f_vternlog<mode>_mask): Ditto.
861 (avx512f_shufps512_mask): Ditto.
862 (avx512f_fixupimm<mode>_mask): Ditto.
863 (avx512f_shufpd512_mask): Ditto.
864 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
865 (avx512f_<code>v8div16qi2_mask/trunc): Ditto.
866 (*avx512f_<code>v8div16qi2_store_mask): Ditto.
867 (ashr<mode>3<mask_name>): Ditto.
868 (avx512f_vinsert<shuffletype>32x4_mask): Ditto.
869 (avx512f_vinsert<shuffletype>64x4_mask): Ditto.
870 (avx512f_shuf_<shuffletype>64x2_mask): Ditto.
871 (avx512f_shuf_<shuffletype>32x4_mask): Ditto.
872 (avx512f_pshufdv3_mask): Ditto.
873 (avx512f_perm<mode>_mask): Ditto.
874 (avx512f_vpermi2var<mode>3_mask): Ditto.
875 (avx512f_vpermt2var<mode>3_mask): Ditto.
876 (avx512f_compress<mode>_mask): Ditto.
877 (avx512f_compressstore<mode>_mask): Ditto.
878 (avx512f_expand<mode>_mask): Ditto.
879 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>): Extend
881 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
882 (<plusminus_insn><mode>3<mask_name>): Ditto.
883 (*<plusminus_insn><mode>3<mask_name>): Ditto.
884 (mul<mode>3<mask_name>): Ditto.
885 (*mul<mode>3<mask_name>): Ditto.
886 (<sse>_div<mode>3<mask_name>): Ditto.
887 (<mask_codefor>rcp14<mode><mask_name>): Ditto.
888 (<sse>_sqrt<mode>2<mask_name>): Ditto.
889 (<mask_codefor>rsqrt14<mode><mask_name>): Ditto.
890 (<code><mode>3<mask_name>/smaxmin): Ditto.
891 (*<code><mode>3_finite<mask_name>/smaxmin): Ditto.
892 (*<code><mode>3<mask_name>/smaxmin): Ditto.
893 (float<sseintvecmodelower><mode>2<mask_name>): Ditto.
894 (ufloatv16siv16sf2<mask_name>): Ditto.
895 (<mask_codefor>avx512f_fix_notruncv16sfv16si<mask_name>): Ditto.
896 (<mask_codefor>avx512f_ufix_notruncv16sfv16si<mask_name>): Ditto.
897 (<fixsuffix>fix_truncv16sfv16si2<mask_name>): Ditto.
898 (float<si2dfmodelower><mode>2<mask_name>): Ditto.
899 (ufloatv8siv8df<mask_name>): Ditto.
900 (<mask_codefor>avx512f_cvtpd2dq512<mask_name>): Ditto.
901 (avx512f_ufix_notruncv8dfv8si<mask_name>): Ditto.
902 (<fixsuffix>fix_truncv8dfv8si2<mask_name>): Ditto.
903 (<mask_codefor>avx512f_cvtpd2ps512<mask_name>): Ditto.
904 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix><mask_name>): Ditto.
905 (<mask_codefor>avx512f_unpckhps512<mask_name>): Ditto.
906 (<mask_codefor>avx512f_unpcklps512<mask_name>): Ditto.
907 (<mask_codefor>avx512f_movshdup512<mask_name>): Ditto.
908 (<mask_codefor>avx512f_movsldup512<mask_name>): Ditto.
909 (<mask_codefor>avx512f_vextract<shuffletype>32x4_1<mask_name>): Ditto.
910 (vec_extract_lo_<mode><mask_name>): Ditto.
911 (vec_extract_hi_<mode><mask_name>): Ditto.
912 (<mask_codefor>avx512f_unpckhpd512<mask_name>): Ditto.
913 (avx512f_movddup512<mask_name>): Ditto.
914 (avx512f_unpcklpd512<mask_name>): Ditto.
915 (*avx512f_unpcklpd512<mask_name>): Ditto.
916 (*avx512f_vmscalef<mode>): Ditto.
917 (avx512f_scalef<mode><mask_name>): Ditto.
918 (avx512f_getexp<mode><mask_name>): Ditto.
919 (<mask_codefor>avx512f_align<mode><mask_name>): Ditto.
920 (avx512f_rndscale<mode><mask_name>): Ditto.
921 (avx512f_shufps512_1<mask_name>): Ditto.
922 (avx512f_shufpd512_1<mask_name>): Ditto.
923 (<plusminus_insn><mode>3<mask_name>): Ditto.
924 (*<plusminus_insn><mode>3<mask_name>): Ditto.
925 (vec_widen_umult_even_v16si<mask_name>): Ditto.
926 (*vec_widen_umult_even_v16si<mask_name>): Ditto.
927 (vec_widen_smult_even_v16si<mask_name>): Ditto.
928 (*vec_widen_smult_even_v16si<mask_name>): Ditto.
929 (mul<mode>3<mask_name>): Ditto.
930 (*<sse4_1_avx2>_mul<mode>3<mask_name>): Ditto.
931 (<shift_insn><mode>3<mask_name>): Ditto.
932 (avx512f_<rotate>v<mode><mask_name>/rotate): Ditto.
933 (avx512f_<rotate><mode><mask_name>): Ditto.
934 (<code><mode>3<mask_name>/maxmin): Ditto.
935 (*avx2_<code><mode>3<mask_name>/maxmin): Ditto.
936 (<sse2_avx2>_andnot<mode>3<mask_name>): Ditto.
937 (*andnot<mode>3<mask_name>): Ditto.
938 (<mask_codefor><code><mode>3<mask_name>/any_logic): Ditto.
939 (<mask_codefor>avx512f_interleave_highv16si<mask_name>): Ditto.
940 (<mask_codefor>avx512f_interleave_lowv16si<mask_name>): Ditto.
941 (<mask_codefor>avx512f_vinsert<shuffletype>32x4_1<mask_name>): Ditto.
942 (vec_set_lo_<mode><mask_name>): Ditto.
943 (vec_set_hi_<mode><mask_name>): Ditto.
944 (avx512f_shuf_<shuffletype>64x2_1<mask_name>): Ditto.
945 (avx512f_shuf_<shuffletype>32x4_1<mask_name>): Ditto.
946 (avx512f_pshufd_1<mask_name>): Ditto.
947 (<mask_codefor>abs<mode>2<mask_name>): Ditto.
948 (<mask_codefor>avx512f_<code>v16qiv16si2<mask_name>): Ditto.
949 (avx512f_<code>v16hiv16si2<mask_name>/any_extend): Ditto.
950 (avx512f_<code>v8qiv8di2<mask_name>/any_extend): Ditto.
951 (avx512f_<code>v8hiv8di2<mask_name>/any_extend): Ditto.
952 (avx512f_<code>v8siv8di2<mask_name>/any_extend): Ditto.
953 (avx512er_exp2<mode><mask_name>): Ditto.
954 (<mask_codefor>avx512er_rcp28<mode><mask_name>): Ditto.
955 (<mask_codefor>avx512er_rsqrt28<mode><mask_name>): Ditto.
956 (<avx2_avx512f>_permvar<mode><mask_name>): Ditto.
957 (<avx2_avx512f>_perm<mode>_1<mask_name>): Ditto.
958 (<mask_codefor>avx512f_vec_dup<mode><mask_name>): Ditto.
959 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V16FI): Ditto.
960 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V8FI): Ditto.
961 (<mask_codefor>avx512f_vec_dup_gpr<mode><mask_name>): Ditto.
962 (<mask_codefor>avx512f_vec_dup_mem<mode><mask_name>): Ditto.
963 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF2): Ditto.
964 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF1): Ditto.
965 (*<sse2_avx_avx512f>_vpermilp<mode><mask_name>): Ditto.
966 (<sse2_avx_avx512f>_vpermilvar<mode>3<mask_name>): Ditto.
967 (<avx2_avx512f>_ashrv<mode><mask_name>): Ditto.
968 (<avx2_avx512f>_<shift_insn>v<mode><mask_name>): Ditto.
969 (<mask_codefor>avx512f_vcvtph2ps512<mask_name>): Ditto.
970 (<mask_codefor>avx512f_vcvtps2ph512<mask_name>): Ditto.
971 (avx512f_getmant<mode><mask_name>): Ditto.
972 (clz<mode>2<mask_name>): Ditto.
973 (<mask_codefor>conflict<mode><mask_name>): Ditto.
974 (*srcp14<mode>): Remove visibility.
975 (*rsqrt14<mode>): Ditto.
976 (*fma_fmsub_<mode>): Ditto.
977 (*fma_fnmadd_<mode>): Ditto.
978 (*avx512f_rndscale<mode>): Ditto.
979 * config/i386/subst.md: New file.
981 2013-11-13 Joseph Myers <joseph@codesourcery.com>
983 * doc/extend.texi (Statement Exprs, Typeof): Discuss __auto_type.
984 * ginclude/stdatomic.h (kill_dependency, atomic_store_explicit)
985 (atomic_load_explicit, atomic_exchange_explicit)
986 (atomic_compare_exchange_strong_explicit)
987 (atomic_compare_exchange_weak_explicit): Use __auto_type to
988 declare variable initialized with PTR argument.
990 2013-11-12 Jeff Law <law@redhat.com>
992 * tree-ssa-threadedge.c (thread_around_empty_blocks): New
993 argument backedge_seen_p. Set, use and pass it to children
995 (thread_through_normal_block): Similarly.
996 (thread_across_edge): Similarly.
998 * gimple-ssa-isolate-paths.c (check_loadstore): Mark discovered
999 memory references as volatile.
1000 (insert_trap_and_remove_trailing_statements): Fix comment.
1002 2013-11-12 Vladimir Makarov <vmakarov@redhat.com>
1005 * ira-costs.c (record_operand_costs): Check operands number for
1008 2013-11-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1011 * config/rs6000/rs6000.md (movdi_internal32): Eliminate
1012 constraints that would allow DImode into the traditional Altivec
1013 registers, but cause undesirable code generation when loading 0 as
1015 (movdi_internal64): Likewise.
1016 (cmp<mode>_fpr): Do not use %x for CR register output.
1017 (extendsfdf2_fpr): Fix constraints when -mallow-upper-df and
1018 -mallow-upper-sf debug switches are used.
1020 2013-11-12 Andrew MacLeod <amacleod@redhat.com>
1022 * gimple-expr.h (create_tmp_var_name, create_tmp_var_raw,
1023 create_tmp_var, create_tmp_reg, mark_addressable, is_gimple_reg_rhs):
1024 Relocate prototypes from gimple.h.
1025 * gimplify.h: New File. Relocate some prototypes from gimple.h here.
1026 (gimple_predicate, enum fallback, enum gimplify_status): Relocate
1028 * gimple.h: Move some prototypes to gimplify.h.
1029 (gimple_predicate, enum fallback, enum gimplify_status): Move to
1031 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location):
1032 Relocate from gimpify.c.
1033 * gimple-expr.c (remove_suffix, tmp_var_id_num, create_tmp_var_name,
1034 create_tmp_var_raw, create_tmp_var, create_tmp_reg, mark_addressable,
1035 is_gimple_reg_rhs) Relocate from gimplify.c.
1036 * gimplify.c (mark_addressable): Move to gimple-expr.c.
1037 (gimple_seq_add_stmt_without_update): Move to gimple.c.
1038 (remove_suffix, tmp_var_id_num, create_tmp_var_name, create_tmp_var_raw,
1039 create_tmp_var, create_tmp_reg, is_gimple_reg_rhs): Move to
1041 (should_carry_location_p): Move to gimple.c.
1042 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location): Move
1044 (annotate_one_with_location, annotate_all_with_location_after,
1045 annotate_all_with_location): Move to gimple.c.
1046 (compare_case_labels, sort_case_labels,
1047 preprocess_case_label_vec_for_gimple): Move to gimple.c.
1048 (rhs_predicate_for): Make static.
1049 (gimplify_assign): Relocate from gimple.c.
1050 * gimple.c (gimplify_assign): Move to gimplify.c.
1051 (gimple_seq_add_stmt_without_update, should_carry_location_p,
1052 annotate_one_with_location, annotate_all_with_location_after,
1053 annotate_all_with_location, compare_case_labels, sort_case_labels,
1054 preprocess_case_label_vec_for_gimple): Relocate from gimplify.c.
1055 * tree.h (unshare_expr, unshare_expr_without_location,
1056 mark_addressable): Move prototypes to gimplify.h.
1057 * Makefile.in (GTFILES): gimple-expr.c now has the GTY tag for
1059 * asan.c: Include gimplify.h rather than gimple.h.
1060 * cfgloopmanip.c: Likewise.
1061 * cgraphunit.c: Likewise.
1062 * cilk-common.c: Likewise.
1063 * dwarf2out.c: Dont include gimple.h.
1064 * fold-const.c: Include gimplify.h rather than gimple.h.
1065 * function.c: Likewise.
1066 * gimple-fold.c: Likewise.
1067 * gimple-ssa-strength-reduction.c: Likewise.
1068 * graphite-clast-to-gimple.c: Likewise.
1069 * graphite-sese-to-poly.c: Likewise.
1070 * ipa-prop.c: Likewise.
1071 * ipa-split.c: Likewise.
1073 * langhooks.c: Dont include gimple.h.
1074 * loop-init.c: Include gimplify.h rather than gimple.h.
1075 * omp-low.c: Likewise.
1077 * stor-layout.c: Likewise.
1078 * targhooks.c: Likewise.
1079 * trans-mem.c: Likewise.
1080 * tree-affine.c: Likewise.
1081 * tree-cfg.c: Likewise.
1082 * tree-cfgcleanup.c: Likewise.
1083 * tree-complex.c: Likewise.
1084 * tree-if-conv.c: Likewise.
1085 * tree-inline.c: Likewise.
1086 * tree-iterator.c: Likewise.
1087 * tree-loop-distribution.c: Likewise.
1088 * tree-nested.c: Likewise.
1089 * tree-parloops.c: Likewise.
1090 * tree-predcom.c: Likewise.
1091 * tree-profile.c: Likewise.
1092 * tree-scalar-evolution.c: Likewise.
1093 * tree-sra.c: Likewise.
1094 * tree-ssa-address.c: Likewise.
1095 * tree-ssa-ccp.c: Likewise.
1096 * tree-ssa-dce.c: Likewise.
1097 * tree-ssa-forwprop.c: Likewise.
1098 * tree-ssa-ifcombine.c: Likewise.
1099 * tree-ssa-loop-im.c: Likewise.
1100 * tree-ssa-loop-ivopts.c: Likewise.
1101 * tree-ssa-loop-manip.c: Likewise.
1102 * tree-ssa-loop-niter.c: Likewise.
1103 * tree-ssa-loop-prefetch.c: Likewise.
1104 * tree-ssa-loop-unswitch.c: Likewise.
1105 * tree-ssa-math-opts.c: Likewise.
1106 * tree-ssa-phiopt.c: Likewise.
1107 * tree-ssa-phiprop.c: Likewise.
1108 * tree-ssa-pre.c: Likewise.
1109 * tree-ssa-propagate.c: Likewise.
1110 * tree-ssa-reassoc.c: Likewise.
1111 * tree-ssa-sccvn.c: Likewise.
1112 * tree-ssa-strlen.c: Likewise.
1113 * tree-ssa.c: Likewise.
1114 * tree-switch-conversio: Likewise.n.c
1115 * tree-tailcall.c: Likewise.
1116 * tree-vect-data-refs.c: Likewise.
1117 * tree-vect-generic.c: Likewise.
1118 * tree-vect-loop-manip.c: Likewise.
1119 * tree-vect-loop.c: Likewise.
1120 * tree-vect-patterns.c: Likewise.
1121 * tree-vect-stmts.c: Likewise.
1123 * value-prof.c: Likewise.
1124 * config/aarch64/aarch64.c: Include gimplify.h instead of gimple.h.
1125 * config/alpha/alpha.c: Likewise.
1126 * config/darwin.c: Likewise.
1127 * config/i386/i386.c: Likewise.
1128 * config/ia64/ia64.c: Likewise.
1129 * config/mep/mep.c: Likewise.
1130 * config/mips/mips.c: Likewise.
1131 * config/rs6000/rs6000.c: Likewise.
1132 * config/s390/s390.c: Likewise.
1133 * config/sh/sh.c: Likewise.
1134 * config/sparc/sparc.c: Likewise.
1135 * config/spu/spu.c: Likewise.
1136 * config/stormy16/stormy16.c: Likewise.
1137 * config/tilegx/tilegx.c: Likewise.
1138 * config/tilepro/tilepro.c: Likewise.
1139 * config/xtensa/xtensa.c: Likewise.
1141 2013-11-12 Adam Butcher <adam@jessamine.co.uk>
1143 * tree.c (grow_tree_vec_stat): New function ...
1144 * tree.h (grow_tree_vec_stat) (grow_tree_vec): ... and its declaration
1145 and macro front-end.
1147 2013-11-12 Marek Polacek <polacek@redhat.com>
1149 * final.c (update_alignments): Initialize label to NULL_RTX.
1151 2013-11-12 Jeff Law <law@redhat.com>
1153 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
1154 (insert_trap_and_remove_trailing_statements): New argument OP which
1155 is the NULL pointer. Emit the trap after the load/store through
1156 the NULL pointer. Simplify the RHS of a store through a NULL pointer
1157 when trivial to do so.
1158 (isolate_path): Corresponding changes.
1159 (gimple_ssa_isolate_erroneous_path): Likewise.
1161 2013-11-12 Teresa Johnson <tejohnson@google.com>
1162 Jan Hubicka <jh@suse.cz>
1164 * predict.c (drop_profile): New function.
1165 (handle_missing_profiles): Ditto.
1166 (counts_to_freqs): Don't overwrite estimated frequencies
1167 when function has no profile counts.
1168 * predict.h (handle_missing_profiles): Declare.
1169 * tree-inline.c (freqs_to_counts): New function.
1170 (copy_cfg_body): Invoke freqs_to_counts as needed.
1171 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
1173 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
1176 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
1178 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
1180 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
1183 * config/i386/i386.c (ix86_option_override_internal): Check
1184 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
1185 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
1186 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
1187 MASK_AVX256_SPLIT_UNALIGNED_STORE.
1189 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
1190 Clear m_COREI7_AVX and update comments.
1191 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
1193 2013-11-12 Martin Jambor <mjambor@suse.cz>
1195 PR rtl-optimization/10474
1196 * ira.c (interesting_dest_for_shprep): New function.
1197 (split_live_ranges_for_shrink_wrap): Likewise.
1198 (find_moveable_pseudos): Move calculation of dominance info,
1199 df_analysios and the final anlyses to...
1200 (ira): ...here, call split_live_ranges_for_shrink_wrap.
1202 2013-11-12 Bin Cheng <bin.cheng@arm.com>
1204 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
1205 Handle type conversion.
1207 2013-11-11 Martin Liska <marxin.liska@gmail.com>
1208 Jan Hubicka <jh@suse.cz>
1210 * cgraph.c (dump_cgraph_node): Profile dump added.
1211 * cgraph.h (struct cgraph_node): New time profile variable added.
1212 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
1213 * gcov-io.h (gcov_type): New profiler type introduced.
1214 * ipa-profile.c (lto_output_node): Streaming for time profile added.
1215 (input_node): Time profiler is read from LTO stream.
1216 * predict.c (maybe_hot_count_p): Hot prediction changed.
1217 * profile.c (instrument_values): New case for time profiler added.
1218 (compute_value_histograms): Read of time profile.
1219 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
1220 * tree-profile.c (init_ic_make_global_vars): Time profiler function added.
1221 (gimple_init_edge_profiler): TP function instrumentation.
1222 (gimple_gen_time_profiler): New.
1223 * value-prof.c (gimple_add_histogram_value): Support for time profiler
1225 (dump_histogram_value): TP type added to dumps.
1226 (visit_hist): More sensitive check that takes TP into account.
1227 (gimple_find_values_to_profile): TP instrumentation.
1228 * value-prof.h (hist_type): New histogram type added.
1229 (struct histogram_value_t): Pointer to struct function added.
1230 * libgcc/Makefile.in: New GCOV merge function for TP added.
1231 * libgcov.c: function_counter variable introduced.
1232 (_gcov_merge_time_profile): New.
1233 (_gcov_time_profiler): New.
1235 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
1236 Jeff Law <law@redhat.com>
1238 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
1239 ao_ref_init_from_ptr_and_size for builtins.
1241 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
1242 H.J. Lu <hongjiu.lu@intel.com>
1245 * config/i386/x86-tune.def
1246 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
1247 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
1248 * config/i386/i386.h
1249 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
1250 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
1251 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
1252 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
1253 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
1254 misaligned_prologue_used. Check that
1255 desired_aling <= epilogue_size_needed.
1257 2013-11-11 Cong Hou <congh@google.com>
1259 PR tree-optimization/59050
1260 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
1262 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
1265 * expmed.c (emit_store_flag): Fail for const-const comparison.
1267 2013-11-11 Tristan Gingold <gingold@adacore.com>
1268 Eric Botcazou <ebotcazou@adacore.com>
1270 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
1271 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
1272 * tree.def (CONSTRUCTOR): Likewise.
1273 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
1274 * gimplify.c (gimplify_init_constructor): Do not clear the object when
1275 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
1277 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
1279 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
1282 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1284 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
1287 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
1289 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
1291 2013-11-08 Jeff Law <law@redhat.com>
1293 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1294 threading paths first, then perform PHI node checks if applicable.
1296 2013-11-10 Karlson2k <k2k@narod.ru>
1297 Kai Tietz <ktietz@redhat.com>
1300 * configure.ac: Adding for exported symbols check
1301 and for rdynamic-check executable-extension.
1302 * configure: Regenerated.
1304 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
1306 * mode-switching.c (optimize_mode_switching): Mark block as
1307 nontransparent, if last_mode at block exit is different from no_mode.
1309 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1311 * function.c (NAME__MAIN): Move to...
1312 * cfgexpand.c (NAME__MAIN): ...here.
1314 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
1316 * target.def (can_use_doloop_p): New hook.
1317 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
1318 * doc/tm.texi: Regenerate.
1319 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
1320 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
1321 * hooks.c (hook_bool_dint_dint_uint_true): New function.
1322 * targhooks.h (can_use_doloop_if_innermost): Declare.
1323 * targhooks.c (can_use_doloop_if_innermost): New function.
1324 * target.h: Include double-int.h.
1325 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
1326 Remove iteration count, maximum iteration count, loop depth and
1327 enter-at-top inputs from doloop_begin and doloop_end.
1328 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
1330 * config/arc/arc.c (arc_can_use_doloop_p): New function.
1331 (TARGET_CAN_USE_DOLOOP_P): Define.
1332 * config/arm/thumb2.md (doloop_end): Update for new interface.
1333 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
1334 * config/bfin/bfin.md (doloop_end): Update for new interface.
1335 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
1336 (TARGET_CAN_USE_DOLOOP_P): Define.
1337 * config/c6x/c6x.md (doloop_end): Update for new interface.
1338 * config/ia64/ia64.md (doloop_end): Update for new interface.
1339 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
1340 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
1342 * config/mep/mep.c (mep_emit_doloop): Likewise.
1343 (TARGET_CAN_USE_DOLOOP_P): Define.
1344 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
1345 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
1346 * config/s390/s390.md (doloop_end): Update for new interface.
1347 * config/sh/sh.md (doloop_end): Likewise.
1348 * config/spu/spu.md (doloop_end): Likewise.
1349 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
1350 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
1351 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
1352 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
1353 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
1354 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
1356 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
1358 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
1361 * doc/extend.texi: Move Cilk Plus Builtins node before Other
1364 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
1365 Joseph Myers <joseph@codesourcery.com>
1367 * ginclude/stdatomic.h: New file.
1368 * Makefile.in (USER_H): Add stdatomic.h.
1370 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1372 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
1375 2013-11-08 Jeff Law <law@redhat.com>
1377 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
1378 * tree-ssa-threadupdate.c (delete_thread_path): New function.
1379 (ssa_redirect_edges, thread_block_1): Use it.
1380 (thread_through_loop_header, mark_threaded_blocks): Likewise.
1381 (thread_through_all_blocks, register_jump_thread): Likewise.
1382 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
1384 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
1386 * config/arm/aarch-common.c
1387 (search_term): New typedef.
1388 (shift_rtx_costs): New array.
1389 (arm_rtx_shift_left_p): New.
1390 (arm_find_sub_rtx_with_search_term): Likewise.
1391 (arm_find_sub_rtx_with_code): Likewise.
1392 (arm_early_load_addr_dep): Add sanity checking.
1393 (arm_no_early_alu_shift_dep): Likewise.
1394 (arm_no_early_alu_shift_value_dep): Likewise.
1395 (arm_no_early_mul_dep): Likewise.
1396 (arm_no_early_store_addr_dep): Likewise.
1398 2013-11-08 Richard Biener <rguenther@suse.de>
1400 PR tree-optimization/59047
1401 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
1403 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
1405 * common.opt (fcheck-pointer-bounds): Move to ...
1406 * c-family/c.opt: ... here.
1407 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
1408 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
1409 * langhooks.h (lang_hooks): Remove chkp_supported field.
1410 * toplev.c (process_options): Remove chkp_supported check.
1412 2013-11-08 Richard Biener <rguenther@suse.de>
1414 PR tree-optimization/59038
1415 PR tree-optimization/58955
1416 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
1417 previous change. Handle known dependences correctly.
1419 2013-11-08 Tom de Vries <tom@codesourcery.com>
1421 * config/rs6000/t-xilinx: Remove duplicate contents.
1423 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
1424 Joseph Myers <joseph@codesourcery.com>
1426 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
1427 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
1428 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
1429 (struct tree_base): Add atomic_flag field.
1430 * tree.h (TYPE_ATOMIC): New accessor macro.
1431 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
1432 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
1433 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
1434 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
1435 * tree.c (set_type_quals): Set TYPE_ATOMIC.
1436 (find_atomic_core_type): New function.
1437 (build_qualified_type): Adjust alignment for qualified types.
1438 (build_atomic_base): New function
1439 (build_common_tree_nodes): Build atomicQI_type_node,
1440 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
1442 * print-tree.c (print_node): Print atomic qualifier.
1443 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
1444 * target.def (atomic_assign_expand_fenv): New hook.
1445 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
1446 * doc/tm.texi: Regenerate.
1447 * targhooks.c (default_atomic_assign_expand_fenv): New function.
1448 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
1449 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
1450 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
1452 * config/i386/i386.c (enum ix86_builtins): Add
1453 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
1454 IX86_BUILTIN_FNCLEX.
1455 (bdesc_special_args): Add __builtin_ia32_fnstenv,
1456 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
1457 (ix86_expand_builtin): Handle the new built-in functions.
1458 (ix86_atomic_assign_expand_fenv): New function.
1459 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
1460 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
1461 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
1462 (fnstenv, fldenv, fnstsw, fnclex): New insns.
1464 2013-11-07 Steve Ellcey <sellcey@mips.com>
1466 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
1467 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
1468 (MULTILIB_DIRNAMES): Add fp64 directory.
1469 (MULTILIB_EXCEPTIONS): Add new exclusions.
1471 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
1473 * gimplify.c (gimple_regimplify_operands): Do not set
1475 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
1476 (rewrite_close_phi_out_of_ssa): Same.
1477 (rewrite_phi_out_of_ssa): Same.
1478 (rewrite_degenerate_phi): Same.
1479 (handle_scalar_deps_crossing_scop_limits): Same.
1480 * tree-if-conv.c (predicate_scalar_phi): Same.
1481 * tree-parloops.c (create_loads_for_reductions): Same.
1482 (create_final_loads_for_reduction): Same.
1483 (create_loads_and_stores_for_name): Same.
1484 (transform_to_exit_first_loop): Same.
1485 (create_parallel_loop): Same.
1486 * tree-ssa-loop-im.c
1487 (move_computations_dom_walker::before_dom_children): Same.
1488 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
1489 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
1490 * tree-ssa-propagate.c (substitute_and_fold): Same.
1491 * tree-vect-loop.c (vect_finalize_reduction): Same.
1492 * tree-vect-stmts.c (vectorizable_call): Same.
1494 2013-11-07 Mike Stump <mikestump@comcast.net>
1496 * config/pdp11/pdp11.c: Include dbxout.h.
1497 * config/picochip/picochip.c: Likewise.
1499 2013-11-07 Cong Hou <congh@google.com>
1501 PR tree-optimization/56764
1502 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
1503 Combine alias checks if it is possible to amortize the runtime
1504 overhead. Return the number of alias checks after merging.
1505 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
1506 Use the function vect_create_cond_for_alias_checks () to check
1507 the number of alias checks.
1509 2013-11-07 Jeff Law <law@redhat.com>
1511 * varpool.c (ctor_for_folding): Fix typo in comment.
1513 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
1515 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
1516 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
1518 2013-11-07 Jeff Law <law@redhat.com>
1520 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
1522 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
1523 No longer check if we have __builtin_trap, assume it's available.
1525 2013-11-07 Diego Novillo <dnovillo@google.com>
1527 * attribs.c (lookup_scoped_attribute_spec): Make static.
1528 (get_attribute_namespace): Likewise.
1529 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
1530 (validate_arglist): Move earlier in the file. Make static.
1531 (expand_stack_restore): Move from stmt.c
1532 (expand_stack_save): Move from stmt.c
1533 (rewrite_call_expr_array): Move earlier in the file.
1534 (rewrite_call_expr_valist): Likewise.
1535 * cfgexpand.c: Include hard-reg-set.h before tree.h
1538 (expand_asm_loc): Move from stmt.c.
1539 (n_occurrences): Move from stmt.c.
1540 (check_operand_nalternatives): Move from stmt.c.
1541 (tree_conflicts_with_clobbers_p): Move from stmt.c.
1542 (expand_asm_operands): Move from stmt.c
1543 (expand_asm_stmt): Move from stmt.c
1544 (expand_computed_goto): Move from stmt.c
1545 (expand_goto): Move from stmt.c
1546 (expand_null_return_1): Move from stmt.c
1547 (expand_null_return): Move from stmt.c
1548 (expand_value_return): Move from stmt.c
1549 (expand_return): Move from stmt.c
1550 (expand_main_function): Move from function.c
1551 (stack_protect_prologue): Move from function.c
1552 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
1553 (build_function_decl_skip_args): Move from tree.c.
1554 * explow.c (tree_expr_size): Move from tree.c.
1555 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
1556 (fields_length): Move from tree.c.
1557 * fold-const.c (size_low_cst): Move from tree.c.
1558 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
1559 (tree_expr_nonzero_p): Make static. Move earlier in the file.
1560 (fold_build3_initializer_loc): Remove.
1561 (tree_invalid_nonnegative_warnv_p): Make static.
1562 * function.c (expand_main_function): Move to cfgexpand.c.
1563 (stack_protect_prologue): Move to cfgexpand.c.
1564 (set_insn_locations): Move earlier in the file.
1565 * gimple-fold.c: Include langhooks.h.
1566 (truth_type_for): Move from tree.c.
1567 * print-tree.c (print_vec_tree): Remove.
1568 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
1569 (expand_goto): Move to cfgexpand.c.
1570 (n_occurrences): Move to cfgexpand.c.
1571 (expand_asm_loc): Move to cfgexpand.c
1572 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
1573 (expand_asm_operands): Move to cfgexpand.c.
1574 (expand_asm_stmt): Move to cfgexpand.c.
1575 (check_operand_nalternatives): Move to cfgexpand.c
1576 (expand_null_return): Move to cfgexpand.c.
1577 (expand_value_return): Move to cfgexpand.c.
1578 (expand_null_return_1): Move to cfgexpand.c.
1579 (expand_return): Move to cfgexpand.c.
1580 (expand_stack_save): Move to builtins.c.
1581 (expand_stack_restore): Move to builtins.c
1582 * symtab.c: Include output.h.
1583 (decl_assembler_name_hash): Move from tree.c.
1584 (decl_assembler_name_equal): Move from tree.c.
1585 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
1586 * tree-eh.c (in_array_bounds_p): Move from tree.c.
1587 (range_in_array_bounds_p): Move from tree.c.
1588 * tree-object-size.c (fini_object_sizes): Make static.
1589 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
1590 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
1591 * tree.c (decl_assembler_name_equal): Move to symtab.c.
1592 (tree_expr_size): Move to explow.c.
1593 (decl_assembler_name_hash): Move to symtab.c.
1594 (real_twop): Remove.
1595 (tree_expr_size): Move to explow.c.
1596 (stabilize_reference_1): Move earlier in the file. Make static.
1597 (omp_remove_redundant_declare_simd_attrs): Remove.
1598 (simple_cst_list_equal): Move earlier in the file. Make static.
1599 (size_low_cst): Move to fold-const.c.
1600 (build_type_no_quals): Remove.
1601 (build_function_type_skip_args): Move to cgraphclones.c.
1602 (build_function_decl_skip_args): Move to cgraphclones.c.
1603 (in_array_bounds_p): Move to tree-eh.c.
1604 (range_in_array_bounds_p): Move to tree-eh.c.
1605 (truth_type_for): Move to gimple-fold.c.
1606 (list_equal_p): Remove.
1607 * tree.h (decl_assembler_name_equal): Remove.
1608 (decl_assembler_name_hash): Remove.
1609 (truth_type_for): Remove.
1610 (build_type_no_quals): Remove.
1611 (build_function_decl_skip_args): Remove.
1612 (in_array_bounds_p): Remove.
1613 (range_in_array_bounds_p): Remove.
1614 (size_low_cst): Remove.
1615 (omp_remove_redundant_declare_simd_attrs): Remove.
1616 (tree_expr_size): Remove.
1617 (fields_length): Remove.
1618 (stabilize_reference_1): Remove.
1619 (expand_goto): Remove.
1620 (expand_stack_save): Remove.
1621 (expand_stack_restore): Remove.
1622 (expand_return): Remove.
1623 (fold_build3_initializer_loc): Remove.
1624 (tree_expr_nonzero_p): Remove.
1625 (tree_invalid_nonnegative_warnv_p): Remove.
1626 (tree_expr_nonzero_warnv_p): Remove.
1627 (fold_builtin_snprintf_chk): Remove.
1628 (validate_arglist): Remove.
1629 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
1630 (simple_cst_list_equal): Remove.
1631 (real_twop): Remove.
1632 (expand_main_function): Remove.
1633 (stack_protect_prologue): Remove.
1634 (print_vec_tree): Remove.
1635 (lookup_scoped_attribute_spec): Remove.
1636 (get_attribute_namespace): Remove.
1637 (expand_computed_goto): Remove.
1638 (expand_asm_stmt): Remove.
1639 (list_equal_p): Remove.
1640 (ssa_name_nonnegative_p): Remove.
1641 (fini_object_sizes): Remove.
1642 (addr_expr_of_non_mem_decl_p): Remove.
1643 (is_tm_safe_or_pure): Move to trans-mem.c.
1644 (more_const_call_expr_args_p): Remove.
1645 (save_vtable_map_decl): Remove.
1647 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
1649 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
1650 linker plugins, too.
1652 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
1653 link to libgomp and its dependencies.
1654 * config/ia64/hpux.h (LIB_SPEC): Likewise.
1655 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
1656 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
1657 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
1659 2013-11-07 Jakub Jelinek <jakub@redhat.com>
1661 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
1662 (determine_value_range): Add loop argument. Use get_range_info to
1664 (bound_difference): Adjust caller.
1666 2013-11-07 Richard Biener <rguenther@suse.de>
1667 Jakub Jelinek <jakub@redhat.com>
1669 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
1670 latches from header PHI arguments from the latch edge.
1672 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
1675 * varasm.c (output_constant): Handle NULLPTR_TYPE.
1677 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
1679 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
1680 misaligned_prologue_used when it has been set.
1682 2013-11-07 Yury Gribov <y.gribov@samsung.com>
1683 Jakub Jelinek <jakub@redhat.com>
1686 * asan.c (get_mem_refs_of_builtin_call): Allow
1687 integer literals as addresses in instrumented builtins.
1689 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1691 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
1692 Explain why plus_constant is not used.
1694 2013-11-07 Richard Biener <rguenther@suse.de>
1696 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
1697 (canonicalize_value): ... this. Also handle stripping of
1699 (get_value, set_lattice_value, get_value_for_expr): Adjust.
1700 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
1701 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
1703 2013-11-07 Richard Biener <rguenther@suse.de>
1705 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
1707 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
1710 * config/i386/i386.md (push peepholer/splitter): Use Pmode
1711 with stack_pointer_rtx.
1713 2013-11-07 Bin Cheng <bin.cheng@arm.com>
1715 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
1716 using operand_equal_p.
1718 2013-11-07 Bin Cheng <bin.cheng@arm.com>
1720 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
1721 * tree-affine.c (get_inner_reference_aff): Return base.
1722 * tree-affine.h (get_inner_reference_aff): Change prototype.
1724 2013-11-06 Tobias Burnus <burnus@net-b.de>
1726 * doc/invoke.texi (Wdate-time): Fix typo.
1728 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
1730 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
1731 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
1732 fp_arith_reg_operand instead of arith_reg_operand.
1734 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
1736 * config/sh/sh.md (adddi3): Remove empty constraints.
1737 Remove can_create_pseudo_p and arith_reg_operand check.
1738 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
1739 Split before reload.
1741 2013-11-06 Jeff Law <law@redhat.com>
1742 Tom Tromey <tromey@redhat.com>
1744 * gdbinit.in: Disable strict type checking.
1746 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
1748 * tree-pass.h (make_pass_live_range_shrinkage): New external.
1749 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
1750 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
1751 (rest_of_handle_live_range_shrinkage): Ditto
1752 (class pass_live_range_shrinkage): Ditto.
1753 (pass_data_live_range_shrinkage): Ditto.
1754 (make_pass_live_range_shrinkage): Ditto.
1755 * sched-int.h (initialize_live_range_shrinkage): New prototype.
1756 (finish_live_range_shrinkage): Ditto.
1757 * sched-deps.c (create_insn_reg_set): Make void return value.
1758 * passes.def: Add pass_live_range_shrinkage.
1759 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
1760 * haifa-sched.c (live_range_shrinkage_p): New.
1761 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
1763 (rank_for_schedule): Add code for pressure relief through live
1765 (schedule_insn): Print more debug info.
1766 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
1767 through live range shrinkage.
1768 * doc/invoke.texi (-flive-range-shrinkage): New.
1769 * common.opt (flive-range-shrinkage): New.
1771 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
1774 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
1775 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
1776 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
1777 RTXes that return in AVX256 register.
1779 2013-11-06 Richard Biener <rguenther@suse.de>
1781 PR tree-optimization/58653
1782 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
1783 (prepare_initializers_chain): Adjust.
1785 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
1787 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
1788 (gimple_in_transaction): Use bb_in_transaction.
1789 * basic-block.h (bb_in_transaction): Relocate here and rename.
1790 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
1792 2013-11-06 Richard Biener <rguenther@suse.de>
1794 * tree.c (drop_tree_overflow): New function.
1795 * tree.h (drop_tree_overflow): Declare.
1796 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
1797 * tree-vrp.c (range_int_cst_singleton_p): Use
1798 is_overflow_infinity instead of testing TREE_OVERFLOW.
1799 (extract_range_from_assert): Likewise.
1800 (zero_nonzero_bits_from_vr): Likewise.
1801 (extract_range_basic): Likewise.
1802 (register_new_assert_for): Use drop_tree_overflow.
1803 (vrp_visit_phi_node): Likewise.
1805 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
1807 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
1808 checking for leaf functions without dynamic stack allocation.
1809 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
1810 (ia64_expand_prologue): Likewise.
1811 * config/mips/mips.c (mips_expand_prologue): Likewise.
1812 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
1813 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
1814 (sparc_flat_expand_prologue): Likewise.
1816 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
1818 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
1819 (__ST3_LANE_FUNC): Likewise.
1820 (__ST4_LANE_FUNC): Likewise.
1822 2013-11-06 Nick Clifton <nickc@redhat.com>
1824 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
1825 name returned by msp430_mcu_name.
1826 (LIB_SPEC): If a -T option has not been specified then set a
1827 default, mcu-specific, linker script.
1828 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
1829 * config/msp430/msp430.c (msp430x_names): Likewise.
1830 Alpha sort the names for ease of comparison.
1831 (msp430_mcu_name): New function: Returns a string suitable for
1832 use as a C preprocessor symbol based upon the name of the MCU
1834 (msp430_option_override): Accept msp430x and msp430xv2 as generic
1836 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
1838 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
1839 unless it is the last part of a spec substring.
1840 * doc/invoke.texi (Spec Files): Document space insertion
1843 2013-11-06 Christian Bruel <christian.bruel@st.com>
1845 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
1846 Factorize probabilities, Use adjust_address instead of
1847 adjust_automodify_address when possible. Enable for optimize.
1848 (sh_expand_strlen): New function.
1849 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
1850 * config/sh/sh.md (strlensi): New pattern.
1851 (UNSPEC_BUILTIN_STRLEN): Define.
1853 2013-11-06 Jakub Jelinek <jakub@redhat.com>
1856 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
1857 (expand_assignment): If *bitpos is negative, set *offset
1858 and adjust *bitpos, so that it is not negative.
1860 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
1862 * config/i386/bdver3.md : Added two additional decoder units
1863 to support issue rate of 4 and remodeled vector unit.
1864 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
1865 architectures is set to 4.
1866 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
1867 lookahead is set to 4 for BD architectures.
1869 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1871 * config/rs6000/rs6000.c (rs6000_option_override_internal):
1872 Remove restriction against use of VSX instructions when generating
1873 code for little endian mode.
1875 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1877 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
1878 for both big and little endian.
1879 (mulv8hi3): Swap input operands for merge high and merge low
1880 instructions for little endian.
1882 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1884 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
1885 define_insn to define_expand that uses even patterns for big
1886 endian and odd patterns for little endian.
1887 (vec_widen_smult_even_v16qi): Likewise.
1888 (vec_widen_umult_even_v8hi): Likewise.
1889 (vec_widen_smult_even_v8hi): Likewise.
1890 (vec_widen_umult_odd_v16qi): Likewise.
1891 (vec_widen_smult_odd_v16qi): Likewise.
1892 (vec_widen_umult_odd_v8hi): Likewise.
1893 (vec_widen_smult_odd_v8hi): Likewise.
1894 (altivec_vmuleub): New define_insn.
1895 (altivec_vmuloub): Likewise.
1896 (altivec_vmulesb): Likewise.
1897 (altivec_vmulosb): Likewise.
1898 (altivec_vmuleuh): Likewise.
1899 (altivec_vmulouh): Likewise.
1900 (altivec_vmulesh): Likewise.
1901 (altivec_vmulosh): Likewise.
1903 2013-11-05 Mike Stump <mikestump@comcast.net>
1905 * Makefile.in (mostlyclean): Remove c-family objects.
1907 2013-11-05 Ian Lance Taylor <iant@google.com>
1909 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
1910 If possible, add .cfi directives to record change to bx.
1911 * config/i386/i386.c (ix86_emit_cfi): New function.
1912 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
1914 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
1916 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
1917 immediately after a label for a tablejump pattern.
1919 * config/arm/arm.c (is_jump_table): Remove.
1920 (create_fix_barrier): Use tablejump_p instead.
1921 (arm_reorg): Likewise.
1922 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
1923 (thumb2_output_casesi): Likewise.
1924 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
1925 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
1926 casesi_shift_media, casesi_load_media): Likewise.
1927 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
1928 * config/microblaze/microblaze.md: Likewise.
1930 2013-11-05 Tobias Burnus <burnus@net-b.de>
1932 * doc/invoke.texi (-Wdate-time): Document.
1934 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
1936 * double-int.c (lshift_double, rshift_double): Remove
1937 SHIFT_COUNT_TRUNCATED handling.
1939 2013-11-05 Jeff Law <law@redhat.com>
1941 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
1942 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
1943 * gimple-ssa-isolate-paths.c: New file.
1944 * gimple.c (check_loadstore): New function.
1945 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
1946 Verify OP is in the argument list and the argument corresponding
1947 to OP is a pointer type. Use operand_equal_p rather than
1948 pointer equality when testing if OP is on the nonnull list.
1949 Use check_loadstore rather than count_ptr_derefs. Handle
1950 GIMPLE_RETURN statements.
1951 * tree-vrp.c (infer_nonnull_range): Remove.
1952 * gimple.h (infer_nonnull_range): Declare.
1953 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
1954 * passes.def: Add pass_isolate_erroneous_paths.
1955 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
1956 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
1957 * tree-ssa.c (struct count_ptr_d): Remove.
1958 (count_ptr_derefs, count_uses_and_derefs): Remove.
1959 * tree-ssa.h (count_uses_and_derefs): Remove.
1961 2013-11-05 Jakub Jelinek <jakub@redhat.com>
1963 PR rtl-optimization/58997
1964 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
1965 get_iv_value to be in iv->mode rather than iv->extend_mode.
1966 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
1967 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
1968 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
1971 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
1973 * gimple.h: Move some prototypes to gimple-expr.h and add to include
1975 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
1976 Move to gimple-expr.h.
1977 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
1978 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
1979 is_gimple_id, virtual_operand_p, is_gimple_addressable,
1980 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
1982 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
1983 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
1984 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
1985 is_gimple_address, is_gimple_invariant_address,
1986 is_gimple_ip_invariant_address, is_gimple_min_invariant,
1987 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
1988 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
1989 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
1990 gimple_decl_printable_name, useless_type_conversion_p,
1991 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
1993 * gimple-expr.c: New File.
1994 (useless_type_conversion_p, gimple_set_body, gimple_body,
1995 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
1996 gimple_can_coalesce_p, extract_ops_from_tree_1,
1997 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
1998 is_gimple_address, is_gimple_invariant_address,
1999 is_gimple_ip_invariant_address, is_gimple_min_invariant,
2000 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
2001 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
2002 is_gimple_mem_ref_addr): Relocate here.
2003 * Makefile.in (OBJS): Add gimple-expr.o.
2005 2013-11-05 David Malcolm <dmalcolm@redhat.com>
2007 * gengtype-parse.c (struct_field_seq): Support empty structs.
2009 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2011 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
2013 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
2015 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
2017 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
2019 (LONG_DOUBLE_TYPE_SIZE): New define.
2020 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
2022 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
2025 * doc/extend.texi [visibility ("visibility_type")]: Add example
2026 about visibility attribute on namespace declaration.
2028 2013-11-05 Richard Biener <rguenther@suse.de>
2031 * passes.def (all_passes): Start with pass_fixup_cfg again.
2033 2013-11-05 Richard Biener <rguenther@suse.de>
2035 PR tree-optimization/58955
2036 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
2039 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2041 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
2043 (vec_pack_ufix_trunc_v2df): Likewise.
2045 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
2048 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
2049 pattern, instead of word_mode.
2051 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
2053 (set_storage_via_setmem): Likewise.
2055 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
2057 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
2058 gimple.h and the rest of the condition in eliminate_build.
2059 (eliminate_build): Call new routine.
2060 * gimple.h (phi_ssa_name_p): Delete.
2062 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
2064 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
2065 case of no prefix and reserving zero slots, because when that's the
2066 case we'll never get here.
2067 * vec.h (va_heap::reserve): Don't try and handle
2068 vec_prefix::calculate_allocation returning zero because that should
2071 2013-11-05 Richard Biener <rguenther@suse.de>
2074 * tree-dfa.c (get_ref_base_and_extent): Merge common code
2075 in MEM_REF and TARGET_MEM_REF handling. Make sure to
2076 process trailing array detection before diving into the
2077 view-converted object (and possibly apply some extra offset).
2079 2013-11-05 Joseph Myers <joseph@codesourcery.com>
2081 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
2083 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
2085 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
2087 PR tree-optimization/58958
2088 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
2089 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
2091 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
2093 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
2094 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
2096 2013-11-05 Jakub Jelinek <jakub@redhat.com>
2098 PR tree-optimization/58984
2099 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
2100 set *SIZE_P if non-NULL on success.
2101 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
2103 (ipcp_transform_function): Likewise. Punt if size of access
2104 is different from TYPE_SIZE on v->value's type.
2106 2013-11-05 Tobias Burnus <burnus@net-b.de>
2108 * doc/invoke.texi (-fopenmp-simd): Document new option.
2109 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
2110 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
2111 * tree.c (attribute_value_equal): Ditto.
2113 2013-11-04 Wei Mi <wmi@google.com>
2115 * sched-rgn.c (add_branch_dependences): Keep insns in
2116 a SCHED_GROUP at the end of BB to remain their location.
2118 2013-11-04 Wei Mi <wmi@google.com>
2120 * config/i386/i386.c (memory_address_length): Extract a part
2121 of code to rip_relative_addr_p.
2122 (rip_relative_addr_p): New Function.
2123 (ix86_macro_fusion_p): Ditto.
2124 (ix86_macro_fusion_pair_p): Ditto.
2125 * config/i386/i386.h: Add new tune features about macro-fusion.
2126 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
2127 * doc/tm.texi: Generated.
2128 * doc/tm.texi.in: Ditto.
2129 * haifa-sched.c (try_group_insn): New Function.
2130 (group_insns_for_macro_fusion): Ditto.
2131 (sched_init): Call group_insns_for_macro_fusion.
2132 * target.def: Add two hooks: macro_fusion_p and
2133 macro_fusion_pair_p.
2135 2013-11-04 Kostya Serebryany <kcc@google.com>
2137 Update to match the changed asan API.
2138 * asan.c (asan_function_start): New function.
2139 (asan_emit_stack_protection): Update the string stored in the
2140 stack red zone to match new API. Store the PC of the current
2141 function in the red zone.
2142 (asan_global_struct): Update the __asan_global definition to match
2144 (asan_add_global): Ditto.
2145 * asan.h (asan_function_start): New prototype.
2146 * final.c (final_start_function): Call asan_function_start.
2147 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
2150 2013-11-04 Wei Mi <wmi@google.com>
2152 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
2153 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
2154 * config/i386/i386.c (ix86_option_override_internal): Ditto.
2155 (ix86_issue_rate): Ditto.
2156 (ix86_adjust_cost): Ditto.
2157 (ia32_multipass_dfa_lookahead): Ditto.
2158 (ix86_sched_init_global): Ditto.
2159 (get_builtin_code_for_version): Ditto.
2160 * config/i386/i386.h (enum target_cpu_default): Ditto.
2161 (enum processor_type): Ditto.
2162 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
2164 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
2166 PR rtl-optimization/58967
2167 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
2168 !lra_in_progress for mode sizes bigger word.
2170 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2172 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
2173 arguments to merge instruction for little endian.
2174 (vec_widen_umult_lo_v16qi): Likewise.
2175 (vec_widen_smult_hi_v16qi): Likewise.
2176 (vec_widen_smult_lo_v16qi): Likewise.
2177 (vec_widen_umult_hi_v8hi): Likewise.
2178 (vec_widen_umult_lo_v8hi): Likewise.
2179 (vec_widen_smult_hi_v8hi): Likewise.
2180 (vec_widen_smult_lo_v8hi): Likewise.
2182 2013-11-04 Ian Lance Taylor <iant@google.com>
2184 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
2185 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
2186 builtins that take pointers.
2187 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
2188 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
2189 (run_gcc): Pass -fnon-call-exceptions.
2191 2013-11-04 Jakub Jelinek <jakub@redhat.com>
2193 * optabs.c (expand_vec_perm): Revert one incorrect line from
2196 PR tree-optimization/58978
2197 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
2198 use_stmt by single_imm_use directly. Only call single_imm_use
2201 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
2203 PR rtl-optimization/58968
2204 * lra-spills.c (return_regno_p): New function.
2205 (lra_final_code_change): Use it.
2207 2013-11-04 Joseph Myers <joseph@codesourcery.com>
2209 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
2210 * target.def (float_exceptions_rounding_supported_p): New hook.
2211 * targhooks.c (default_float_exceptions_rounding_supported_p): New
2213 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
2214 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
2216 * doc/tm.texi: Regenerate.
2217 * config.gcc (powerpc*-*-linux*): Set extra_objs.
2218 * config/rs6000/rs6000-linux.c: New file.
2219 * config/rs6000/rs6000-protos.h
2220 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
2221 * config/rs6000/linux.h
2222 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
2223 * config/rs6000/linux64.h
2224 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
2225 * config/rs6000/t-linux (rs6000-linux.o): New rule.
2226 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
2228 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2230 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
2231 Replace the define_insn_and_split with a define_insn and two
2232 define_splits, with the split after reload re-permuting the source
2233 register to its original value.
2234 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
2235 (*vsx_le_perm_store_v8hi): Likewise.
2236 (*vsx_le_perm_store_v16qi): Likewise.
2238 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2240 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
2243 2013-11-04 Jakub Jelinek <jakub@redhat.com>
2245 PR tree-optimization/58946
2246 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
2247 bbs with bbinfo[idx].op != NULL before all blocks with
2248 bbinfo[idx].op == NULL.
2250 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
2252 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
2253 (avr_dump_double_int_hex): Likewise.
2254 (avr_log_vadump): Remove %D and %X handling.
2255 * config/avr/avr.c (avr_double_int_push_digit): Delete.
2256 (avr_map_op_t): Change map from double_int to unsigned int.
2257 (avr_map_op): Update accordingly.
2258 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
2259 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
2260 unsigned ints rather than double_ints.
2262 2013-11-03 Marek Polacek <polacek@redhat.com>
2264 Implement -fsanitize=vla-bound.
2265 * opts.c (common_handle_option): Handle vla-bound.
2266 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
2267 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
2268 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
2270 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2272 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
2275 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
2277 * config/i386/constraints.md (Ts, Tv): New address constrains.
2278 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
2279 constraint for address_no_seg_operand.
2280 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
2281 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
2282 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
2283 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
2284 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
2285 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
2286 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
2287 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
2288 constraint for vsib_address_operand.
2290 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
2292 * gcse.c (pre_delete): Remove references to regmove from comments.
2293 * recog.c: (validate_replace_rtx_1): Likewise.
2294 * config/rl78/rl78.c: Likewise.
2295 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
2296 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
2297 * common/config/mmix/mmix-common.c: Likewise.
2299 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
2301 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
2302 * gimplify.c (gimplify_compound_lval): Likewise.
2303 * graphite-clast-to-gimple.c (gloog): Likewise.
2304 * graphite-dependences.c (loop_is_parallel_p): Likewise.
2305 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
2306 (limit_scop); Likewise.
2307 (build_scops): Likewise.
2308 (dot_scop): Likewise.
2309 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
2310 (build_scop_drs): Likewise.
2311 (insert_stmts): Likewise.
2312 (insert_out_of_ssa_copy): Likewise.
2313 (remove_phi): Likewise.
2314 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
2315 * hw-doloop.c (discover_loop): Likewise.
2316 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
2317 * tree-dfa.c (dump_enumerated_decls): Likewise.
2318 * tree-if-conv.c (if_convertable_loop_p): Likewise.
2319 * tree-inline.c (tree_function_versioning): Likewise.
2320 * tree-loop-distribution.c (build_rdg): Likewise.
2321 (rdg_flag_vertex_and_dependent): Likewise.
2322 (distribute_loop): Likewise.
2323 * tree-parloops.c (loop_parallel_p): Likewise.
2324 (eliminate_local_variables): Likewise.
2325 (separate_decls_in_region): Likewise.
2326 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
2327 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
2328 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
2329 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
2330 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
2331 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
2332 (vectorizable_condition): Likewise.
2334 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
2336 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
2337 * configure: Regenerate.
2338 * config/i386/i386.md (*movdi_internal): Change
2339 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
2340 (*movdf_internal): Ditto.
2341 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
2342 * config/i386/sse.md (vec_concatv2di): Output interunit movq
2343 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
2345 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
2347 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
2350 2013-10-31 David Malcolm <dmalcolm@redhat.com>
2352 Automated part of renaming of symtab_node_base to symtab_node.
2354 Patch autogenerated by rename_symtab.py from
2355 https://github.com/davidmalcolm/gcc-refactoring-scripts
2356 revision 58bb219cc090b2f4516a9297d868c245495ee622
2357 with ChangeLog entry fixed up by hand.
2359 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
2361 (cgraph_node_for_asm): Likewise.
2362 * cgraph.h (symtab_node_base): Likewise.
2363 (cgraph_node): Likewise.
2364 (varpool_node): Likewise.
2365 (is_a_helper <cgraph_node>::test): Likewise.
2366 (is_a_helper <varpool_node>::test): Likewise.
2367 (symtab_nodes): Likewise.
2368 (symtab_register_node): Likewise.
2369 (symtab_unregister_node): Likewise.
2370 (symtab_remove_node): Likewise.
2371 (symtab_get_node): Likewise.
2372 (symtab_node_for_asm): Likewise.
2373 (symtab_node_asm_name): Likewise.
2374 (symtab_node_name): Likewise.
2375 (symtab_insert_node_to_hashtable): Likewise.
2376 (symtab_add_to_same_comdat_group): Likewise.
2377 (symtab_dissolve_same_comdat_group_list): Likewise.
2378 (dump_symtab_node): Likewise.
2379 (debug_symtab_node): Likewise.
2380 (dump_symtab_base): Likewise.
2381 (verify_symtab_node): Likewise.
2382 (verify_symtab_base): Likewise.
2383 (symtab_used_from_object_file_p): Likewise.
2384 (symtab_alias_ultimate_target): Likewise.
2385 (symtab_resolve_alias): Likewise.
2386 (fixup_same_cpp_alias_visibility): Likewise.
2387 (symtab_for_node_and_aliases): Likewise.
2388 (symtab_nonoverwritable_alias): Likewise.
2389 (availability symtab_node_availability): Likewise.
2390 (symtab_semantically_equivalent_p): Likewise.
2391 (fixup_same_cpp_alias_visibility): Likewise.
2392 (symtab_prevail_in_asm_name_hash): Likewise.
2394 (varpool): Likewise.
2395 (varpool_first_variable): Likewise.
2396 (varpool_next_variable): Likewise.
2397 (varpool_first_static_initializer): Likewise.
2398 (varpool_next_static_initializer): Likewise.
2399 (varpool_first_defined_variable): Likewise.
2400 (varpool_next_defined_variable): Likewise.
2401 (cgraph_first_defined_function): Likewise.
2402 (cgraph_next_defined_function): Likewise.
2403 (cgraph_first_function): Likewise.
2404 (cgraph_next_function): Likewise.
2405 (cgraph_first_function_with_gimple_body): Likewise.
2406 (cgraph_next_function_with_gimple_body): Likewise.
2407 (symtab_alias_target): Likewise.
2408 (symtab_real_symbol_p): Likewise.
2409 (symtab_can_be_discarded): Likewise.
2410 * cgraphbuild.c (mark_address): Likewise.
2411 (mark_load): Likewise.
2412 (mark_store): Likewise.
2413 * cgraphunit.c (decide_is_symbol_needed): Likewise.
2415 (enqueue_node): Likewise.
2416 (referred_to_p): Likewise.
2417 (cgraph_process_same_body_aliases): Likewise.
2418 (analyze_functions): Likewise.
2419 (handle_alias_pairs): Likewise.
2420 (output_weakrefs): Likewise.
2421 (compile): Likewise.
2422 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2423 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2424 * ipa-prop.c (remove_described_reference): Likewise.
2425 (try_decrement_rdesc_refcount): Likewise.
2426 (ipa_edge_duplication_hook): Likewise.
2427 * ipa-ref.c (ipa_record_reference): Likewise.
2428 (ipa_maybe_record_reference): Likewise.
2429 (ipa_clone_ref): Likewise.
2430 (ipa_clone_references): Likewise.
2431 (ipa_clone_referring): Likewise.
2432 (ipa_find_reference): Likewise.
2433 (ipa_remove_stmt_references): Likewise.
2434 (ipa_clear_stmts_in_references): Likewise.
2435 * ipa-ref.h (symtab_node_base): Likewise.
2436 (ipa_ref): Likewise.
2437 (ipa_record_reference): Likewise.
2438 (ipa_maybe_record_reference): Likewise.
2439 (ipa_clone_references): Likewise.
2440 (ipa_clone_referring): Likewise.
2441 (ipa_clone_ref): Likewise.
2442 (ipa_find_reference): Likewise.
2443 (ipa_remove_stmt_references): Likewise.
2444 (ipa_clear_stmts_in_references): Likewise.
2445 * ipa-reference.c (ipa_reference_write_optimization_summary):
2447 * ipa.c (enqueue_node): Likewise.
2448 (process_references): Likewise.
2449 (walk_polymorphic_call_targets): Likewise.
2450 (symtab_remove_unreachable_nodes): Likewise.
2451 (address_taken_from_non_vtable_p): Likewise.
2452 (comdat_can_be_unshared_p_1): Likewise.
2453 (comdat_can_be_unshared_p): Likewise.
2454 (can_replace_by_local_alias): Likewise.
2455 (function_and_variable_visibility): Likewise.
2456 * is-a.h: Likewise (within example in comment).
2457 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
2458 (lto_symtab_encoder_encode): Likewise.
2459 (lto_symtab_encoder_delete_node): Likewise.
2460 (lto_symtab_encoder_in_partition_p): Likewise.
2461 (lto_set_symtab_encoder_in_partition): Likewise.
2462 (output_refs): Likewise.
2463 (compute_ltrans_boundary): Likewise.
2464 (output_symtab): Likewise.
2465 (input_node): Likewise.
2466 (input_ref): Likewise.
2467 (input_edge): Likewise.
2468 (input_cgraph_1): Likewise.
2469 (input_refs): Likewise.
2470 (output_cgraph_opt_summary): Likewise.
2471 (input_node_opt_summary): Likewise.
2472 (input_cgraph_opt_section): Likewise.
2473 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
2475 * lto-streamer-out.c (lto_output): Likewise.
2476 (output_symbol_p): Likewise.
2477 (produce_symtab): Likewise.
2478 * lto-streamer.h (lto_encoder_entry): Likewise.
2479 (lto_free_function_in_decl_state_for_node): Likewise.
2480 (lto_symtab_encoder_encode): Likewise.
2481 (lto_symtab_encoder_delete_node): Likewise.
2482 (lto_symtab_encoder_in_partition_p): Likewise.
2483 (lto_set_symtab_encoder_in_partition): Likewise.
2484 (lto_symtab_encoder_lookup): Likewise.
2485 (lsei_node): Likewise.
2486 (lto_symtab_encoder_deref): Likewise.
2487 * symtab.c (symtab_hash): Likewise.
2488 (assembler_name_hash): Likewise.
2489 (symtab_nodes): Likewise.
2490 (hash_node): Likewise.
2491 (eq_node): Likewise.
2492 (hash_node_by_assembler_name): Likewise.
2493 (eq_assembler_name): Likewise.
2494 (insert_to_assembler_name_hash): Likewise.
2495 (unlink_from_assembler_name_hash): Likewise.
2496 (symtab_prevail_in_asm_name_hash): Likewise.
2497 (symtab_register_node): Likewise.
2498 (symtab_insert_node_to_hashtable): Likewise.
2499 (symtab_unregister_node): Likewise.
2500 (symtab_get_node): Likewise.
2501 (symtab_remove_node): Likewise.
2502 (symtab_initialize_asm_name_hash): Likewise.
2503 (symtab_node_for_asm): Likewise.
2504 (symtab_add_to_same_comdat_group): Likewise.
2505 (symtab_dissolve_same_comdat_group_list): Likewise.
2506 (symtab_node_asm_name): Likewise.
2507 (symtab_node_name): Likewise.
2508 (dump_symtab_base): Likewise.
2509 (dump_symtab_node): Likewise.
2510 (dump_symtab): Likewise.
2511 (debug_symtab_node): Likewise.
2512 (verify_symtab_base): Likewise.
2513 (verify_symtab_node): Likewise.
2514 (verify_symtab): Likewise.
2515 (symtab_used_from_object_file_p): Likewise.
2516 (symtab_node_availability): Likewise.
2517 (symtab_alias_ultimate_target): Likewise.
2518 (fixup_same_cpp_alias_visibility): Likewise.
2519 (symtab_resolve_alias): Likewise.
2520 (symtab_for_node_and_aliases): Likewise.
2521 (symtab_for_node_and_aliases): Likewise.
2522 (symtab_nonoverwritable_alias_1): Likewise.
2523 (symtab_nonoverwritable_alias): Likewise.
2524 (symtab_semantically_equivalent_p): Likewise.
2525 * value-prof.c (init_node_map): Likewise.
2526 * varasm.c (find_decl): Likewise.
2527 * varpool.c (varpool_node_for_asm): Likewise.
2528 (varpool_remove_unreferenced_decls): Likewise.
2530 2013-10-31 David Malcolm <dmalcolm@redhat.com>
2532 Manual part of renaming of symtab_node_base to symtab_node.
2534 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
2535 clashes with the preferred name for the base class.
2536 (const_symtab_node): Remove redundant typedef.
2538 2013-10-31 Jakub Jelinek <jakub@redhat.com>
2540 * optabs.c (expand_vec_perm): Avoid vector mode punning
2541 SUBREGs in SET_DEST.
2542 * expmed.c (store_bit_field_1): Likewise.
2543 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
2544 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
2545 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
2546 * config/i386/i386.c (ix86_expand_vector_move_misalign,
2547 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
2548 ix86_expand_sse_unpack, ix86_expand_args_builtin,
2549 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
2550 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
2551 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
2552 expand_vec_perm_vpshufb2_vpermq,
2553 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
2554 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
2555 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
2556 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
2559 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
2561 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
2562 Use gcc_unreachable for unhandled modes. Do not check results of
2563 expand_simple_binop. If not expanded to target, move the result.
2565 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
2566 Shiva Chen <shiva0217@gmail.com>
2568 * config.gcc (nds32*-*-*): Add nds32 target.
2569 * config/nds32/nds32.c: New file.
2570 * config/nds32/nds32.h: New file.
2571 * config/nds32/nds32.md: New file.
2572 * config/nds32/constants.md: New file.
2573 * config/nds32/constraints.md: New file.
2574 * config/nds32/iterators.md: New file.
2575 * config/nds32/nds32-doubleword.md: New file.
2576 * config/nds32/nds32-intrinsic.md: New file.
2577 * config/nds32/nds32_intrinsic.h: New file.
2578 * config/nds32/nds32-modes.def: New file.
2579 * config/nds32/nds32-multiple.md: New file.
2580 * config/nds32/nds32.opt: New file.
2581 * config/nds32/nds32-opts.h: New file.
2582 * config/nds32/nds32-protos.h: New file.
2583 * config/nds32/nds32-peephole2.md: New file.
2584 * config/nds32/pipelines.md: New file.
2585 * config/nds32/predicates.md: New file.
2586 * config/nds32/t-mlibs: New file.
2587 * common/config/nds32: New directory and files.
2589 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
2590 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
2591 * doc/install.texi (Cross-Compiler-Specific Options): Document
2592 --with-nds32-lib for nds32 target.
2593 * doc/extend.texi (Function Attributes, Target Builtins): Document
2594 nds32 specific attributes.
2596 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
2598 * lra-constraints (process_alt_operands): Use the result
2599 elimination register for operand when matching constraints.
2601 2013-10-31 Jakub Jelinek <jakub@redhat.com>
2603 * tree-vrp.c (maybe_set_nonzero_bits): New function.
2604 (remove_range_assertions): Call it.
2606 * tree.c (tree_ctz): New function.
2607 * tree.h (tree_ctz): New prototype.
2608 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
2609 first argument from tree to const_tree.
2610 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
2611 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
2612 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
2614 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
2615 * tree-vect-loop.c (vect_analyze_loop_operations,
2616 vect_transform_loop): Don't force scalar loop for bound just because
2617 number of iterations is unknown, only do it if it is not known to be
2618 a multiple of vectorization_factor.
2619 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
2621 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
2622 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
2623 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
2624 (set_nonzero_bits, get_nonzero_bits): New prototypes.
2625 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
2626 a default def isn't partially constant.
2627 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
2628 is known to be partially zero.
2629 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
2630 to see if a default def isn't partially constant.
2631 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
2632 creation of a range, if VR_RANGE, try to improve nonzero_bits from
2634 (set_nonzero_bits, get_nonzero_bits): New functions.
2636 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
2637 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
2638 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
2639 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
2640 uses but in the condition and ASSERT_EXPR and the other successor of
2641 the predecessor bb is __builtin_unreachable (), set_range_info of the
2642 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
2644 2013-10-31 Martin Jambor <mjambor@suse.cz>
2646 PR rtl-optimization/58934
2648 2013-10-30 Martin Jambor <mjambor@suse.cz>
2649 PR rtl-optimization/10474
2650 * ira.c (find_moveable_pseudos): Do not calculate dominance info
2652 (interesting_dest_for_shprep): New function.
2653 (split_live_ranges_for_shrink_wrap): Likewise.
2654 (ira): Calculate dominance info and df analysis. Call
2655 split_live_ranges_for_shrink_wrap.
2657 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2658 Yury Gribov <y.gribov@samsung.com>
2661 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
2662 shadow pointer to avoid clobbering the main one.
2664 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2666 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
2668 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
2671 * ira-color.c (update_costs_from_copies): Add new parameter. Use
2672 it for calling update_costs_from_allocno.
2673 (assign_hard_reg): Call restore_costs_from_copies only for
2674 !retry_p. Pass new argument to update_costs_from_copies.
2675 (color_pass): Pass new argument to update_costs_from_copies.
2676 (ira_mark_allocation_change): Ditto.
2678 2013-10-30 Sharad Singhai <singhai@google.com>
2681 * opts.c (common_handle_option): Remove deprecated option
2682 -ftree-vectorizer-verbose.
2683 * doc/invoke.texi (Debugging Options): Ditto.
2684 * opts-global.c (handle_common_deferred_options): Ditto.
2685 (dump_remap_tree_vectorizer_verbose): Delete.
2686 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
2688 2013-10-30 DJ Delorie <dj@redhat.com>
2690 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
2691 that take no arguments.
2693 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
2696 * reload1.c (update_eliminables_and_spill): New function, broken
2698 (reload): Use it. Check for frame size change after frame size
2699 alignment, and call update_eliminables_and_spill first if continue-ing.
2701 2013-10-30 Cong Hou <congh@google.com>
2704 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
2705 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
2706 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
2708 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
2710 PR rtl-optimization/58369
2711 * reload1.c (compute_reload_subreg_offset): New function.
2712 (choose_reload_regs): Use it to pass endian-correct
2713 offset to subreg_regno_offset.
2715 2013-10-30 Tobias Burnus <burnus@net-b.de>
2718 * tree-cfg.c (replace_loop_annotate): Replace warning by
2721 2013-10-30 Jason Merrill <jason@redhat.com>
2723 * configure.ac (loose_warn): Add -Wno-format if
2724 --disable-build-format-warnings.
2726 2013-10-30 David Malcolm <dmalcolm@redhat.com>
2728 * cgraphunit.c (analyze_functions): Split symtab_node declarations
2729 onto multiple lines to make things easier for rename_symtab.py.
2731 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
2732 (symtab_semantically_equivalent_p): Likewise.
2734 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
2737 * lra.c (check_rtl): Remove address check before LRA work.
2739 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
2741 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
2742 POINTER_PLUS_EXPR in the defining statement.
2744 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
2746 * regmove.c: Remove.
2747 * tree-pass.h (make_pass_regmove): Remove.
2748 * timevar.def (TV_REGMOVE): Remove.
2749 * passes.def (pass_regmove): Remove.
2750 * opts.c (default_options_table): Remove entry for regmove.
2751 * doc/passes.texi: Remove regmove pass description.
2752 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
2754 (-fdump-rtl-regmove): Ditto.
2755 * common.opt (foptimize-register-move, fregmove): Ignore.
2756 * Makefile.in (OBJS): Remove regmove.o.
2757 * regmove.c: Remove.
2758 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
2760 (struct ira_allocno) New member allocno_prefs.
2761 (ALLOCNO_PREFS): New macro.
2762 (ira_prefs, ira_prefs_num): New external vars.
2763 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
2764 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
2765 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
2767 (ira_add_allocno_copy_to_list): Remove prototype.
2768 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
2769 (ira_pref_iterator): New type.
2770 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
2771 (FOR_EACH_PREF): New macro.
2772 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
2773 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
2775 (ira_setup_alts): New function.
2776 (decrease_live_ranges_number): New function.
2777 (ira): Call the above function.
2778 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
2779 (ira_create_allocno): Initialize allocno prefs.
2780 (pref_pool, pref_vec): New static vars.
2781 (initiate_prefs, find_allocno_pref, ira_create_pref): New
2783 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
2784 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
2785 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
2786 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
2787 (ira_add_allocno_copy_to_list): Make static. Rename to
2788 add_allocno_copy_to_list.
2789 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
2790 swap_allocno_copy_ends_if_necessary.
2791 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
2792 ira_remove_allocno_prefs.
2793 (ira_flattening): Ditto.
2794 (ira_build): Call initiate_prefs, print_prefs.
2795 (ira_destroy): Call finish_prefs.
2796 * ira-color.c (struct update_cost_record): New.
2797 (struct allocno_color_data): Add new member update_cost_records.
2798 (update_cost_record_pool): New static var.
2799 (init_update_cost_records, get_update_cost_record): New functions.
2800 (free_update_cost_record_list, finish_update_cost_records): Ditto.
2801 (struct update_cost_queue_elem): Add member from.
2802 (initiate_cost_update): Call init_update_cost_records.
2803 (finish_cost_update): Call finish_update_cost_records.
2804 (queue_update_cost, get_next_update_cost): Add new param from.
2805 (Update_allocno_cost, update_costs_from_allocno): New functions.
2806 (update_costs_from_prefs): Ditto.
2807 (update_copy_costs): Rename to update_costs_from_copies.
2808 (restore_costs_from_copies): New function.
2809 (update_conflict_hard_regno_costs): Don't go back.
2810 (assign_hard_reg): Call restore_costs_from_copies. Add printing
2812 (pop_allocnos): Add priniting more debug info.
2813 (color_allocnos): Remove prefs for conflicting hard regs.
2814 Call update_costs_from_prefs.
2815 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
2816 (get_dup_num): Rename, modify, and move to ira.c
2817 (process_regs_for_copy): Add prefs.
2818 (add_insn_allocno_copies): Put src as first arg of
2819 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
2820 * ira-costs.c (record_reg_classes): Modify and move code into
2821 record_operands_costs.
2822 (find_costs_and_classes): Create prefs for the hard reg of small
2824 (process_bb_node_for_hard_reg_moves): Add prefs.
2826 2013-10-30 Richard Biener <rguenther@suse.de>
2829 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
2830 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
2832 (pre_and_rev_post_order_compute): ... which now wraps it.
2833 * graph.c (draw_cfg_nodes_no_loops): Use
2834 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
2837 2013-10-30 Christian Bruel <christian.bruel@st.com>
2839 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
2840 (sh_expand_cmpstr): Handle known align and schedule improvements.
2841 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
2842 * config/sh/sh.md (cmpstrnsi): New pattern.
2844 2013-10-30 Martin Jambor <mjambor@suse.cz>
2846 PR rtl-optimization/10474
2847 * ira.c (find_moveable_pseudos): Do not calculate dominance info
2849 (interesting_dest_for_shprep): New function.
2850 (split_live_ranges_for_shrink_wrap): Likewise.
2851 (ira): Calculate dominance info and df analysis. Call
2852 split_live_ranges_for_shrink_wrap.
2854 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2857 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
2859 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
2861 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
2862 * tree.h (POINTER_BOUNDS_P): New.
2863 (BOUNDED_TYPE_P): New.
2865 (pointer_bounds_type_node): New.
2866 * tree.c (build_common_tree_nodes): Initialize
2867 pointer_bounds_type_node.
2868 * gimple.h (gimple_call_get_nobnd_arg_index): New.
2869 (gimple_call_num_nobnd_args): New.
2870 (gimple_call_nobnd_arg): New.
2871 (gimple_return_retbnd): New.
2872 (gimple_return_set_retbnd): New
2873 * gimple.c (gimple_build_return): Increase number of ops
2874 for return statement.
2875 (gimple_call_get_nobnd_arg_index): New.
2876 * gimple-pretty-print.c (dump_gimple_return): Print second op.
2878 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
2880 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
2881 with "chkp ctor" and "bnd_legacy" attributes.
2882 * gimplify.c (gimplify_init_constructor): Avoid infinite
2883 loop during gimplification of bounds initializer.
2885 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
2887 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
2888 (handle_bnd_legacy): New.
2889 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
2890 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
2893 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
2895 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
2896 (BT_FN_PTR_CONST_PTR): New.
2897 (BT_FN_CONST_PTR_CONST_PTR): New.
2898 (BT_FN_PTR_CONST_PTR_SIZE): New.
2899 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
2900 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
2901 (BT_FN_VOID_CONST_PTR_SIZE): New.
2902 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
2903 * chkp-builtins.def: New.
2904 * builtins.def: include chkp-builtins.def.
2905 (DEF_CHKP_BUILTIN): New.
2906 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
2907 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
2908 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
2909 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
2910 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
2911 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
2912 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
2913 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
2914 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
2915 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
2916 * common.opt (fcheck-pointer-bounds): New.
2917 * toplev.c (process_options): Check Pointer Bounds Checker is
2919 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
2921 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
2923 * target.def (builtin_chkp_function): New.
2924 (chkp_bound_type): New.
2925 (chkp_bound_mode): New.
2926 (fn_abi_va_list_bounds_size): New.
2927 (load_bounds_for_arg): New.
2928 (store_bounds_for_arg): New.
2929 * targhooks.h (default_load_bounds_for_arg): New.
2930 (default_store_bounds_for_arg): New.
2931 (default_fn_abi_va_list_bounds_size): New.
2932 (default_chkp_bound_type): New.
2933 (default_chkp_bound_mode): New.
2934 (default_builtin_chkp_function): New.
2935 * targhooks.c (default_load_bounds_for_arg): New.
2936 (default_store_bounds_for_arg): New.
2937 (default_fn_abi_va_list_bounds_size): New.
2938 (default_chkp_bound_type): New.
2939 (default_chkp_bound_mode); New.
2940 (default_builtin_chkp_function): New.
2941 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
2942 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
2943 (TARGET_STORE_BOUNDS_FOR_ARG): New.
2944 (TARGET_BUILTIN_CHKP_FUNCTION): New.
2945 (TARGET_CHKP_BOUND_TYPE): New.
2946 (TARGET_CHKP_BOUND_MODE): New.
2947 * doc/tm.texi: Regenerated.
2948 * langhooks.h (lang_hooks): Add chkp_supported field.
2949 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
2950 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
2952 2013-10-29 Andrew Pinski <apinski@cavium.com>
2954 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
2955 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
2956 fails, combining the branches anyways.
2957 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
2958 increases the number of combinings.
2959 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
2961 2013-10-29 Mike Stump <mikestump@comcast.net>
2963 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
2964 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
2965 (make_vector_mode): Increase namebuf to 16.
2966 (emit_insn_modes_h): When processing BImode, don't
2967 also match partial int modes.
2968 (emit_class_narrowest_mode): Likewise.
2970 * config/bfin/bfin-modes.def: Add precision to PDI.
2971 * config/m32c/m32c-modes.def: Add precision to PSI.
2972 * config/msp430/msp430-modes.def: Add precision to PSI.
2973 * config/rs6000/rs6000-modes.def: Add precision to PTI.
2974 * config/sh/sh-modes.def: Add precision to PSI and PDI.
2976 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
2979 * config/sh/sh.md (*addc): Rename existing variations to ...
2980 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
2981 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
2982 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
2983 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
2985 2013-10-29 Teresa Johnson <tejohnson@google.com>
2988 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
2989 insanities when updating probabilities.
2991 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2993 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
2994 prettyprinter for cgraph_node to reflect the conversion of the
2995 symtable types to a C++ class hierarchy: it now *is* a
2996 symtab_node_base, rather than having one (named "symbol").
2998 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
3000 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
3001 __cilkrts_pop_frame. If matched, then return true for built-in
3003 (expand_builtin): Added BUILT_IN_CILK_DETACH and
3004 BUILT_IN_CILK_POP_FRAME case.
3005 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
3006 (lhs_cilk_detect_spawn): Likewise.
3007 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
3008 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
3009 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
3010 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
3011 (LANG_HOOKS_CILKPLUS): Likewise.
3012 * tree.h (CILK_SPAWN_FN): Likewise.
3013 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
3014 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
3015 (OBJS): Added cilk-common.o.
3016 (BUILTINS_DEF): Added cilk-builtins.def.
3017 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
3018 (lhd_cilk_detect_spawn): Likewise.
3019 * langhooks.h (lang_hooks_for_cilkplus): New struct.
3020 (struct lang_hooks): Added new field called "cilkplus."
3021 * cilk-common.c: New file.
3023 * cilk-builtins.def: Likewise.
3024 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
3025 "__cilk" macro and set it to 200.
3026 * function.h (struct function::cilk_frame_decl): New field.
3027 (struct function::is_cilk_function): Likewise.
3028 (struct function::calls_cilk_spawn): Likewise.
3029 * gimplify.c (gimplify_call_expr): Added a check if the function call
3030 being gimplified is a spawn detach point. If so, then add pop_frame
3031 and detach function calls.
3032 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
3033 for gimplifying _Cilk_spawn and _Cilk_sync statements.
3034 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
3035 function. If so, added a _Cilk_sync and gimplified it.
3036 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
3037 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
3038 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
3040 (can_inline_edge_p): Prevent inling of spawnee function.
3041 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
3042 for functions that use Cilk keywords.
3043 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
3044 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
3045 CILK_SYNC_STMT cases.
3046 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
3048 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
3049 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
3050 * passes.texi (Cilk Keywords): New section that describes the compiler
3051 code changes for handling Cilk Keywords.
3053 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3055 Patch autogenerated by refactor_symtab.py from
3056 https://github.com/davidmalcolm/gcc-refactoring-scripts
3057 revision 58bb219cc090b2f4516a9297d868c245495ee622
3059 * asan.c (asan_finish_file): Update for conversion of symtab types to
3060 a true class hierarchy.
3061 * cfgexpand.c (estimated_stack_frame_size): Likewise.
3062 * cgraph.c (cgraph_get_body): Likewise.
3063 (cgraph_get_create_real_symbol_node): Likewise.
3064 (verify_cgraph_node): Likewise.
3065 (verify_edge_corresponds_to_fndecl): Likewise.
3066 (verify_edge_count_and_frequency): Likewise.
3067 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
3068 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3069 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
3070 (cgraph_node_cannot_return): Likewise.
3071 (cgraph_set_pure_flag_1): Likewise.
3072 (cgraph_set_const_flag_1): Likewise.
3073 (cgraph_set_nothrow_flag_1): Likewise.
3074 (cgraph_make_node_local_1): Likewise.
3075 (cgraph_for_node_and_aliases): Likewise.
3076 (cgraph_for_node_thunks_and_aliases): Likewise.
3077 (cgraph_node_can_be_local_p): Likewise.
3078 (cgraph_node_cannot_be_local_p_1): Likewise.
3079 (cgraph_function_body_availability): Likewise.
3080 (dump_cgraph_node): Likewise.
3081 (cgraph_rtl_info): Likewise.
3082 (cgraph_mark_address_taken_node): Likewise.
3083 (cgraph_remove_node): Likewise.
3084 (cgraph_release_function_body): Likewise.
3085 (cgraph_update_edges_for_call_stmt_node): Likewise.
3086 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
3087 (cgraph_make_edge_direct): Likewise.
3088 (cgraph_resolve_speculation): Likewise.
3089 (cgraph_speculative_call_info): Likewise.
3090 (cgraph_turn_edge_to_speculative): Likewise.
3091 (cgraph_create_edge_1): Likewise.
3092 (cgraph_set_call_stmt): Likewise.
3093 (cgraph_node_for_asm): Likewise.
3094 (cgraph_add_thunk): Likewise.
3095 (cgraph_same_body_alias): Likewise.
3096 (cgraph_create_function_alias): Likewise.
3097 (cgraph_create_node): Likewise.
3098 (cgraph_create_empty_node): Likewise.
3099 (record_function_versions): Likewise.
3100 (used_from_object_file_p): Likewise.
3101 * cgraph.h (symtab_can_be_discarded): Likewise.
3102 (symtab_real_symbol_p): Likewise.
3103 (cgraph_mark_force_output_node): Likewise.
3104 (cgraph_edge_recursive_p): Likewise.
3105 (symtab_alias_target): Likewise.
3106 (varpool_all_refs_explicit_p): Likewise.
3107 (varpool_can_remove_if_no_refs): Likewise.
3108 (cgraph_only_called_directly_or_aliased_p): Likewise.
3109 (cgraph_next_function_with_gimple_body): Likewise.
3110 (cgraph_first_function_with_gimple_body): Likewise.
3111 (cgraph_function_with_gimple_body_p): Likewise.
3112 (cgraph_next_function): Likewise.
3113 (cgraph_first_function): Likewise.
3114 (cgraph_next_defined_function): Likewise.
3115 (cgraph_first_defined_function): Likewise.
3116 (varpool_next_defined_variable): Likewise.
3117 (varpool_first_defined_variable): Likewise.
3118 (varpool_next_static_initializer): Likewise.
3119 (varpool_first_static_initializer): Likewise.
3120 (varpool_next_variable): Likewise.
3121 (varpool_first_variable): Likewise.
3122 (varpool_node_name): Likewise.
3123 (varpool): Likewise.
3125 (is_a_helper <varpool_node>::test): Likewise.
3126 (is_a_helper <cgraph_node>::test): Likewise.
3127 (varpool_variable_node): Likewise.
3128 (cgraph_function_or_thunk_node): Likewise.
3129 (varpool_alias_target): Likewise.
3130 (cgraph_alias_target): Likewise.
3131 (cgraph_node_name): Likewise.
3132 (varpool_node_asm_name): Likewise.
3133 (cgraph_node_asm_name): Likewise.
3134 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
3135 (cgraph_rebuild_references): Likewise.
3136 (rebuild_cgraph_edges): Likewise.
3137 (record_eh_tables): Likewise.
3138 (build_cgraph_edges): Likewise.
3139 (mark_store): Likewise.
3140 (mark_load): Likewise.
3141 (mark_address): Likewise.
3142 (record_type_list): Likewise.
3143 (record_reference): Likewise.
3144 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
3145 (cgraph_materialize_clone): Likewise.
3146 (cgraph_function_versioning): Likewise.
3147 (cgraph_copy_node_for_versioning): Likewise.
3148 (update_call_expr): Likewise.
3149 (cgraph_find_replacement_node): Likewise.
3150 (cgraph_create_virtual_clone): Likewise.
3151 (cgraph_clone_node): Likewise.
3152 * cgraphunit.c (compile): Likewise.
3153 (output_weakrefs): Likewise.
3154 (output_in_order): Likewise.
3155 (expand_function): Likewise.
3156 (assemble_thunks_and_aliases): Likewise.
3157 (expand_thunk): Likewise.
3158 (mark_functions_to_output): Likewise.
3159 (handle_alias_pairs): Likewise.
3160 (analyze_functions): Likewise.
3161 (walk_polymorphic_call_targets): Likewise.
3162 (varpool_finalize_decl): Likewise.
3163 (process_function_and_variable_attributes): Likewise.
3164 (cgraph_process_same_body_aliases): Likewise.
3165 (analyze_function): Likewise.
3166 (cgraph_add_new_function): Likewise.
3167 (cgraph_finalize_function): Likewise.
3168 (referred_to_p): Likewise.
3169 (cgraph_reset_node): Likewise.
3170 (cgraph_process_new_functions): Likewise.
3171 (enqueue_node): Likewise.
3172 (decide_is_symbol_needed): Likewise.
3173 * coverage.c (coverage_compute_profile_id): Likewise.
3174 * dbxout.c (dbxout_expand_expr): Likewise.
3175 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
3176 (reference_to_unused): Likewise.
3177 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3178 * gimplify.c (unvisit_body): Likewise.
3179 (unshare_body): Likewise.
3180 * ipa-cp.c (ipcp_generate_summary): Likewise.
3181 (ipcp_decision_stage): Likewise.
3182 (identify_dead_nodes): Likewise.
3183 (decide_whether_version_node): Likewise.
3184 (decide_about_value): Likewise.
3185 (perhaps_add_new_callers): Likewise.
3186 (create_specialized_node): Likewise.
3187 (update_profiling_info): Likewise.
3188 (ipcp_propagate_stage): Likewise.
3189 (estimate_local_effects): Likewise.
3190 (good_cloning_opportunity_p): Likewise.
3191 (devirtualization_time_bonus): Likewise.
3192 (propagate_constants_accross_call): Likewise.
3193 (initialize_node_lattices): Likewise.
3194 (ipcp_cloning_candidate_p): Likewise.
3195 (determine_versionability): Likewise.
3196 (print_all_lattices): Likewise.
3197 (print_lattice): Likewise.
3198 (ipcp_discover_new_direct_edges): Likewise.
3199 * ipa-devirt.c (ipa_devirt): Likewise.
3200 (likely_target_p): Likewise.
3201 (update_type_inheritance_graph): Likewise.
3202 (possible_polymorphic_call_target_p): Likewise.
3203 (dump_possible_polymorphic_call_targets): Likewise.
3204 (devirt_variable_node_removal_hook): Likewise.
3205 (record_binfo): Likewise.
3206 (maybe_record_node): Likewise.
3207 (build_type_inheritance_graph): Likewise.
3208 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3209 (inline_generate_summary): Likewise.
3210 (inline_analyze_function): Likewise.
3211 (do_estimate_growth): Likewise.
3212 (simple_edge_hints): Likewise.
3213 (estimate_node_size_and_time): Likewise.
3214 (estimate_edge_devirt_benefit): Likewise.
3215 (compute_inline_parameters): Likewise.
3216 (estimate_function_body_sizes): Likewise.
3217 (compute_bb_predicates): Likewise.
3218 (initialize_inline_failed): Likewise.
3219 (dump_inline_summary): Likewise.
3220 (dump_inline_edge_summary): Likewise.
3221 * ipa-inline-transform.c (inline_transform): Likewise.
3222 (preserve_function_body_p): Likewise.
3223 (save_inline_function_body): Likewise.
3224 (inline_call): Likewise.
3225 (clone_inlined_nodes): Likewise.
3226 (can_remove_node_now_p): Likewise.
3227 (can_remove_node_now_p_1): Likewise.
3228 * ipa-inline.c (early_inliner): Likewise.
3229 (early_inline_small_functions): Likewise.
3230 (inline_always_inline_functions): Likewise.
3231 (ipa_inline): Likewise.
3232 (flatten_function): Likewise.
3233 (inline_small_functions): Likewise.
3234 (speculation_useful_p): Likewise.
3235 (recursive_inlining): Likewise.
3236 (update_caller_keys): Likewise.
3237 (reset_edge_caches): Likewise.
3238 (update_edge_key): Likewise.
3239 (edge_badness): Likewise.
3240 (relative_time_benefit): Likewise.
3241 (want_inline_self_recursive_call_p): Likewise.
3242 (want_inline_small_function_p): Likewise.
3243 (want_early_inline_function_p): Likewise.
3244 (num_calls): Likewise.
3245 (can_early_inline_edge_p): Likewise.
3246 (can_inline_edge_p): Likewise.
3247 (report_inline_failed_reason): Likewise.
3248 * ipa-profile.c (ipa_profile): Likewise.
3249 (ipa_propagate_frequency): Likewise.
3250 (ipa_propagate_frequency_1): Likewise.
3251 (ipa_profile_generate_summary): Likewise.
3252 * ipa-prop.c (ipcp_transform_function): Likewise.
3253 (read_replacements_section): Likewise.
3254 (ipa_prop_read_section): Likewise.
3255 (ipa_modify_call_arguments): Likewise.
3256 (ipa_print_node_params): Likewise.
3257 (propagate_controlled_uses): Likewise.
3258 (update_indirect_edges_after_inlining): Likewise.
3259 (remove_described_reference): Likewise.
3260 (ipa_make_edge_direct_to_target): Likewise.
3261 (ipa_analyze_node): Likewise.
3262 (ipa_analyze_params_uses): Likewise.
3263 (ipa_compute_jump_functions): Likewise.
3264 (ipa_get_callee_param_type): Likewise.
3265 (ipa_print_node_jump_functions): Likewise.
3266 (ipa_initialize_node_params): Likewise.
3267 (ipa_populate_param_decls): Likewise.
3268 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
3269 (write_agg_replacement_chain): Likewise.
3270 (ipa_write_node_info): Likewise.
3271 (ipa_edge_duplication_hook): Likewise.
3272 (try_decrement_rdesc_refcount): Likewise.
3273 * ipa-pure-const.c (propagate_nothrow): Likewise.
3274 (propagate_pure_const): Likewise.
3275 (pure_const_read_summary): Likewise.
3276 (pure_const_write_summary): Likewise.
3277 (analyze_function): Likewise.
3278 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
3279 (ipa_ref_referring_ref_list): Likewise.
3280 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
3281 (ipa_remove_stmt_references): Likewise.
3282 (ipa_find_reference): Likewise.
3283 (ipa_dump_referring): Likewise.
3284 (ipa_dump_references): Likewise.
3285 (ipa_record_reference): Likewise.
3286 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
3287 (ipa_reference_write_optimization_summary): Likewise.
3288 (write_node_summary_p): Likewise.
3289 (propagate): Likewise.
3290 (read_write_all_from_decl): Likewise.
3291 (generate_summary): Likewise.
3292 (analyze_function): Likewise.
3293 (propagate_bits): Likewise.
3294 (ipa_reference_get_not_written_global): Likewise.
3295 (ipa_reference_get_not_read_global): Likewise.
3296 * ipa-split.c (execute_split_functions): Likewise.
3297 (split_function): Likewise.
3298 * ipa-utils.c (ipa_merge_profiles): Likewise.
3299 (dump_cgraph_node_set): Likewise.
3300 (ipa_reverse_postorder): Likewise.
3301 (ipa_edge_within_scc): Likewise.
3302 (ipa_get_nodes_in_cycle): Likewise.
3303 (ipa_free_postorder_info): Likewise.
3304 (ipa_reduced_postorder): Likewise.
3305 (searchc): Likewise.
3306 (recursive_call_p): Likewise.
3307 * ipa.c (ipa_cdtor_merge): Likewise.
3308 (record_cdtor_fn): Likewise.
3309 (function_and_variable_visibility): Likewise.
3310 (varpool_externally_visible_p): Likewise.
3311 (cgraph_externally_visible_p): Likewise.
3312 (comdat_can_be_unshared_p): Likewise.
3313 (comdat_can_be_unshared_p_1): Likewise.
3314 (address_taken_from_non_vtable_p): Likewise.
3315 (ipa_discover_readonly_nonaddressable_vars): Likewise.
3316 (symtab_remove_unreachable_nodes): Likewise.
3317 (walk_polymorphic_call_targets): Likewise.
3318 (process_references): Likewise.
3319 (enqueue_node): Likewise.
3320 (has_addr_references_p): Likewise.
3321 (cgraph_non_local_node_p_1): Likewise.
3322 * is-a.h (varpool_analyze_node): Likewise.
3323 * lto-cgraph.c (input_symtab): Likewise.
3324 (merge_profile_summaries): Likewise.
3325 (input_cgraph_1): Likewise.
3326 (input_edge): Likewise.
3327 (input_varpool_node): Likewise.
3328 (input_node): Likewise.
3329 (input_overwrite_node): Likewise.
3330 (compute_ltrans_boundary): Likewise.
3331 (output_refs): Likewise.
3332 (lto_output_varpool_node): Likewise.
3333 (lto_output_node): Likewise.
3334 (reachable_from_other_partition_p): Likewise.
3335 (referenced_from_other_partition_p): Likewise.
3336 (lto_output_edge): Likewise.
3337 (output_node_opt_summary): Likewise.
3338 (add_node_to): Likewise.
3339 (reachable_from_this_partition_p): Likewise.
3340 (lto_set_symtab_encoder_in_partition): Likewise.
3341 (lto_symtab_encoder_in_partition_p): Likewise.
3342 (lto_set_symtab_encoder_encode_initializer): Likewise.
3343 (lto_symtab_encoder_encode_initializer_p): Likewise.
3344 (lto_set_symtab_encoder_encode_body): Likewise.
3345 (lto_symtab_encoder_encode_body_p): Likewise.
3346 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
3348 * lto-streamer-in.c (lto_read_body): Likewise.
3349 (fixup_call_stmt_edges): Likewise.
3350 (fixup_call_stmt_edges_1): Likewise.
3351 * lto-streamer-out.c (produce_symtab): Likewise.
3352 (output_symbol_p): Likewise.
3353 (write_symbol): Likewise.
3354 (lto_output): Likewise.
3355 (copy_function): Likewise.
3356 (output_function): Likewise.
3357 * passes.c (function_called_by_processed_nodes_p): Likewise.
3358 (ipa_write_optimization_summaries): Likewise.
3359 (ipa_write_summaries): Likewise.
3360 (do_per_function_toporder): Likewise.
3361 (do_per_function): Likewise.
3362 (dump_passes): Likewise.
3363 * symtab.c (symtab_semantically_equivalent_p): Likewise.
3364 (symtab_nonoverwritable_alias): Likewise.
3365 (symtab_nonoverwritable_alias_1): Likewise.
3366 (symtab_for_node_and_aliases): Likewise.
3367 (symtab_resolve_alias): Likewise.
3368 (fixup_same_cpp_alias_visibility): Likewise.
3369 (symtab_alias_ultimate_target): Likewise.
3370 (symtab_used_from_object_file_p): Likewise.
3371 (verify_symtab_base): Likewise.
3372 (dump_symtab_base): Likewise.
3373 (symtab_node_name): Likewise.
3374 (symtab_node_asm_name): Likewise.
3375 (symtab_dissolve_same_comdat_group_list): Likewise.
3376 (symtab_add_to_same_comdat_group): Likewise.
3377 (symtab_unregister_node): Likewise.
3378 (symtab_insert_node_to_hashtable): Likewise.
3379 (symtab_register_node): Likewise.
3380 (unlink_from_assembler_name_hash): Likewise.
3381 (insert_to_assembler_name_hash): Likewise.
3382 (eq_assembler_name): Likewise.
3383 (hash_node_by_assembler_name): Likewise.
3384 (eq_node): Likewise.
3385 (hash_node): Likewise.
3386 * toplev.c (wrapup_global_declaration_2): Likewise.
3387 * trans-mem.c (ipa_tm_execute): Likewise.
3388 (ipa_tm_transform_clone): Likewise.
3389 (ipa_tm_transform_transaction): Likewise.
3390 (ipa_tm_transform_calls_redirect): Likewise.
3391 (ipa_tm_insert_gettmclone_call): Likewise.
3392 (ipa_tm_insert_irr_call): Likewise.
3393 (ipa_tm_create_version): Likewise.
3394 (ipa_tm_create_version_alias): Likewise.
3395 (ipa_tm_mark_forced_by_abi_node): Likewise.
3396 (ipa_tm_mark_force_output_node): Likewise.
3397 (ipa_tm_diagnose_tm_safe): Likewise.
3398 (ipa_tm_mayenterirr_function): Likewise.
3399 (ipa_tm_scan_irr_function): Likewise.
3400 (ipa_tm_note_irrevocable): Likewise.
3401 (ipa_tm_scan_calls_clone): Likewise.
3402 (get_cg_data): Likewise.
3403 * tree-eh.c (tree_could_trap_p): Likewise.
3404 * tree-emutls.c (ipa_lower_emutls): Likewise.
3405 (create_emultls_var): Likewise.
3406 (lower_emutls_function_body): Likewise.
3407 (gen_emutls_addr): Likewise.
3408 (emutls_decl): Likewise.
3409 (new_emutls_decl): Likewise.
3410 * tree-inline.c (tree_function_versioning): Likewise.
3411 (optimize_inline_calls): Likewise.
3412 (expand_call_inline): Likewise.
3413 (estimate_num_insns): Likewise.
3414 (copy_bb): Likewise.
3415 (delete_unreachable_blocks_update_callgraph): Likewise.
3416 * tree-nested.c (gimplify_all_functions): Likewise.
3417 (create_nesting_tree): Likewise.
3418 (check_for_nested_with_variably_modified): Likewise.
3419 * tree-pretty-print.c (dump_function_header): Likewise.
3420 * tree-profile.c (tree_profiling): Likewise.
3421 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
3422 (modify_function): Likewise.
3423 (convert_callers): Likewise.
3424 (convert_callers_for_node): Likewise.
3425 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
3426 (associate_varinfo_to_alias): Likewise.
3427 (create_variable_info_for): Likewise.
3428 (get_constraint_for_ssa_var): Likewise.
3429 * tree-vectorizer.c (increase_alignment): Likewise.
3430 * tree.c (find_decls_types_in_var): Likewise.
3431 (find_decls_types_in_node): Likewise.
3432 (free_lang_data_in_decl): Likewise.
3433 * value-prof.c (gimple_ic_transform): Likewise.
3434 (gimple_ic): Likewise.
3435 (check_ic_target): Likewise.
3436 (init_node_map): Likewise.
3437 * varasm.c (decl_binds_to_current_def_p): Likewise.
3438 (default_binds_local_p_1): Likewise.
3439 (dump_tm_clone_pairs): Likewise.
3440 (assemble_alias): Likewise.
3441 (find_decl): Likewise.
3442 (mark_decl_referenced): Likewise.
3443 * varpool.c (varpool_for_node_and_aliases): Likewise.
3444 (varpool_extra_name_alias): Likewise.
3445 (varpool_create_variable_alias): Likewise.
3446 (add_new_static_var): Likewise.
3447 (varpool_finalize_named_section_flags): Likewise.
3448 (varpool_remove_unreferenced_decls): Likewise.
3449 (enqueue_node): Likewise.
3450 (varpool_assemble_decl): Likewise.
3451 (assemble_aliases): Likewise.
3452 (varpool_analyze_node): Likewise.
3453 (cgraph_variable_initializer_availability): Likewise.
3454 (varpool_add_new_variable): Likewise.
3455 (ctor_for_folding): Likewise.
3456 (dump_varpool_node): Likewise.
3457 (varpool_remove_initializer): Likewise.
3458 (varpool_remove_node): Likewise.
3459 (varpool_node_for_decl): Likewise.
3460 (varpool_create_empty_node): Likewise.
3461 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
3462 (ix86_get_function_versions_dispatcher): Likewise.
3464 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3466 * cgraph.h (symtab_node_base): Convert to a class;
3467 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
3468 chain_next/prev from symtab_node_def.
3469 (cgraph_node): Inherit from symtab_node; add GTY option
3470 tag ("SYMTAB_FUNCTION").
3471 (varpool_node): Inherit from symtab_node; add GTY option
3472 tag ("SYMTAB_VARIABLE").
3473 (symtab_node_def): Remove.
3474 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
3475 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
3476 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
3477 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
3479 * ipa-ref.h (symtab_node_def): Drop.
3480 (symtab_node): Change underlying type from symtab_node_def to
3482 (const_symtab_node): Likwise.
3484 * is-a.h: Update examples in comment.
3486 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
3487 (assembler_name_hash): Likewise.
3489 2013-10-29 Martin Liska <marxin.liska@gmail.com>
3491 * doc/tree-ssa.texi (gimple_phi_result): Document.
3492 (gimple_phi_num_args, gimple_phi_arg): Likewise.
3493 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
3494 (PHI_RESULT, PHI_NUM_ARGS): Remove.
3495 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
3497 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
3499 * expr.h: Revert change and include tree-core.h.
3500 * rtl.h: Revert change and don't include tree-core.h.
3502 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
3504 * config/darwin.c: Include gimple.h.
3505 * config/i386/winnt.c: Likewise.
3507 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
3509 PR tree-optimization/19831
3510 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
3512 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
3514 * tree-outof-ssa.h: Remove include files.
3515 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
3517 * tree-ssa-coalesce.c: Likewise.
3518 * cfgexpand.c: Likewise.
3519 * tree-ssa-ter.c: Likewise.
3520 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
3521 * lto-streamer.h: Likewise.
3522 * cgraphbuild.c: Add gimple.h to include list.
3523 * data-streamer-in.c: Likewise.
3524 * ipa-cp.c: Likewise.
3525 * tree-streamer.c: Likewise.
3526 * lto-compress.c: Likewise.
3527 * ipa-reference.c: Likewise.
3528 * data-streamer-out.c: Likewise.
3529 * lto-cgraph.c: Likewise.
3530 * cgraphclones.c: Likewise.
3531 * ipa-utils.c: Likewise.
3532 * data-streamer.c: Likewise.
3533 * ipa-split.c: Likewise.
3534 * lto-section-in.c: Likewise.
3535 * tree-streamer-out.c: Likewise.
3536 * ipa-prop.c: Likewise.
3537 * tree-streamer-in.c: Likewise.
3538 * symtab.c: Likewise.
3539 * opts-global.c: Likewise.
3540 * lto-opts.c: Likewise.
3541 * lto-section-out.c: Likewise.
3542 * lto-streamer.c: Likewise.
3543 * rtl.h: Add tree-core.h to include list.
3544 * expr.h: Remove tree-core.h from include list.
3545 * gimple.h: Likewise.
3546 * ipa-utils.h: Likewise.
3547 * streamer-hooks.h: Likewise.
3548 * streamer-hooks.c: Include input.h.
3550 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3552 * config/arm/arm.c (cortexa7_extra_costs): New table.
3553 (arm_cortex_a7_tune): New.
3554 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
3556 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
3558 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
3560 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3562 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
3563 to use autogenerated markers for a class-hierarchy, every class
3564 must have a GTY marker.
3565 * gengtype.h (struct type): Add linked list of subclasses to
3566 the "s" member of the union.
3567 (add_subclass): New decl.
3568 * gengtype-state.c (read_state_struct_type): Set up subclass
3570 * gengtype.c (get_ultimate_base_class): New.
3571 (add_subclass): New.
3572 (new_structure): Set up subclass linked list.
3573 (set_gc_used_type): Propagate usage information to subclasses.
3574 (output_mangled_typename): Use get_ultimate_base_class.
3575 (walk_subclasses): Use the subclass linked list, avoiding an
3576 O(N^2) when writing out all types.
3577 (walk_type): Issue an error if the base class is missing a tag,
3578 rather than generating bogus C code. Add a gcc_unreachable
3579 default case, in case people omit tags from concrete subclasses,
3580 or get the values wrong.
3581 (write_func_for_structure): Issue an error for subclasses for
3582 which the base doesn't have a "desc", since otherwise the
3583 autogenerated routines for the base would silently fail to visit
3584 any subclass fields.
3585 (write_root): Use get_ultimate_base_class, tweaking constness of
3586 tp to match that function's signature.
3588 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3590 * doc/gty.texi (GTY Options): Add note about inheritance to
3591 description of desc and tag.
3592 (Inheritance and GTY): New.
3594 2013-10-29 David Malcolm <dmalcolm@redhat.com>
3596 * gengtype-parse.c (opts_have): Drop "static" so that
3597 we can use this from gengtype.c.
3598 * gengtype.c (set_gc_used_type): Mark any base class as used;
3599 update field traversal to visit inherited fields.
3600 (output_mangled_typename): Convert references to classes within
3601 an inheritance hierarchy to reference the ultimate base class,
3602 since only it will have gt_ functions.
3603 (get_string_option): New.
3604 (walk_subclasses): New.
3605 (walk_type): Treat GTY structs that have a "desc" as being the
3606 root of an inheritance hierarchy. Generate a switch on it
3607 within the marking function which walks all subclasses, adding
3608 cases for them via walk_subclasses. For subclasses, visit all
3609 fields of the type (including inherited ones).
3610 (write_func_for_structure): Don't write fns for subclasses, only
3611 for the ultimate base class within an inheritance hierarchy.
3612 Subclasses-marking will be handled by the base class marking functions.
3613 (write_types): Likewise.
3614 (write_local_func_for_structure): Likewise.
3615 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
3616 a "tag" option (and are thus concrete subclasses).
3617 (write_root): Use the marker function for the ultimate base class.
3618 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
3619 (opts_have): Add declaration.
3621 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
3623 * lra-spills.c (lra_final_code_change): Remove useless move insns
3624 originated from moves of pseudos.
3626 2013-10-28 Jeff Law <law@redhat.com>
3628 * lower-subreg.c (resolve_simple_move): Fix comment typo.
3630 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
3632 * df-scan.c (df_collection_rec): Adjust.
3633 (copy_defs): New constant.
3634 (copy_uses): Likewise.
3635 (copy_eq_uses): Likewise.
3636 (copy_mw): Likewise.
3637 (copy_all): Likewise.
3638 (df_insn_rescan): Adjust.
3639 (df_notes_rescan): Likewise.
3640 (df_swap_refs): Likewise.
3641 (df_sort_and_compress_refs): Likewise.
3642 (df_sort_and_compress_mws): Likewise.
3643 (df_install_refs): Likewise.
3644 (df_install_mws): Likewise.
3645 (df_refs_add_to_chains): Add flags parameter controlling which vectors
3647 (df_bb_refs_record): Adjust.
3648 (df_record_entry_block_defs): Likewise.
3649 (df_record_exit_block_defs): Likewise.
3650 (df_refs_verify): Likewise.
3651 (df_mws_verify): Likewise.
3652 (df_insn_refs_verify): Likewise.
3653 (df_bb_verify): Likewise.
3654 * ipa-pure-const.c (finish_state): Remove.
3655 (propagate): Adjust.
3656 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
3657 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
3658 var-tracking.c: Adjust.
3659 * vec.c (stack_vecs): Remove.
3660 (register_stack_vec): Likewise.
3661 (stack_vec_register_index): Likewise.
3662 (unregister_stack_vec): Likewise.
3663 * vec.h (struct va_stack): Remove.
3664 (struct vec<T, A, vl_ptr>): Specialize as
3665 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
3666 allocation strategy compatable with the vl_ptr layout.
3667 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
3668 specialization anyway.
3669 (class stack_vec): New class.
3670 (vec_stack_alloc): Remove.
3671 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
3673 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
3674 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3675 Sergey Lega <sergey.s.lega@intel.com>
3676 Anna Tikhonova <anna.tikhonova@intel.com>
3677 Ilya Tocar <ilya.tocar@intel.com>
3678 Andrey Turetskiy <andrey.turetskiy@intel.com>
3679 Ilya Verbin <ilya.verbin@intel.com>
3680 Kirill Yukhin <kirill.yukhin@intel.com>
3681 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3683 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
3684 (*prefetch_avx512pf_<mode>): New.
3685 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
3686 (avx512f_maskcmp<mode>3): Ditto.
3687 (vashrv16si3): Ditto.
3689 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
3690 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3691 Sergey Lega <sergey.s.lega@intel.com>
3692 Anna Tikhonova <anna.tikhonova@intel.com>
3693 Ilya Tocar <ilya.tocar@intel.com>
3694 Andrey Turetskiy <andrey.turetskiy@intel.com>
3695 Ilya Verbin <ilya.verbin@intel.com>
3696 Kirill Yukhin <kirill.yukhin@intel.com>
3697 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3699 * config/i386/i386.md (any_truncate): New.
3700 (trunsuffix): Ditto.
3701 * config/i386/predicates.md (const_8_to_9_operand): New.
3702 (const_10_to_11_operand): Ditto.
3703 (const_12_to_13_operand): Ditto.
3704 (const_14_to_15_operand): Ditto.
3705 (const_16_to_19_operand): Ditto.
3706 (const_20_to_23_operand): Ditto.
3707 (const_24_to_27_operand): Ditto.
3708 (const_28_to_31_operand): Ditto.
3709 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
3710 (cvtusi2<ssescalarmodesuffix>32): New.
3711 (cvtusi2<ssescalarmodesuffix>64): Ditto.
3712 (ufloatv16siv16sf2): Ditto.
3713 (avx512f_fix_notruncv16sfv16si): Ditto.
3714 (avx512f_ufix_notruncv16sfv16si): Ditto.
3715 (avx512f_vcvtss2usi): Ditto.
3716 (avx512f_vcvtss2usiq): Ditto.
3717 (avx512f_vcvttss2usi): Ditto.
3718 (avx512f_vcvttss2usiq): Ditto.
3719 (avx512f_vcvtsd2usi): Ditto.
3720 (avx512f_vcvtsd2usiq): Ditto.
3721 (avx512f_vcvttsd2usi): Ditto.
3722 (avx512f_vcvttsd2usiq): Ditto.
3723 (ufloatv8siv8df): Ditto.
3724 (avx512f_cvtdq2pd512_2): Ditto.
3725 (avx512f_cvtpd2dq512): Ditto.
3726 (avx512f_ufix_notruncv8dfv8si): Ditto.
3727 (avx512f_cvtpd2ps512): Ditto.
3728 (vec_unpacks_lo_v16sf): Ditto.
3729 (vec_unpacks_hi_v16sf): Ditto.
3730 (vec_unpacks_float_hi_v16si): Ditto.
3731 (vec_unpacks_float_lo_v16si): Ditto.
3732 (avx512f_unpckhps512): Ditto.
3733 (avx512f_unpcklps512): Ditto.
3734 (avx512f_movshdup512): Ditto.
3735 (avx512f_movsldup512): Ditto.
3736 (vec_extract_lo_v32hi): Ditto.
3737 (vec_extract_hi_v32hi): Ditto.
3738 (vec_extract_lo_v64qi): Ditto.
3739 (vec_extract_hi_v64qi): Ditto.
3740 (avx512f_unpckhpd512): Ditto.
3741 (avx512f_movddup512): Ditto.
3742 (avx512f_unpcklpd512): Ditto.
3743 (*avx512f_unpcklpd512): Ditto.
3744 (avx512f_shufps512_1): Ditto.
3745 (avx512f_shufpd512_1): Ditto.
3746 (avx512f_interleave_highv8di): Ditto.
3747 (avx512f_interleave_lowv8di): Ditto.
3748 (PMOV_DST_MODE): Ditto.
3749 (pmov_src_mode): Ditto.
3750 (pmov_src_lower): Ditto.
3752 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
3753 (*avx512f_<code>v8div16qi2): Ditto.
3754 (*avx512f_<code>v8div16qi2_store): Ditto.
3755 (vec_widen_umult_even_v16si): Ditto.
3756 (*vec_widen_umult_even_v16si): Ditto.
3757 (vec_widen_smult_even_v16si): Ditto.
3758 (*vec_widen_smult_even_v16si): Ditto.
3759 (avx512f_interleave_highv16si): Ditto.
3760 (avx512f_interleave_lowv16si): Ditto.
3761 (avx512f_<code>v16qiv16si2): Ditto.
3762 (avx512f_<code>v16hiv16si2): Ditto.
3763 (avx512f_<code>v8qiv8di2): Ditto.
3764 (avx512f_<code>v8hiv8di2): Ditto.
3765 (avx512f_<code>v8siv8di2): Ditto.
3766 (avx512cd_maskb_vec_dupv8di): Ditto.
3767 (avx512cd_maskw_vec_dupv16si): Ditto.
3768 (avx512f_vcvtph2ps512): Ditto.
3769 (avx512f_vcvtps2ph512): Ditto.
3770 (VEC_EXTRACT_MODE): Extened with wider modes.
3771 (VEC_PERM_AVX2): Ditto.
3772 (VEC_PERM_CONST): Ditto.
3774 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
3776 * config/arc/arc.c (arc_ccfsm_post_advance):
3777 Add comment about TYPE_RETURN.
3779 2013-10-28 Bin Cheng <bin.cheng@arm.com>
3781 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
3782 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
3783 (strip_offset): Convert offset to unsigned number.
3785 2013-10-27 Tom de Vries <tom@codesourcery.com>
3787 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
3788 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
3790 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
3792 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
3793 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
3794 (TARGET_BIG_ENDIAN): New macro.
3795 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
3796 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
3797 * config/sh/sh.c: Likewise.
3798 * config/sh/sh.h: Likewise.
3800 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
3802 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
3803 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
3805 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
3807 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
3809 (cmpstr_t, cmpstrsi): Fix formatting.
3811 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
3814 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
3815 addressing, do not use general_operand for memory operands.
3817 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
3820 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
3821 * lra-spills.c (lra_final_code_change): Remove useless move insns.
3823 2013-10-26 Jeff Law <law@redhat.com>
3825 * predict.c (PRED_MUDFLAP): Remove.
3826 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
3828 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
3829 (OBJS): Remove tree-nomudflap.o
3830 (GTFILES): Remove tree-mudflap.c
3831 * builtins.c (expand_builtin_alloc): Remove mudflap support.
3832 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
3833 (mfwrap_spec, mflib_spec): Likewise.
3834 (cpp_unique_options, cc1_options, static_specs): Likewise.
3835 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
3836 * passes.def: Likewise.
3837 * toplev.c (compile_file, process_options): Likewise.
3838 * tree-inline.c (copy_tree_r): Likewise.
3839 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
3840 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
3841 (build_constant_desc, output_constant_def_contents): Likewise.
3842 (categorize_decl_for_section): Likewise.
3843 * tree-mudflap.c: Removed.
3844 * tree-mudflap.h: Removed.
3845 * tree-nomudflap.c: Removed.
3846 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
3847 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
3848 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
3849 * config/sol2.h (MFLIB_SPEC): Likewise.
3850 * doc/install.texi: Remove mudflap references.
3851 * doc/passes.texi: Similarly.
3852 * doc/sourcebuild.texi: Similarly.
3853 * doc/invoke.texi: Remove mudlfap related options.
3855 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
3857 PR rtl-optimization/58759
3858 * lra-constraints.c (lra_constraints): Remove wrong condition to
3859 remove insn setting up an equivalent pseudo.
3861 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
3863 * config/rs6000/rs6000-protos.h
3864 (rs6000_secondary_memory_needed_mode): New prototype.
3865 * config/rs6000/rs6000.c: Include ira.h.
3866 (TARGET_LRA_P): Redefine.
3867 (rs6000_legitimate_offset_address_p): Call
3868 legitimate_constant_pool_address_p in strict mode for LRA.
3869 (rs6000_legitimate_address_p): Ditto.
3870 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
3871 (rs6000_emit_move): Add LRA version of code to generate load/store
3873 (rs6000_secondary_memory_needed_mode): New.
3874 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
3875 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
3876 constants, memory, and FP registers.
3877 (rs6000_lra_p): New.
3878 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
3879 * config/rs6000/rs6000.opt (mlra): New option.
3880 * lra-spills.c (lra_final_code_change): Remove useless move insns.
3882 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
3884 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
3885 has_single_use () and not do the conversion if has_single_use ()
3886 returns false for the multiplication result.
3888 2013-10-25 David Malcolm <dmalcolm@redhat.com>
3890 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
3891 to "NODE", since this works on a "tree", not an
3893 (CONSTANT_CLASS_P): Likewise.
3896 (INDIRECT_REF_P): Likewise.
3897 (REFERENCE_CLASS_P): Likewise.
3898 (COMPARISON_CLASS_P): Likewise.
3899 (UNARY_CLASS_P): Likewise.
3900 (BINARY_CLASS_P): Likewise.
3901 (STATEMENT_CLASS_P): Likewise.
3902 (VL_EXP_CLASS_P): Likewise.
3903 (EXPRESSION_CLASS_P): Likewise.
3904 (IS_TYPE_OR_DECL_P): Likewise.
3906 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
3908 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
3909 ADDR_EXPR in the defining statement.
3911 2013-10-25 Richard Biener <rguenther@suse.de>
3913 PR tree-optimization/58626
3914 * tree-loop-distribution.c (enum rdg_dep_type): Remove
3915 anti_dd, output_dd and input_dd.
3916 (struct rdg_edge): Remove level and relation members.
3917 (RDGE_LEVEL, RDGE_RELATION): Remove.
3918 (dot_rdg_1): Adjust.
3919 (create_rdg_edge_for_ddr): Remove.
3920 (create_rdg_edges_for_scalar): Adjust.
3921 (create_edge_for_control_dependence): Likewise.
3922 (create_rdg_edges): Split into ...
3923 (create_rdg_flow_edges): ... this
3924 (create_rdg_cd_edges): ... and this.
3926 (build_rdg): Likewise, do not compute data dependences or
3928 (pg_add_dependence_edges): New function.
3930 (distribute_loop): First apply all non-dependence based
3931 partition mergings. Then compute dependences between partitions
3932 and merge and order partitions according to them.
3934 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
3936 PR rtl-optimization/58831
3937 * alias.c (init_alias_analysis): At the beginning of each iteration,
3938 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
3940 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
3942 * recog.c (search_ofs): New static variable moved from...
3943 (peep2_find_free_register): ...here.
3944 (peephole2_optimize): Initialize it.
3946 2013-10-25 Tobias Burnus <burnus@net-b.de>
3948 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
3950 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
3952 * config/i386/i386.h (TARGET_MPX): New define.
3953 (TARGET_MPX_P): Ditto.
3955 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
3957 * config/i386/constraints.md (B): New.
3960 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
3961 * config/i386/i386-modes.def (BND32): New.
3963 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
3964 * config/i386/i386.c (isa_opts): Add mmpx.
3965 (regclass_map): Add bound registers.
3966 (dbx_register_map): Likewise.
3967 (dbx64_register_map): Likewise.
3968 (svr4_dbx_register_map): Likewise.
3970 (ix86_option_override_internal): Support MPX ISA.
3971 (ix86_conditional_register_usage): Support bound registers.
3972 (print_reg): Likewise.
3973 (ix86_code_end): Add MPX bnd prefix.
3974 (output_set_got): Likewise.
3975 (ix86_output_call_insn): Likewise.
3976 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
3977 (ix86_print_operand_punct_valid_p): Likewise.
3978 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
3980 (ix86_class_likely_spilled_p): Add bound regs support.
3981 (ix86_hard_regno_mode_ok): Likewise.
3982 (x86_order_regs_for_local_alloc): Likewise.
3983 (ix86_bnd_prefixed_insn_p): New.
3984 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
3985 (FIXED_REGISTERS): Add bound registers.
3986 (CALL_USED_REGISTERS): Likewise.
3987 (REG_ALLOC_ORDER): Likewise.
3988 (HARD_REGNO_NREGS): Likewise.
3990 (VALID_BND_REG_MODE): New.
3991 (FIRST_BND_REG): New.
3992 (LAST_BND_REG): New.
3993 (reg_class): Add BND_REGS.
3994 (REG_CLASS_NAMES): Likewise.
3995 (REG_CLASS_CONTENTS): Likewise.
3997 (ANY_BND_REG_P): New.
3999 (HI_REGISTER_NAMES): Add bound registers.
4000 * config/i386/i386.md (UNSPEC_BNDMK): New.
4001 (UNSPEC_BNDMK_ADDR): New.
4002 (UNSPEC_BNDSTX): New.
4003 (UNSPEC_BNDLDX): New.
4004 (UNSPEC_BNDLDX_ADDR): New.
4005 (UNSPEC_BNDCL): New.
4006 (UNSPEC_BNDCU): New.
4007 (UNSPEC_BNDCN): New.
4008 (UNSPEC_MPX_FENCE): New.
4011 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
4012 (length_immediate): Likewise.
4013 (prefix_0f): Likewise.
4015 (prefix_rep): Check for bnd prefix.
4016 (length_nobnd): New.
4017 (length): Use length_nobnd if specified.
4022 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
4025 (simple_return_internal): Likewise.
4026 (simple_return_pop_internal): Likewise.
4027 (*indirect_jump): Add MPX bnd prefix.
4028 (*tablejump_1): Likewise.
4029 (simple_return_internal_long): Likewise.
4030 (simple_return_indirect_internal): Likewise.
4034 (*mov<mode>_internal_mpx): New.
4035 (<mode>_<bndcheck>): New.
4036 (*<mode>_<bndcheck>): New.
4041 * config/i386/predicates.md (lea_address_operand): Rename to...
4042 (address_no_seg_operand): ... this.
4043 (address_mpx_no_base_operand): New.
4044 (address_mpx_no_index_operand): New.
4045 (bnd_mem_operator): New.
4046 * config/i386/i386.opt (mmpx): New.
4047 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
4048 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
4050 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
4052 * mode-classes.def (MODE_POINTER_BOUNDS): New.
4053 * tree.def (POINTER_BOUNDS_TYPE): New.
4054 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
4055 (POINTER_BOUNDS_MODE): New.
4056 (make_pointer_bounds_mode): New.
4057 * machmode.h (POINTER_BOUNDS_MODE_P): New.
4058 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
4059 (layout_type): Support POINTER_BOUNDS_TYPE.
4060 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
4061 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
4062 (type_contains_placeholder_1): Likewise.
4063 * tree.h (POINTER_BOUNDS_TYPE_P): New.
4064 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
4065 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
4067 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
4069 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
4070 mode of address computation when calling memory_address_addr_space.
4072 2013-08-24 Richard Henderson <rth@twiddle.net>
4075 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
4076 instead of create_convert_operand_to.
4077 (maybe_emit_sync_lock_test_and_set): Likewise.
4078 (expand_atomic_compare_and_swap): Likewise.
4079 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
4081 2013-08-24 Sriraman Tallam <tmsriram@google.com>
4083 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
4084 Change param_is to use the struct and not the pointer to the struct.
4086 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
4088 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
4089 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
4090 Move to targhooks.c.
4091 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
4092 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
4093 Relocate from gimplify.c.
4094 * targhooks.h: Add 2 prototypes.
4095 * tree.h: Delete 2 prototypes.
4097 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
4099 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
4100 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
4102 2013-10-24 Cong Hou <congh@google.com>
4104 * convert.c (convert_to_real): Guard those unsafe math function
4105 convertions with flag_unsafe_math_optimizations. Handle sqrt()
4108 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
4111 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
4113 (cgraph_create_edge): Use the new argument.
4114 (cgraph_create_indirect_edge): Likewise.
4116 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
4118 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
4120 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
4121 changes statep->state.
4123 2013-10-24 Tobias Burnus <burnus@net-b.de>
4126 * tree-cfg.c (replace_loop_annotate): New function.
4127 (execute_build_cfg): Call it.
4128 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
4129 * internal-fn.c (expand_ANNOTATE): New function.
4130 * internal-fn.def (ANNOTATE): Define as new internal function.
4131 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
4132 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
4133 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
4134 * doc/extend.texi (Pragmas): Document #pragma ivdep.
4135 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
4137 2013-10-17 Ian Bolton <ian.bolton@arm.com>
4138 Marcus Shawcroft <marcus.shawcroft@arm.com>
4140 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
4141 Special case reload SP+C into none GENERAL_REGS.
4143 2013-10-24 Michael Matz <matz@suse.de>
4145 * gengtype.c (is_file_equal): Check that files will be same length.
4147 2013-10-25 Christian Bruel <christian.bruel@st.com>
4149 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
4150 * config/sh/t-sh (sh-mem.o): New rule.
4151 * config/sh/sh-mem.cc (expand_block_move): Moved here.
4152 (sh_expand_cmpstr): New function.
4153 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
4154 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
4155 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
4156 (rotlhi3_8): Rename.
4158 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4160 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
4161 * Makefile.in (CXXDEPMODE): Assign and change users.
4162 (CCDEPMODE): Delete.
4163 * configure: Regenerate.
4165 2013-10-23 David Malcolm <dmalcolm@redhat.com>
4167 * gengtype-parse.c (require_without_advance): New.
4168 (type): For GTY-marked types that are not GTY((user)), parse any
4169 base classes, requiring them to be single-inheritance, and not
4170 be templates. For non-GTY-marked types and GTY((user)),
4171 continue to skip over any C++ inheritance specification.
4172 * gengtype-state.c (state_writer::write_state_struct_type):
4173 Write base class of type (if any).
4174 (read_state_struct_type): Read base class of type (if any).
4175 * gengtype.c (new_structure): Add a "base_class" parameter.
4176 (create_optional_field_): Update for new parameter to new_structure.
4177 (adjust_field_rtx_def): Likewise.
4178 (adjust_field_tree_exp): Likewise.
4179 * gengtype.h (struct type): Add "base_class" field to the s
4181 (new_structure): Add "base" parameter.
4183 2013-10-23 Sriraman Tallam <tmsriram@google.com>
4186 * config/i386/i386.c (ix86_option_override_internal):
4187 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
4188 (ix86_valid_target_attribute_tree):
4189 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
4190 Change TARGET_SSE to TARGET_SSE_P (opts->...)
4192 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
4194 * tree-ssa-loop.h: Remove include files.
4195 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
4196 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
4197 * graphite-clast-to-gimple.c: Likewise.
4198 * graphite-scop-detection.c: Likewise.
4199 * graphite-sese-to-poly.c: Likewise.
4200 * ipa-inline-analysis.c: Likewise.
4201 * ipa-pure-const.c: Likewise.
4202 * loop-init.c: Likewise.
4203 * passes.c: Likewise.
4204 * predict.c: Likewise.
4205 * tree-cfg.c: Likewise.
4206 * tree-cfgcleanup.c: Likewise.
4207 * tree-chrec.c: Likewise.
4208 * tree-data-ref.c: Likewise.
4209 * tree-loop-distribution.c: Likewise.
4210 * tree-parloops.c: Likewise.
4211 * tree-predcom.c: Likewise.
4212 * tree-scalar-evolution.c: Likewise.
4213 * tree-ssa-address.c: Likewise.
4214 * tree-ssa.c: Likewise.
4215 * tree-ssa-dce.c: Likewise.
4216 * tree-ssa-loop.c: Likewise.
4217 * tree-ssa-loop-im.c: Likewise.
4218 * tree-ssa-loop-ivcanon.c: Likewise.
4219 * tree-ssa-loop-ivopts.c: Likewise.
4220 * tree-ssa-loop-manip.c: Likewise.
4221 * tree-ssa-loop-niter.c: Likewise.
4222 * tree-ssa-loop-prefetch.c: Likewise.
4223 * tree-ssa-loop-unswitch.c: Likewise.
4224 * tree-ssa-reassoc.c: Likewise.
4225 * tree-vect-data-refs.c: Likewise.
4226 * tree-vect-loop.c: Likewise.
4227 * tree-vect-loop-manip.c: Likewise.
4228 * tree-vectorizer.c: Likewise.
4229 * tree-vect-stmts.c: Likewise.
4230 * tree-vrp.c: Likewise.
4232 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4234 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
4236 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4238 PR tree-optimization/58775
4239 PR tree-optimization/58791
4240 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
4241 (insert_stmt_after): Rewritten, don't move the stmt, but really
4243 (get_stmt_uid_with_default): Remove.
4244 (build_and_add_sum): Use insert_stmt_after and
4245 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
4246 (update_range_test): Set uid on stmts added by
4247 force_gimple_operand_gsi. Don't immediately modify statements
4248 in inter-bb optimization, just update oe->op values.
4249 (optimize_range_tests): Return bool whether any changed have been made.
4250 (update_ops): New function.
4251 (struct inter_bb_range_test_entry): New type.
4252 (maybe_optimize_range_tests): Perform statement changes here.
4253 (not_dominated_by, appears_later_in_bb, get_def_stmt,
4254 ensure_ops_are_available): Remove.
4255 (find_insert_point): Rewritten.
4256 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
4257 return LHS of the (new resp. old) stmt. Don't call
4258 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
4259 instead of last, move new stmt at the right place.
4260 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
4261 (negate_value): Likewise. Set uids.
4262 (break_up_subtract_bb): Initialize uids.
4263 (reassociate_bb): Adjust rewrite_expr_tree caller.
4264 (do_reassoc): Don't call renumber_gimple_stmt_uids.
4266 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
4269 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
4270 TARGET_32BIT final condition.
4271 (mulsi3_internal2 and splitter): Same.
4273 2013-10-23 Jeff Law <law@redhat.com>
4275 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
4276 through joiner blocks with abnormal outgoing edges.
4278 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
4279 Add parameter JOINERS, to allow/disallow threading through joiner
4281 (thread_block): New. Call thread_block_1.
4282 (mark_threaded_blocks): Remove code to filter out certain cases
4283 of threading through joiner blocks.
4284 (thread_through_all_blocks): Document how we can have a dangling
4285 edge AUX field and clear it.
4287 2013-10-23 Ian Lance Taylor <iant@google.com>
4289 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
4290 (C++ Dialect Options): Likewise.
4291 (Optimize Options): Likewise.
4293 2013-10-23 Tom de Vries <tom@codesourcery.com>
4295 PR tree-optimization/58805
4296 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
4298 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4300 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
4301 sequence based on get_range_info returned range.
4303 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
4305 * tree-ssa.h: Remove all #include's
4306 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
4307 * alias.c: Move required includes from tree-ssa.h.
4309 * builtins.c: Likewise.
4310 * calls.c: Likewise.
4311 * cfgexpand.c: Likewise.
4312 * cfghooks.c: Likewise.
4313 * cfgloop.c: Likewise.
4314 * cfgloopmanip.c: Likewise.
4315 * cgraph.c: Likewise.
4316 * cgraphbuild.c: Likewise.
4317 * cgraphclones.c: Likewise.
4318 * cgraphunit.c: Likewise.
4320 * except.c: Likewise.
4322 * final.c: Likewise.
4323 * fold-const.c: Likewise.
4324 * ggc-page.c: Likewise.
4325 * gimple-builder.c: Likewise.
4326 * gimple-fold.c: Likewise.
4327 * gimple-iterator.c: Likewise.
4328 * gimple-low.c: Likewise.
4329 * gimple-pretty-print.c: Likewise.
4330 * gimple-ssa-strength-reduction.c: Likewise.
4331 * gimple-streamer-in.c: Likewise.
4332 * gimple-streamer-out.c: Likewise.
4333 * gimplify.c: Likewise.
4334 * graphite-blocking.c: Likewise.
4335 * graphite-clast-to-gimple.c: Likewise.
4336 * graphite-dependences.c: Likewise.
4337 * graphite-interchange.c: Likewise.
4338 * graphite-optimize-isl.c: Likewise.
4339 * graphite-poly.c: Likewise.
4340 * graphite-scop-detection.c: Likewise.
4341 * graphite-sese-to-poly.c: Likewise.
4342 * graphite.c: Likewise.
4343 * ipa-cp.c: Likewise.
4344 * ipa-inline-analysis.c: Likewise.
4345 * ipa-inline-transform.c: Likewise.
4346 * ipa-inline.c: Likewise.
4347 * ipa-prop.c: Likewise.
4348 * ipa-pure-const.c: Likewise.
4349 * ipa-reference.c: Likewise.
4350 * ipa-split.c: Likewise.
4351 * ipa-utils.c: Likewise.
4352 * loop-init.c: Likewise.
4353 * lto-cgraph.c: Likewise.
4354 * lto-section-in.c: Likewise.
4355 * lto-section-out.c: Likewise.
4356 * lto-streamer-in.c: Likewise.
4357 * lto-streamer-out.c: Likewise.
4358 * lto-streamer.c: Likewise.
4359 * omp-low.c: Likewise.
4360 * passes.c: Likewise.
4361 * predict.c: Likewise.
4362 * print-tree.c: Likewise.
4363 * profile.c: Likewise.
4365 * targhooks.c: Likewise.
4366 * tracer.c: Likewise.
4367 * trans-mem.c: Likewise.
4368 * tree-call-cdce.c: Likewise.
4369 * tree-cfg.c: Likewise.
4370 * tree-cfgcleanup.c: Likewise.
4371 * tree-chrec.c: Likewise.
4372 * tree-complex.c: Likewise.
4373 * tree-data-ref.c: Likewise.
4374 * tree-dfa.c: Likewise.
4375 * tree-eh.c: Likewise.
4376 * tree-emutls.c: Likewise.
4377 * tree-if-conv.c: Likewise.
4378 * tree-inline.c: Likewise.
4379 * tree-into-ssa.c: Likewise.
4380 * tree-loop-distribution.c: Likewise.
4381 * tree-mudflap.c: Likewise.
4382 * tree-nested.c: Likewise.
4383 * tree-nrv.c: Likewise.
4384 * tree-object-size.c: Likewise.
4385 * tree-outof-ssa.c: Likewise.
4386 * tree-parloops.c: Likewise.
4387 * tree-phinodes.c: Likewise.
4388 * tree-predcom.c: Likewise.
4389 * tree-pretty-print.c: Likewise.
4390 * tree-profile.c: Likewise.
4391 * tree-scalar-evolution.c: Likewise.
4392 * tree-sra.c: Likewise.
4393 * tree-ssa-address.c: Likewise.
4394 * tree-ssa-alias.c: Likewise.
4395 * tree-ssa-ccp.c: Likewise.
4396 * tree-ssa-coalesce.c: Likewise.
4397 * tree-ssa-copy.c: Likewise.
4398 * tree-ssa-copyrename.c: Likewise.
4399 * tree-ssa-dce.c: Likewise.
4400 * tree-ssa-dom.c: Likewise.
4401 * tree-ssa-dse.c: Likewise.
4402 * tree-ssa-forwprop.c: Likewise.
4403 * tree-ssa-ifcombine.c: Likewise.
4404 * tree-ssa-live.c: Likewise.
4405 * tree-ssa-loop-ch.c: Likewise.
4406 * tree-ssa-loop-im.c: Likewise.
4407 * tree-ssa-loop-ivcanon.c: Likewise.
4408 * tree-ssa-loop-ivopts.c: Likewise.
4409 * tree-ssa-loop-manip.c: Likewise.
4410 * tree-ssa-loop-niter.c: Likewise.
4411 * tree-ssa-loop-prefetch.c: Likewise.
4412 * tree-ssa-loop-unswitch.c: Likewise.
4413 * tree-ssa-loop.c: Likewise.
4414 * tree-ssa-math-opts.c: Likewise.
4415 * tree-ssa-operands.c: Likewise.
4416 * tree-ssa-phiopt.c: Likewise.
4417 * tree-ssa-phiprop.c: Likewise.
4418 * tree-ssa-pre.c: Likewise.
4419 * tree-ssa-propagate.c: Likewise.
4420 * tree-ssa-reassoc.c: Likewise.
4421 * tree-ssa-sccvn.c: Likewise.
4422 * tree-ssa-sink.c: Likewise.
4423 * tree-ssa-strlen.c: Likewise.
4424 * tree-ssa-structalias.c: Likewise.
4425 * tree-ssa-tail-merge.c: Likewise.
4426 * tree-ssa-ter.c: Likewise.
4427 * tree-ssa-threadedge.c: Likewise.
4428 * tree-ssa-threadupdate.c: Likewise.
4429 * tree-ssa-uncprop.c: Likewise.
4430 * tree-ssa-uninit.c: Likewise.
4431 * tree-ssa.c: Likewise.
4432 * tree-ssanames.c: Likewise.
4433 * tree-stdarg.c: Likewise.
4434 * tree-streamer-in.c: Likewise.
4435 * tree-switch-conversion.c: Likewise.
4436 * tree-tailcall.c: Likewise.
4437 * tree-vect-data-refs.c: Likewise.
4438 * tree-vect-generic.c: Likewise.
4439 * tree-vect-loop-manip.c: Likewise.
4440 * tree-vect-loop.c: Likewise.
4441 * tree-vect-patterns.c: Likewise.
4442 * tree-vect-slp.c: Likewise.
4443 * tree-vect-stmts.c: Likewise.
4444 * tree-vectorizer.c: Likewise.
4445 * tree-vrp.c: Likewise.
4448 * value-prof.c: Likewise.
4449 * var-tracking.c: Likewise.
4450 * varpool.c: Likewise.
4451 * vtable-verify.c: Likewise.
4453 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4455 * config/tilegx/tilegx.c: Include "tree.h".
4457 2013-10-23 Jakub Jelinek <jakub@redhat.com>
4459 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
4460 the info, not after it.
4461 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
4462 dump_ssaname_info call.
4463 (pp_gimple_stmt_1): Adjust caller.
4464 (dump_phi_nodes): Likewise. Don't print "# " here.
4466 2013-10-22 Jan Hubicka <jh@suse.cz>
4468 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
4470 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
4471 * i386.c (expand_small_movmem_or_setmem): New function.
4472 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
4474 (alg_usable_p): Add support for value ranges; cleanup.
4475 (ix86_expand_set_or_movmem): Add support for misaligned moves.
4477 2013-10-22 Sterling Augustine <saugustine@google.com>
4479 * doc/invoke.texi: Document -ggnu-pubnames.
4480 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
4482 * dwarf2out.c: Include gdb/gdb-index.h.
4483 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
4484 debug_generate_pub_sections.
4485 (is_java, output_pubtables, output_pubname): New functions.
4486 (include_pubname_in_output): Handle debug_generate_pub_sections at
4488 (size_of_pubnames): Use new local space_for_flags based on
4489 debug_generate_pub_sections.
4490 (output_pubnames): Unify pubnames and pubtypes output logic.
4491 Genericize comments. Call output_pubname.
4492 (dwarf2out_finish): Move logic to output_pubnames and call it.
4494 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
4497 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
4498 Remove CCCmode handling.
4499 <case LTU>: Return 'c' suffix for CCCmode.
4500 <case GEU>: Return 'nc' suffix for CCCmode.
4501 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
4502 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
4503 (*sub<mode>3_cc_overflow): Ditto.
4504 (*subsi3_zext_cc_overflow): Ditto.
4506 2013-10-22 Steve Ellcey <sellcey@mips.com>
4508 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
4509 (AND (NOT OP1) (NOT OP2)).
4511 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
4513 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
4514 meaning of merge-high and merge-low masks for little endian; avoid
4515 use of vector-pack masks for little endian for mismatched modes.
4517 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4519 * config/tilepro/tilepro.c: Include "tree.h".
4521 2013-10-22 Andreas Schwab <schwab@suse.de>
4523 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
4526 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
4528 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
4529 [Ed Smith-Rowland]): New entries.
4530 ([Stephen M. Webb]): Update.
4532 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
4534 * tree-ssa-ter.h: Remove duplicate copy of file contents.
4536 2013-10-21 Marek Polacek <polacek@redhat.com>
4539 * fold-const.c (fold_range_test): Return 0 if the type is not
4542 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
4544 * system.h: Move hwint.h include further down.
4545 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
4546 gcc_checking_asserts.
4547 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
4549 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
4551 Fix volatile issues in optimize_bit_field_compare.
4552 * fold-const.c (optimize_bit_field_compare): Bail out if
4553 lvolatilep or rvolatilep.
4555 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
4557 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
4558 and get_inner_reference returning different pmode for non-volatile
4559 bit-field members dependent on flag_strict_volatile_bitfields.
4560 * stor-layout.c (layout_decl): Remove special handling of
4561 flag_strict_volatile_bitfields.
4562 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
4563 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
4565 2013-10-21 Paulo Matos <pmatos@broadcom.com>
4567 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
4570 2013-10-21 Jeff Law <law@redhat.com>
4572 * tree-ssa-threadedge.c (thread_through_normal_block): New
4573 argument VISITED. Remove VISISTED as a local variable. When we
4574 have a threadable jump, verify the destination of the jump has not
4576 (thread_across_edge): Allocate VISITED bitmap once at function
4577 scope and use it throughout. Make sure to set appropriate bits in
4578 VISITED for E (start of jump thread path).
4579 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
4580 through a joiner if any edge on the path has a recorded jump thread.
4582 2013-10-21 Ian Lance Taylor <iant@google.com>
4584 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
4585 don't imply that attributes can solve all problems.
4586 (Directory Options): Fix typo.
4588 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4590 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
4591 extend and extend_add.
4593 2013-10-21 Richard Biener <rguenther@suse.de>
4595 PR tree-optimization/58794
4596 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
4597 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
4599 2013-10-21 Richard Biener <rguenther@suse.de>
4602 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
4603 to (T) X for sign-changing conversions (or no conversion).
4605 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
4607 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
4609 2013-10-20 Jan Hubicka <jh@suse.cz>
4611 * config/i386/i386-tune.def: Add comment; organize into categories
4613 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
4615 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
4616 argument. Update function comment.
4617 (expand_set_or_movmem_via_rep): New function combining
4618 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
4619 (expand_movmem_via_rep_mov): Remove.
4620 expand_setmem_via_rep_stos): Remove.
4621 (expand_movmem_epilogue): Update calls correspondingly.
4622 (expand_setmem_epilogue_via_loop): Likewise.
4624 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
4625 (expand_set_or_movmem_prologue): New function combining
4626 expand_movmem_prologue and expand_setmem_prologue.
4627 (expand_movmem_prologue): Remove.
4628 (expand_setmem_prologue): Remove.
4629 (expand_set_or_movmem_constant_prologue): New function combining
4630 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
4631 (expand_constant_movmem_prologue): Remove.
4632 (expand_constant_setmem_prologue): Remove.
4633 (promote_duplicated_reg): Allow vector-const0 value.
4634 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
4635 and ix86_expand_setmem.
4636 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
4637 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
4639 2013-10-21 Diego Novillo <dnovillo@google.com>
4641 * asan.c: Include tree.h
4642 * bb-reorder.c: Likewise.
4643 * cfgcleanup.c: Likewise.
4644 * cfgloopmanip.c: Likewise.
4645 * data-streamer-in.c: Likewise.
4646 * data-streamer-out.c: Likewise.
4647 * data-streamer.c: Likewise.
4648 * dwarf2cfi.c: Likewise.
4649 * graphite-blocking.c: Likewise.
4650 * graphite-clast-to-gimple.c: Likewise.
4651 * graphite-dependences.c: Likewise.
4652 * graphite-interchange.c: Likewise.
4653 * graphite-optimize-isl.c: Likewise.
4654 * graphite-poly.c: Likewise.
4655 * graphite-scop-detection.c: Likewise.
4656 * graphite-sese-to-poly.c: Likewise.
4657 * graphite.c: Likewise.
4658 * ipa-devirt.c: Likewise.
4659 * ipa-profile.c: Likewise.
4662 * loop-init.c: Likewise.
4663 * loop-unroll.c: Likewise.
4664 * lower-subreg.c: Likewise.
4665 * lto/lto-object.c: Likewise.
4666 * recog.c: Likewise.
4667 * reginfo.c: Likewise.
4668 * tree-loop-distribution.c: Likewise.
4669 * tree-parloops.c: Likewise.
4670 * tree-ssa-strlen.c: Likewise.
4671 * tree-streamer.c: Likewise.
4672 * value-prof.c: Likewise.
4673 * target-globals.c: Likewise.
4674 * expr.h: Include tree-core.h instead of tree.h.
4675 * gimple.h: Likewise.
4676 * ipa-prop.h: Likewise.
4677 * ipa-utils.h: Likewise.
4678 * lto-streamer.h: Likewise.
4679 * streamer-hooks.h: Likewise.
4680 * ipa-reference.h: Include cgraph.h instead of tree.h.
4681 * cgraph.h: Include basic-block.h instead of tree.h.
4682 * tree-streamer.h: Do not include tree.h.
4683 * genattrtab.c (write_header): Generate inclusion of
4685 * genautomata.c (main): Likewise.
4686 * genemit.c: Likewise.
4687 * genopinit.c: Likewise.
4688 * genoutput.c (output_prologue): Likewise.
4689 * genpeep.c: Likewise.
4691 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4693 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
4695 (vec_unpacku_hi_v8hi): Likewise.
4696 (vec_unpacku_lo_v16qi): Likewise.
4697 (vec_unpacku_lo_v8hi): Likewise.
4699 2013-10-20 Jan Hubicka <jh@suse.cz>
4701 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
4702 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
4703 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
4705 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
4707 * config/mips/mips.h (ISA_HAS_WSBH): Define.
4708 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
4710 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
4712 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
4715 * system.h: Undef m_slot.
4717 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4719 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
4720 all elements for both endian flavors.
4722 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
4725 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
4726 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
4727 and SI_REG for 64bit SYSV ABI targets.
4729 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
4731 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
4732 to multi_reg_return. Clarify that we are skipping USEs of multiple
4733 return registers. Use bool type where appropriate.
4735 2013-10-18 Jan Hubicka <jh@suse.cz>
4737 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
4739 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
4740 (X86_TUNE_PUSH_MEMORY): Likewise.
4741 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
4742 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
4743 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
4745 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
4746 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
4748 (ix86_option_override_internal): Update to use tune features instead
4751 2013-10-18 Cong Hou <congh@google.com>
4753 PR tree-optimization/58508
4754 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
4755 statement that contains data refs with zero-step.
4757 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
4759 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
4760 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
4761 * gimple-low.c (gimple_check_call_arg,
4762 gimple_check_call_matching_types): Move to cgraph.c.
4763 * gimple-low.h: Remove prototype.
4764 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
4765 Relocate from gimple-low.c.
4766 * cgraph.h: Add prototype. Don't include basic-block.h.
4767 * gimplify.c: Add gimple-low to include list.
4768 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
4769 * tree-eh.c: Add gimple-low to include list.
4770 * tree-nested.c: Likewise.
4771 * cfgexpand.c: Add tree-ssa-address.h to include list.
4773 * gimple-fold.c: Likewise.
4774 * gimple-ssa-strength-reduction.c: Likewise.
4775 * trans-mem.c: Likewise.
4776 * tree-mudflap.c: Likewise.
4777 * tree-ssa-loop-ivopts.c: Likewise.
4778 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
4779 (degenerate_phi_result): Move to tree-phinodes.c.
4780 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
4781 * tree-ssa-threadedge.c: Likewise.
4782 * tree-vrp.c: Likewise.
4783 * tree-phinodes.c (degenerate_phi_result): Relocate here.
4784 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
4785 * tree-phinodes.h (degenerate_phi_result): Add prototype.
4786 * tree-ssa-copy.c: Include tree-ssa-dom.h.
4787 * tree-ssa-forwprop.c: Likewise.
4788 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
4789 pass_data_cleanup_cfg_post_optimizing,
4790 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
4791 * tree-optimize.c: Delete File.
4792 * graphite.c: Include tree-cfgcleanup.h.
4793 * passes.c: Likewise.
4794 * tree-cfg.c: Likewise.
4795 * tree-profile.c: Likewise.
4796 * tree-ssa-dse.c: Likewise.
4797 * tree-ssa-loop-ivcanon.c: Likewise.
4798 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
4799 * tree-outof-ssa.c: Include sbitmap.h.
4800 * tree-ssa-live.c: Likewise.
4801 * tree-ssa-propagate.c: Likewise.
4802 * tree-ssa-structalias.c: Likewise.
4803 * tree-stdarg.c: Likewise.
4804 * Makefile.in (OBJS): Delete tree-optimize.o.
4805 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
4806 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
4807 * varasm.c: Include basic-block.h.
4808 * cfgloop.h: Include function.h instead of basic-block.h
4809 (bb_loop_depth): Move to cfgloop.c.
4810 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
4812 2013-10-18 Teresa Johnson <tejohnson@google.com>
4814 * predict.c (probably_never_executed): Compare frequency-based
4815 count to number of training runs.
4816 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
4818 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4820 * config/arm/arm.c (cortexa9_extra_costs): New table.
4821 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
4823 2013-10-18 Jeff Law <law@redhat.com>
4825 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
4827 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
4828 (register_jump_thread): Add "registered_jump_thread" debug
4830 * dbgcnt.def (registered_jump_thread): New debug counter.
4832 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
4834 * config/rs6000/rs6000.c: Include cgraph.h.
4836 2013-10-18 Teresa Johnson <tejohnson@google.com>
4838 * tree-ssa-tail-merge.c (replace_block_by): Update edge
4839 weights during merging.
4841 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
4843 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
4844 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
4845 * tree-ssa-operands.h: Remove prototype.
4846 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
4847 * gimple.c (virtual_operand_p): Relocate from gimple.c.
4848 * gimple.h: Add prototype.
4849 * gimple-ssa.h: Include tree-ssa-operands.h.
4850 * tree-dump.c: Add tree-cfg.h to include list.
4851 * tree-ssa-alias.c: Add ipa-reference.h to include list.
4852 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
4853 * config/i386/i386.c: Don't include tree-flow.h.
4854 * config/rs6000/rs6000.c: Likewise.
4856 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4858 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
4860 2013-10-18 Richard Biener <rguenther@suse.de>
4862 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
4863 or TYPE_UNSIGNED of integral types.
4864 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
4865 NULL_TREE for zero-precision integral types.
4867 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
4869 * config/aarch64/arm_neon.h
4870 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
4872 2013-10-17 Sriraman Tallam <tmsriram@google.com>
4875 * opth-gen.awk: Define target_flags_explicit.
4877 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4879 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
4880 fields to the reg_addr array that describes the valid addressing
4881 mode for any register, general purpose registers, floating point
4882 registers, and Altivec registers.
4883 (FIRST_RELOAD_REG_CLASS): Likewise.
4884 (LAST_RELOAD_REG_CLASS): Likewise.
4885 (struct reload_reg_map_type): Likewise.
4886 (reload_reg_map_type): Likewise.
4887 (RELOAD_REG_VALID): Likewise.
4888 (RELOAD_REG_MULTIPLE): Likewise.
4889 (RELOAD_REG_INDEXED): Likewise.
4890 (RELOAD_REG_OFFSET): Likewise.
4891 (RELOAD_REG_PRE_INCDEC): Likewise.
4892 (RELOAD_REG_PRE_MODIFY): Likewise.
4893 (reg_addr): Likewise.
4894 (mode_supports_pre_incdec_p): New helper functions to say whether
4895 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
4896 (mode_supports_pre_modify_p): Likewise.
4897 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
4898 print the valid address mode bits for each mode.
4899 (rs6000_debug_print_mode): Likewise.
4900 (rs6000_debug_reg_global): Likewise.
4901 (rs6000_setup_reg_addr_masks): New function to set up the address
4902 mask bits for each type.
4903 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
4904 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
4905 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
4906 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
4907 PRE_MODIFY are supported.
4908 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
4909 registers, instead of {src,dest}_av_p.
4910 (rs6000_print_options_internal): Tweak the debug output slightly.
4912 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
4914 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
4917 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
4919 * tree-flow.h (struct omp_region): Move to omp-low.c.
4920 Remove omp_ prototypes and variables.
4921 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
4922 (copy_var_decl): Relocate prototype from tree-flow.h.
4923 * gimple.c (copy_var_decl): Relocate from omp-low.c.
4924 * tree.h: Move prototype to omp-low.h.
4925 * omp-low.h: New File. Relocate prototypes here.
4926 * omp-low.c (struct omp_region): Make local here.
4927 (root_omp_region): Make static.
4928 (copy_var_decl) Move to gimple.c.
4929 (new_omp_region): Make static.
4930 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
4931 * tree-cfg.c: Include omp-low.h.
4932 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
4933 * gimplify.c: Include omp-low.h.
4934 * tree-parloops.c: Likewise.
4936 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
4938 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
4939 operands are in memory, prefer to force non-matched operand 1 to
4942 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4945 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
4946 restrict TImode addresses to single indirect registers if both
4947 -mquad-memory and -mvsx-timode are used.
4948 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
4949 we should emit load/store quad. Remove using %y for quad memory
4952 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
4953 constraints to allow load/store quad on machines where TImode is
4954 not allowed in VSX registers. Use 'n' instead of 'F' constraint
4955 for TImode to load integer constants.
4957 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4959 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
4961 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
4963 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
4964 handling of STACK_REG.
4966 2013-10-17 Richard Biener <rguenther@suse.de>
4968 PR tree-optimization/58143
4969 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
4971 (rewrite_to_defined_overflow): Likewise.
4972 (move_computations_dom_walker::before_dom): Rewrite stmts
4973 with undefined signed overflow that are not always executed
4974 into unsigned arithmetic.
4976 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
4979 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
4980 explicit isa flag to be an options variable, instead of using
4981 global_options_set. Remove define from rs6000.h.
4982 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
4984 * config/rs6000/rs6000.c (rs6000_option_override_internal):
4985 Initialize rs6000_isa_flags_explicit.
4986 (rs6000_function_specific_save): Add gcc_options* parameter, so
4987 that the powerpc builds after the 2013-10-15 changes.
4988 (rs6000_function_specific_restore): Likewise.
4990 2013-10-16 DJ Delorie <dj@redhat.com>
4992 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
4993 op is a REG before checking REGNO.
4994 (rl78_alloc_physical_registers): Verify pattern is a SET before
4997 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4999 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
5001 (vec_unpacks_lo_v4sf): Likewise.
5002 (vec_unpacks_float_hi_v4si): Likewise.
5003 (vec_unpacks_float_lo_v4si): Likewise.
5004 (vec_unpacku_float_hi_v4si): Likewise.
5005 (vec_unpacku_float_lo_v4si): Likewise.
5007 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5009 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
5010 (vsx_concat_v2sf): Likewise.
5012 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
5014 * config/aarch64/aarch64.md
5015 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
5017 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
5019 PR tree-optimization/58697
5020 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
5021 estimated_loop_iterations_int.
5022 (max_stmt_executions_int): Call get_max_loop_iterations_int.
5023 (get_max_loop_iterations_int): New. HWINT version of
5024 get_max_loop_iterations.
5025 * cfgloop.h: Add prototypes.
5026 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
5027 * loop-unroll.c (decide_peel_once_rolling): Call
5028 get_estimated_loop_iterations_int.
5029 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
5030 * tree-ssa-loop-niter.h: Tweak prototypes.
5032 2013-10-16 David Malcolm <dmalcolm@redhat.com>
5034 * gengtype-parse.c (struct_field_seq): Ignore access-control
5035 keywords ("public:" etc).
5037 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
5039 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
5040 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
5042 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
5044 * config/arm/arm.opt (mlra): New option.
5045 * config/arm/arm.c (arm_lra_p): New function.
5046 (TARGET_LRA_P): Define.
5048 2013-10-16 Paulo Matos <pmatos@broadcom.com>
5050 * tree-core.h (tree_code_name): Remove.
5051 * tree.h (get_tree_code_name): New prototype.
5052 * tree.c (tree_code_name): Make static.
5053 (get_tree_code_name): New function.
5054 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
5055 tree_class_check_failed, tree_range_check_failed,
5056 tree_not_class_check_failed, omp_clause_check_failed,
5057 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
5058 wrapper get_tree_code_name instead of calling tree_code_name directly.
5059 * tree-vrp.c (dump_asserts_for): Likewise.
5060 * tree-dump.c (dequeue_and_dump): Likewise.
5061 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
5062 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
5063 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
5064 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
5065 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
5066 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
5067 dump_gimple_omp_for): Likewise.
5068 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
5069 * tree-ssa-pre.c (print_pre_expr): Likewise.
5070 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
5071 * print-tree.c (print_node_brief, print_node): Likewise.
5072 * gimple.c (gimple_check_failed): Likewise.
5073 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
5074 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
5075 * config/mep/mep.c (mep_validate_vliw): Likewise.
5076 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
5077 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
5079 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
5081 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
5084 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
5086 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
5087 (MULTILIB_MATCHES): Add multilib for -march=v8.
5089 2013-10-15 Sriraman Tallam <tmsriram@google.com>
5092 * optc-save-gen.awk: Add extra parameter to the save and restore
5094 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
5095 * tree.c (build_optimization_node): New parameter. Add extra parameter
5096 to call to cl_optimization_save.
5097 (build_target_option_node): New parameter. Add extra parameter
5098 to call to cl_target_option_save.
5099 * tree.h (build_optimization_node): New parameter.
5100 (build_target_option_node): New parameter.
5101 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
5102 build_optimization_node and build_target_option_node.
5103 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
5104 (handle_pragma_push_options): Ditto.
5105 * toplev.c (process_options): Ditto.
5106 * opts.c (init_options_struct): Check for opts_set non-null.
5107 * target.def (target_option.save): New parameter.
5108 (target_option.restore): New parameter.
5109 * tm.texi: Generate.
5110 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
5111 (ix86_pragma_target_parse): Ditto.
5112 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
5114 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
5115 to build_optimization_node and build_target_option_node.
5116 (rs6000_valid_attribute_p): Ditto.
5117 (rs6000_pragma_target_parse): Ditto.
5118 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
5120 * config/i386/i386.h:
5121 TARGET_64BIT_P: New Macro
5122 TARGET_MMX_P: New Macro.
5123 TARGET_3DNOW_P: New Macro.
5124 TARGET_3DNOW_A_P: New Macro.
5125 TARGET_SSE_P: New Macro.
5126 TARGET_SSE2_P: New Macro.
5127 TARGET_SSE3_P: New Macro.
5128 TARGET_SSSE3_P: New Macro.
5129 TARGET_SSE4_1_P: New Macro.
5130 TARGET_SSE4_2_P: New Macro.
5131 TARGET_AVX_P: New Macro.
5132 TARGET_AVX2_P: New Macro.
5133 TARGET_AVX512F_P: New Macro.
5134 TARGET_AVX512PF_P: New Macro.
5135 TARGET_AVX512ER_P: New Macro.
5136 TARGET_AVX512CD_P: New Macro.
5137 TARGET_FMA_P: New Macro.
5138 TARGET_SSE4A_P: New Macro.
5139 TARGET_FMA4_P: New Macro.
5140 TARGET_XOP_P: New Macro.
5141 TARGET_LWP_P: New Macro.
5142 TARGET_ABM_P: New Macro.
5143 TARGET_BMI_P: New Macro.
5144 TARGET_BMI2_P: New Macro.
5145 TARGET_LZCNT_P: New Macro.
5146 TARGET_TBM_P: New Macro.
5147 TARGET_POPCNT_P: New Macro.
5148 TARGET_SAHF_P: New Macro.
5149 TARGET_MOVBE_P: New Macro.
5150 TARGET_CRC32_P: New Macro.
5151 TARGET_AES_P: New Macro.
5152 TARGET_PCLMUL_P: New Macro.
5153 TARGET_CMPXCHG16B_P: New Macro.
5154 TARGET_FSGSBASE_P: New Macro.
5155 TARGET_RDRND_P: New Macro.
5156 TARGET_F16C_P: New Macro.
5157 TARGET_RTM_P: New Macro.
5158 TARGET_HLE_P: New Macro.
5159 TARGET_RDSEED_P: New Macro.
5160 TARGET_PRFCHW_P: New Macro.
5161 TARGET_ADX_P: New Macro.
5162 TARGET_FXSR_P: New Macro.
5163 TARGET_XSAVE_P: New Macro.
5164 TARGET_XSAVEOPT_P: New Macro.
5165 TARGET_LP64_P: New Macro.
5166 TARGET_X32_P: New Macro.
5167 TARGET_FPMATH_DEFAULT_P: New Macro.
5168 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
5169 * config/i386/i386.c (ix86_option_override_internal): New parameters.
5171 Change ix86_tune_string to access opts->x_ix86_tune_string.
5172 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
5173 Change ix86_arch_string to access opts->x_ix86_arch_string.
5174 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
5175 Change ix86_pmode to access opts->x_ix86_pmode.
5176 Change ix86_abi to access opts->x_ix86_abi.
5177 Change ix86_cmodel to access opts->x_ix86_cmodel.
5178 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
5179 Change ix86_isa_flags_explicit to access
5180 opts->x_ix86_isa_flags_explicit.
5181 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
5182 Change ix86_regparm to access opts->x_ix86_regparm.
5183 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
5184 Change ix86_preferred_stack_boundary_arg to access
5185 opts->x_ix86_preferred_stack_boundary_arg.
5186 Change ix86_force_align_arg_pointer to access
5187 opts->x_ix86_force_align_arg_pointer.
5188 Change ix86_incoming_stack_boundar_arg to access
5189 opts->x_ix86_incoming_stack_boundar_arg.
5190 Change ix86_fpmath to access opts->x_ix86_fpmath.
5191 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
5192 Change ix86_recip_name to access opts->x_ix86_recip_name.
5193 Change ix86_stack_protector_guard to access
5194 opts->x_ix86_stack_protector_guard.
5195 Change ix86_tune_memcpy_strategy to access
5196 opts->x_ix86_tune_memcpy_strategy.
5197 Change ix86_tune_memset_strategy to access
5198 opts->x_ix86_tune_memset_strategy.
5199 Change global_options to access opts.
5200 Change global_options_set to access opts_set.
5201 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
5202 Change TARGET_MMX to TARGET_MMX_P (opts->...).
5203 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
5204 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
5205 Change TARGET_SSE to TARGET_SSE_P (opts->...).
5206 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
5207 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
5208 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
5209 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
5210 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
5211 Change TARGET_AVX to TARGET_AVX_P (opts->...).
5212 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
5213 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
5214 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
5215 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
5216 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
5217 Change TARGET_FMA to TARGET_FMA_P (opts->...).
5218 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
5219 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
5220 Change TARGET_XOP to TARGET_XOP_P (opts->...).
5221 Change TARGET_LWP to TARGET_LWP_P (opts->...).
5222 Change TARGET_ABM to TARGET_ABM_P (opts->...).
5223 Change TARGET_BMI to TARGET_BMI_P (opts->...).
5224 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
5225 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
5226 Change TARGET_TBM to TARGET_TBM_P (opts->...).
5227 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
5228 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
5229 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
5230 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
5231 Change TARGET_AES to TARGET_AES_P (opts->...).
5232 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
5233 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
5234 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
5235 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
5236 Change TARGET_F16C to TARGET_F16C_P (opts->...).
5237 Change TARGET_RTM to TARGET_RTM_P (opts->...).
5238 Change TARGET_HLE to TARGET_HLE_P (opts->...).
5239 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
5240 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
5241 Change TARGET_ADX to TARGET_ADX_P (opts->...).
5242 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
5243 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
5244 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
5245 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
5246 Change TARGET_X32 to TARGET_X32_P (opts->...).
5247 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
5248 Change TARGET_FLOAT_RETURNS_IN_80387 to
5249 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
5250 (ix86_function_specific_save): New parameter. Use opts-> fields
5251 to replace global fields.
5252 (ix86_function_specific_restore): Ditto.
5253 (ix86_valid_target_attribute_inner_p): New parameters.
5255 Fix call to ix86_handle_option and set_option.
5256 (ix86_valid_target_attribute_tree): New parameters.
5257 Change global_options to access opts.
5258 Change global_options_set to access opts_set.
5259 Fix call to ix86_valid_target_attribute_inner_p.
5260 Change ix86_tune_string to access opts->x_ix86_tune_string.
5261 Change ix86_arch_string to access opts->x_ix86_arch_string.
5262 Change ix86_fpmath to access opts->x_ix86_fpmath
5263 Fix call to ix86_option_override_internal.
5264 Fix call to ix86_add_new_builtins.
5265 Fix calls to build_optimization_node and build_target_option_node.
5266 (ix86_valid_target_attribute_p): Remove access to global_options.
5267 Use new gcc_options structure func_options.
5268 Fix call to ix86_valid_target_attribute_tree.
5269 Fix call to build_optimization_node.
5270 (get_builtin_code_for_version): Fix call to
5271 ix86_valid_target_attribute_tree.
5273 2013-10-15 David Malcolm <dmalcolm@redhat.com>
5275 * Makefile.in (PICFLAG): New.
5276 (enable_host_shared): New.
5277 (INTERNAL_CFLAGS): Use PICFLAG.
5278 (LIBIBERTY): Use pic build of libiberty.a if configured with
5279 --enable-host-shared.
5280 * configure.ac: Add --enable-host-shared, setting up new
5282 * configure: Regenerate.
5283 * doc/install.texi (--enable-shared): Add note contrasting it with ...
5284 (--enable-host-shared): New option.
5286 2013-10-15 Richard Biener <rguenther@suse.de>
5288 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
5289 for built-in functions.
5291 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
5293 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
5294 (optimize_range_tests_1): New function,
5295 extracted from optimize_range_tests.
5296 (optimize_range_tests_xor): Similarly.
5297 (optimize_range_tests_diff): New function.
5298 (optimize_range_tests): Use optimize_range_tests_1.
5300 2013-10-15 Cong Hou <congh@google.com>
5302 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
5303 requirement of the reduction pattern so that one operand of the
5304 reduction operation can come from outside of the loop.
5306 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5308 * config/arm/neon-schedgen.ml: Remove.
5309 * config/arm/cortex-a9-neon.md: Remove comment regarding
5312 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5314 * config/arm/types: Remove old neon types.
5316 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5318 * config/arm/cortex-a7.md
5319 (cortex_a7_neon_type): New.
5320 (cortex_a7_neon_mul): Update for new types.
5321 (cortex_a7_neon_mla): Likewise.
5322 (cortex_a7_neon): Likewise.
5324 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5326 * config/arm/cortex-a15-neon.md
5327 (cortex_a15_neon_type): New,
5329 (cortex_a15_neon_int_1): Remove.
5330 (cortex_a15_neon_int_2): Likewise.
5331 (cortex_a15_neon_int_3): Likewise.
5332 (cortex_a15_neon_int_4): Likewise.
5333 (cortex_a15_neon_int_5): Likewise.
5334 (cortex_a15_neon_vqneg_vqabs): Likewise.
5335 (cortex_a15_neon_vmov): Likewise.
5336 (cortex_a15_neon_vaba): Likewise.
5337 (cortex_a15_neon_vaba_qqq): Likewise.
5338 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5339 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5340 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
5342 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5343 (cortex_a15_neon_mla_qqq_8_16): Likewise.
5344 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
5345 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5346 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5347 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
5348 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5349 (cortex_a15_neon_shift_1): Likewise.
5350 (cortex_a15_neon_shift_2): Likewise.
5351 (cortex_a15_neon_shift_3): Likewise.
5352 (cortex_a15_neon_vshl_ddd): Likewise.
5353 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5354 (cortex_a15_neon_vsra_vrsra): Likewise.
5355 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5356 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5357 (cortex_a15_neon_bp_3cycle): Likewise.
5358 (cortex_a15_neon_ldm_2): Likewise.
5359 (cortex_a15_neon_stm_2): Likewise.
5360 (cortex_a15_neon_mcr): Likewise.
5361 (cortex_a15_neon_mrc): Likewise.
5362 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
5363 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
5364 (cortex_a15_neon_fp_vmul_ddd): Likewise.
5365 (cortex_a15_neon_fp_vmul_qqd): Likewise.
5366 (cortex_a15_neon_fp_vmla_ddd): Likewise.
5367 (cortex_a15_neon_fp_vmla_qqq): Likewise.
5368 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5369 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5370 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5371 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5372 (cortex_a15_neon_bp_simple): Likewise.
5373 (cortex_a15_neon_bp_2cycle): Likewise.
5374 (cortex_a15_neon_bp_3cycle): Likewise.
5375 (cortex_a15_neon_vld1_1_2_regs): Likewise.
5376 (cortex_a15_neon_vld1_3_4_regs): Likewise.
5377 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5378 (cortex_a15_neon_vld2_4_regs): Likewise.
5379 (cortex_a15_neon_vld3_vld4): Likewise.
5380 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5381 (cortex_a15_neon_vst1_3_4_regs): Likewise.
5382 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
5383 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
5384 (cortex_a15_neon_vst3_vst4): Rename to...
5385 (cortex_a15_neon_vst4): This, update for new attributes.
5386 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
5387 (cortex_a15_neon_vld3_vld4_lane): Likewise.
5388 (cortex_a15_neon_vst1_vst2_lane): Likewise.
5389 (cortex_a15_neon_vst3_vst4_lane): Likewise.
5390 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
5391 (cortex_a15_neon_ldm_2): Likewise.
5392 (cortex_a15_neon_stm_2): Likewise.
5393 (cortex_a15_neon_mcr): Likewise.
5394 (cortex_a15_neon_mcr_2_mcrr): Likewise.
5395 (cortex_a15_neon_mrc): Likewise.
5396 (cortex_a15_neon_mrrc): Likewise.
5398 (cortex_a15_neon_abd): New.
5399 (cortex_a15_neon_abd_q): Likewise.
5400 (cortex_a15_neon_aba): Likewise.
5401 (cortex_a15_neon_aba_q): Likewise.
5402 (cortex_a15_neon_acc): Likewise.
5403 (cortex_a15_neon_acc_q): Likewise.
5404 (cortex_a15_neon_arith_basic): Likewise.
5405 (cortex_a15_neon_arith_complex): Likewise.
5406 (cortex_a15_neon_multiply): Likewise.
5407 (cortex_a15_neon_multiply_q): Likewise.
5408 (cortex_a15_neon_mla): Likewise.
5409 (cortex_a15_neon_mla_q): Likewise.
5410 (cortex_a15_neon_sat_mla_long): Likewise.
5411 (cortex_a15_neon_shift_acc): Likewise.
5412 (cortex_a15_neon_shift_imm_basic): Likewise.
5413 (cortex_a15_neon_shift_imm_complex): Likewise.
5414 (cortex_a15_neon_shift_reg_basic): Likewise.
5415 (cortex_a15_neon_shift_reg_basic_q): Likewise.
5416 (cortex_a15_neon_shift_reg_complex): Likewise.
5417 (cortex_a15_neon_shift_reg_complex_q): Likewise.
5418 (cortex_a15_neon_fp_negabs): Likewise
5419 (cortex_a15_neon_fp_arith): Likewise
5420 (cortex_a15_neon_fp_arith_q): Likewise
5421 (cortex_a15_neon_fp_cvt_int): Likewise
5422 (cortex_a15_neon_fp_cvt_int_q): Likewise
5423 (cortex_a15_neon_fp_cvt_16): Likewise
5424 (cortex_a15_neon_fp_mul): Likewise
5425 (cortex_a15_neon_fp_mul_q): Likewise
5426 (cortex_a15_neon_fp_mla): Likewise
5427 (cortex_a15_neon_fp_mla_q): Likewise
5428 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
5429 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
5430 (cortex_a15_neon_bitops): Likewise.
5431 (cortex_a15_neon_bitops_q): Likewise.
5432 (cortex_a15_neon_from_gp): Likewise.
5433 (cortex_a15_neon_from_gp_q): Likewise.
5434 (cortex_a15_neon_tbl3_tbl4): Likewise.
5435 (cortex_a15_neon_zip_q): Likewise.
5436 (cortex_a15_neon_to_gp): Likewise.
5437 (cortex_a15_neon_load_a): Likewise.
5438 (cortex_a15_neon_load_b): Likewise.
5439 (cortex_a15_neon_load_c): Likewise.
5440 (cortex_a15_neon_load_d): Likewise.
5441 (cortex_a15_neon_load_e): Likewise.
5442 (cortex_a15_neon_load_f): Likewise.
5443 (cortex_a15_neon_store_a): Likewise.
5444 (cortex_a15_neon_store_b): Likewise.
5445 (cortex_a15_neon_store_c): Likewise.
5446 (cortex_a15_neon_store_d): Likewise.
5447 (cortex_a15_neon_store_e): Likewise.
5448 (cortex_a15_neon_store_f): Likewise.
5449 (cortex_a15_neon_store_g): Likewise.
5450 (cortex_a15_neon_store_h): Likewise.
5451 (cortex_a15_vfp_to_from_gp): Likewise.
5453 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5455 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
5457 (cortex_a9_neon_vshl_ddd): Remove.
5458 (cortex_a9_neon_vst3_vst4): Likewise.
5459 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5461 (cortex_a9_neon_bit_ops_q): New.
5463 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
5464 (cortex_a9_neon_int_2): Likewise.
5465 (cortex_a9_neon_int_3): Likewise.
5466 (cortex_a9_neon_int_4): Likewise.
5467 (cortex_a9_neon_int_5): Likewise.
5468 (cortex_a9_neon_vqneg_vqabs): Likewise.
5469 (cortex_a9_neon_vmov): Likewise.
5470 (cortex_a9_neon_vaba): Likewise.
5471 (cortex_a9_neon_vaba_qqq): Likewise.
5472 (cortex_a9_neon_shift_1): Likewise.
5473 (cortex_a9_neon_shift_2): Likewise.
5474 (cortex_a9_neon_shift_3): Likewise.
5475 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5476 (cortex_a9_neon_vsra_vrsra): Likewise.
5477 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5478 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5479 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
5481 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5482 (cortex_a9_neon_mla_qqq_8_16): Likewise.
5483 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
5485 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5486 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5487 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
5488 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5489 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
5490 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
5491 (cortex_a9_neon_fp_vsum): Likewise.
5492 (cortex_a9_neon_fp_vmul_ddd): Likewise.
5493 (cortex_a9_neon_fp_vmul_qqd): Likewise.
5494 (cortex_a9_neon_fp_vmla_ddd): Likewise.
5495 (cortex_a9_neon_fp_vmla_qqq): Likewise.
5496 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
5497 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
5498 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5499 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5500 (cortex_a9_neon_bp_simple): Likewise.
5501 (cortex_a9_neon_bp_2cycle): Likewise.
5502 (cortex_a9_neon_bp_3cycle): Likewise.
5503 (cortex_a9_neon_ldr): Likewise.
5504 (cortex_a9_neon_str): Likewise.
5505 (cortex_a9_neon_vld1_1_2_regs): Likewise.
5506 (cortex_a9_neon_vld1_3_4_regs): Likewise.
5507 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5508 (cortex_a9_neon_vld2_4_regs): Likewise.
5509 (cortex_a9_neon_vld3_vld4): Likewise.
5510 (cortex_a9_neon_vld1_vld2_lane): Likewise.
5511 (cortex_a9_neon_vld3_vld4_lane): Likewise.
5512 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5513 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5514 (cortex_a9_neon_vst1_3_4_regs): Likewise.
5515 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
5516 (cortex_a9_neon_vst1_vst2_lane): Likewise.
5517 (cortex_a9_neon_vst3_vst4_lane): Likewise.
5518 (cortex_a9_neon_mcr): Likewise.
5519 (cortex_a9_neon_mcr_2_mcrr): Likewise.
5520 (cortex_a9_neon_mrc): Likewise.
5521 (cortex_a9_neon_mrrc): Likewise.
5523 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5525 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
5527 (cortex_a8_neon_vshl_ddd): Remove.
5528 (cortex_a8_neon_vst3_vst4): Likewise.
5529 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
5531 (cortex_a8_neon_bit_ops_q): New.
5533 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
5534 (cortex_a8_neon_int_2): Likewise..
5535 (cortex_a8_neon_int_3): Likewise.
5536 (cortex_a8_neon_int_5): Likewise.
5537 (cortex_a8_neon_vqneg_vqabs): Likewise.
5538 (cortex_a8_neon_int_4): Likewise.
5539 (cortex_a8_neon_vaba): Likewise.
5540 (cortex_a8_neon_vaba_qqq): Likewise.
5541 (cortex_a8_neon_shift_1): Likewise.
5542 (cortex_a8_neon_shift_2): Likewise.
5543 (cortex_a8_neon_shift_3): Likewise.
5544 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5545 (cortex_a8_neon_vsra_vrsra): Likewise.
5546 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5547 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5548 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
5550 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5551 (cortex_a8_neon_mla_qqq_8_16): Likewise.
5552 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
5554 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5555 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5556 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
5557 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5558 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
5559 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
5560 (cortex_a8_neon_fp_vsum): Likewise.
5561 (cortex_a8_neon_fp_vmul_ddd): Likewise.
5562 (cortex_a8_neon_fp_vmul_qqd): Likewise.
5563 (cortex_a8_neon_fp_vmla_ddd): Likewise.
5564 (cortex_a8_neon_fp_vmla_qqq): Likewise.
5565 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
5566 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
5567 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5568 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5569 (cortex_a8_neon_bp_simple): Likewise.
5570 (cortex_a8_neon_bp_2cycle): Likewise.
5571 (cortex_a8_neon_bp_3cycle): Likewise.
5572 (cortex_a8_neon_ldr): Likewise.
5573 (cortex_a8_neon_str): Likewise.
5574 (cortex_a8_neon_vld1_1_2_regs): Likewise.
5575 (cortex_a8_neon_vld1_3_4_regs): Likewise.
5576 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5577 (cortex_a8_neon_vld2_4_regs): Likewise.
5578 (cortex_a8_neon_vld3_vld4): Likewise.
5579 (cortex_a8_neon_vld1_vld2_lane): Likewise.
5580 (cortex_a8_neon_vld3_vld4_lane): Likewise.
5581 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5582 (cortex_a8_neon_vst1_3_4_regs): Likewise.
5583 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
5584 (cortex_a8_neon_vst1_vst2_lane): Likewise.
5585 (cortex_a8_neon_vst3_vst4_lane): Likewise.
5586 (cortex_a8_neon_mcr): Likewise.
5587 (cortex_a8_neon_mcr_2_mcrr): Likewise.
5588 (cortex_a8_neon_mrc): Likewise.
5589 (cortex_a8_neon_mrrc): Likewise.
5591 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5593 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
5595 * config/aarch64/aarch64-simd.md (neon_type): Remove.
5596 (aarch64_simd_dup<mode>): Add "type" attribute.
5597 (aarch64_dup_lane<mode>): Likewise.
5598 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
5599 (*aarch64_simd_mov<mode>): Likewise.
5600 (aarch64_simd_mov_from_<mode>low): Likewise.
5601 (aarch64_simd_mov_from_<mode>high): Likewise.
5602 (orn<mode>3): Likewise.
5603 (bic<mode>3): Likewise.
5604 (add<mode>3): Likewise.
5605 (sub<mode>3): Likewise.
5606 (mul<mode>3): Likewise.
5607 (*aarch64_mul3_elt<mode>): Likewise.
5608 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5609 (*aarch64_mul3_elt_to_128df): Likewise.
5610 (*aarch64_mul3_elt_to_64v2df): Likewise.
5611 (neg<mode>2): Likewise.
5612 (abs<mode>2): Likewise.
5613 (abd<mode>_3): Likewise.
5614 (aba<mode>_3): Likewise.
5615 (fabd<mode>_3): Likewise.
5616 (*fabd_scalar<mode>3): Likewise.
5617 (and<mode>3): Likewise.
5618 (ior<mode>3): Likewise.
5619 (xor<mode>3): Likewise.
5620 (one_cmpl<mode>2): Likewise.
5621 (aarch64_simd_vec_set<mode>): Likewise.
5622 (aarch64_simd_lshr<mode>): Likewise.
5623 (aarch64_simd_ashr<mode>): Likewise.
5624 (aarch64_simd_imm_shl<mode>): Likewise.
5625 (aarch64_simd_reg_sshl<mode): Likewise.
5626 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
5627 (aarch64_simd_reg_shl<mode>_signed): Likewise.
5628 (aarch64_simd_vec_setv2di): Likewise.
5629 (aarch64_simd_vec_set<mode>): Likewise.
5630 (aarch64_mla<mode>): Likewise.
5631 (*aarch64_mla_elt<mode>): Likewise.
5632 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5633 (aarch64_mls<mode>): Likewise.
5634 (*aarch64_mls_elt<mode>): Likewise.
5635 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5636 (<su><maxmin><mode>3): Likewise.
5637 (move_lo_quad_<mode>): Likewise.
5638 (aarch64_simd_move_hi_quad_<mode>): Likewise.
5639 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
5640 (vec_pack_trunc_<mode>): Likewise.
5641 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
5642 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
5643 (*aarch64_<su>mlal_lo<mode>): Likewise.
5644 (*aarch64_<su>mlal_hi<mode>): Likewise.
5645 (*aarch64_<su>mlsl_lo<mode>): Likewise.
5646 (*aarch64_<su>mlsl_hi<mode>): Likewise.
5647 (*aarch64_<su>mlal<mode>): Likewise.
5648 (*aarch64_<su>mlsl<mode>): Likewise.
5649 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
5650 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
5651 (add<mode>3): Likewise.
5652 (sub<mode>3): Likewise.
5653 (mul<mode>3): Likewise.
5654 (div<mode>3): Likewise.
5655 (neg<mode>2): Likewise.
5656 (abs<mode>2): Likewise.
5657 (fma<mode>4): Likewise.
5658 (*aarch64_fma4_elt<mode>): Likewise.
5659 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5660 (*aarch64_fma4_elt_to_128df): Likewise.
5661 (*aarch64_fma4_elt_to_64v2df): Likewise.
5662 (fnma<mode>4): Likewise.
5663 (*aarch64_fnma4_elt<mode>): Likewise.
5664 (*aarch64_fnma4_elt_<vswap_width_name><mode>
5665 (*aarch64_fnma4_elt_to_128df): Likewise.
5666 (*aarch64_fnma4_elt_to_64v2df): Likewise.
5667 (<frint_pattern><mode>2): Likewise.
5668 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
5669 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
5670 (vec_unpacks_lo_v4sf): Likewise.
5671 (aarch64_float_extend_lo_v2df): Likewise.
5672 (vec_unpacks_hi_v4sf): Likewise.
5673 (aarch64_float_truncate_lo_v2sf): Likewise.
5674 (aarch64_float_truncate_hi_v4sf): Likewise.
5675 (aarch64_vmls<mode>): Likewise.
5676 (<su><maxmin><mode>3): Likewise.
5677 (<maxmin_uns><mode>3): Likewise.
5678 (reduc_<sur>plus_<mode>): Likewise.
5679 (reduc_<sur>plus_v2di): Likewise.
5680 (reduc_<sur>plus_v2si): Likewise.
5681 (reduc_<sur>plus_<mode>): Likewise.
5682 (aarch64_addpv4sf): Likewise.
5683 (clz<mode>2): Likewise.
5684 (reduc_<maxmin_uns>_<mode>): Likewise.
5685 (reduc_<maxmin_uns>_v2di): Likewise.
5686 (reduc_<maxmin_uns>_v2si): Likewise.
5687 (reduc_<maxmin_uns>_<mode>): Likewise.
5688 (reduc_<maxmin_uns>_v4sf): Likewise.
5689 (aarch64_simd_bsl<mode>_internal): Likewise.
5690 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
5691 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
5692 (aarch64_get_lane<mode>): Likewise.
5693 (*aarch64_combinez<mode>): Likewise.
5694 (aarch64_combine<mode>): Likewise.
5695 (aarch64_simd_combine<mode>): Likewise.
5696 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
5697 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
5698 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
5699 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
5700 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
5701 (aarch64_<sur>h<addsub><mode>): Likewise.
5702 (aarch64_<sur><addsub>hn<mode>): Likewise.
5703 (aarch64_<sur><addsub>hn2<mode>): Likewise.
5704 (aarch64_pmul<mode>): Likewise.
5705 (aarch64_<su_optab><optab><mode>): Likewise.
5706 (aarch64_<sur>qadd<mode>): Likewise.
5707 (aarch64_sqmovun<mode>): Likewise.
5708 (aarch64_<sur>qmovn<mode>): Likewise.
5709 (aarch64_s<optab><mode>): Likewise.
5710 (aarch64_sq<r>dmulh<mode>): Likewise.
5711 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
5712 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
5713 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
5714 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
5715 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
5716 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
5717 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
5718 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
5719 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
5720 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
5721 (aarch64_sqdmull<mode>): Likewise.
5722 (aarch64_sqdmull_lane<mode>_internal): Likewise.
5723 (aarch64_sqdmull_n<mode>): Likewise.
5724 (aarch64_sqdmull2<mode>_internal): Likewise.
5725 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
5726 (aarch64_sqdmull2_n<mode>_internal): Likewise.
5727 (aarch64_<sur>shl<mode>): Likewise.
5728 (aarch64_<sur>q<r>shl<mode>
5729 (aarch64_<sur>shll_n<mode>): Likewise.
5730 (aarch64_<sur>shll2_n<mode>): Likewise.
5731 (aarch64_<sur>shr_n<mode>): Likewise.
5732 (aarch64_<sur>sra_n<mode>): Likewise.
5733 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
5734 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
5735 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
5736 (aarch64_cm<optab><mode>): Likewise.
5737 (aarch64_cm<optab>di): Likewise.
5738 (aarch64_cm<optab><mode>): Likewise.
5739 (aarch64_cm<optab>di): Likewise.
5740 (aarch64_cmtst<mode>): Likewise.
5741 (aarch64_cmtstdi): Likewise.
5742 (aarch64_cm<optab><mode>): Likewise.
5743 (*aarch64_fac<optab><mode>): Likewise.
5744 (aarch64_addp<mode>): Likewise.
5745 (aarch64_addpdi): Likewise.
5746 (sqrt<mode>2): Likewise.
5747 (vec_load_lanesoi<mode>): Likewise.
5748 (vec_store_lanesoi<mode>): Likewise.
5749 (vec_load_lanesci<mode>): Likewise.
5750 (vec_store_lanesci<mode>): Likewise.
5751 (vec_load_lanesxi<mode>): Likewise.
5752 (vec_store_lanesxi<mode>): Likewise.
5753 (*aarch64_mov<mode>): Likewise.
5754 (aarch64_ld2<mode>_dreg): Likewise.
5755 (aarch64_ld2<mode>_dreg): Likewise.
5756 (aarch64_ld3<mode>_dreg): Likewise.
5757 (aarch64_ld3<mode>_dreg): Likewise.
5758 (aarch64_ld4<mode>_dreg): Likewise.
5759 (aarch64_ld4<mode>_dreg): Likewise.
5760 (aarch64_tbl1<mode>): Likewise.
5761 (aarch64_tbl2v16qi): Likewise.
5762 (aarch64_combinev16qi): Likewise.
5763 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
5764 (aarch64_st2<mode>_dreg): Likewise.
5765 (aarch64_st2<mode>_dreg): Likewise.
5766 (aarch64_st3<mode>_dreg): Likewise.
5767 (aarch64_st3<mode>_dreg): Likewise.
5768 (aarch64_st4<mode>_dreg): Likewise.
5769 (aarch64_st4<mode>_dreg): Likewise.
5770 (*aarch64_simd_ld1r<mode>): Likewise.
5771 (aarch64_frecpe<mode>): Likewise.
5772 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
5773 (aarch64_frecps<mode>): Likewise.
5775 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5777 * config/arm/iterators.md (V_elem_ch): New.
5779 (VQH_type): Likewise.
5780 * config/arm/arm.md (is_neon_type): New.
5781 (conds): Use is_neon_type.
5782 (anddi3_insn): Update type attribute.
5783 (xordi3_insn): Likewise.
5784 (one_cmpldi2): Likewise.
5785 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
5786 * config/arm/neon.md (neon_mov): Update type attribute.
5787 (*movmisalign<mode>_neon_store): Likewise.
5788 (*movmisalign<mode>_neon_load): Likewise.
5789 (vec_set<mode>_internal): Likewise.
5790 (vec_set<mode>_internal): Likewise.
5791 (vec_setv2di_internal): Likewise.
5792 (vec_extract<mode>): Likewise.
5793 (vec_extract<mode>): Likewise.
5794 (vec_extractv2di): Likewise.
5795 (*add<mode>3_neon): Likewise.
5796 (adddi3_neon): Likewise.
5797 (*sub<mode>3_neon): Likewise.
5798 (subdi3_neon): Likewise.
5799 (fma<VCVTF:mode>4): Likewise.
5800 (fma<VCVTF:mode>4_intrinsic): Likewise.
5801 (*fmsub<VCVTF:mode>4): Likewise.
5802 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
5803 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
5804 (ior<mode>3): Likewise.
5805 (and<mode>3): Likewise.
5806 (orn<mode>3_neon): Likewise.
5807 (orndi3_neon): Likewise.
5808 (bic<mode>3_neon): Likewise.
5809 (bicdi3_neon): Likewise.
5810 (xor<mode>3): Likewise.
5811 (one_cmpl<mode>2): Likewise.
5812 (abs<mode>2): Likewise.
5813 (neg<mode>2): Likewise.
5814 (negdi2_neon): Likewise.
5815 (*umin<mode>3_neon): Likewise.
5816 (*umax<mode>3_neon): Likewise.
5817 (*smin<mode>3_neon): Likewise.
5818 (*smax<mode>3_neon): Likewise.
5819 (vashl<mode>3): Likewise.
5820 (vashr<mode>3_imm): Likewise.
5821 (vlshr<mode>3_imm): Likewise.
5822 (ashl<mode>3_signed): Likewise.
5823 (ashl<mode>3_unsigned): Likewise.
5824 (neon_load_count): Likewise.
5825 (ashldi3_neon_noclobber): Likewise.
5826 (ashldi3_neon): Likewise.
5827 (signed_shift_di3_neon): Likewise.
5828 (unsigned_shift_di3_neon): Likewise.
5829 (ashrdi3_neon_imm_noclobber): Likewise.
5830 (lshrdi3_neon_imm_noclobber): Likewise.
5831 (<shift>di3_neon): Likewise.
5832 (widen_ssum<mode>3): Likewise.
5833 (widen_usum<mode>3): Likewise.
5834 (quad_halves_<code>v4si): Likewise.
5835 (quad_halves_<code>v4sf): Likewise.
5836 (quad_halves_<code>v8hi): Likewise.
5837 (quad_halves_<code>v16qi): Likewise.
5838 (reduc_splus_v2di): Likewise.
5839 (neon_vpadd_internal<mode>): Likewise.
5840 (neon_vpsmin<mode>): Likewise.
5841 (neon_vpsmax<mode>): Likewise.
5842 (neon_vpumin<mode>): Likewise.
5843 (neon_vpumax<mode>): Likewise.
5844 (*ss_add<mode>_neon): Likewise.
5845 (*us_add<mode>_neon): Likewise.
5846 (*ss_sub<mode>_neon): Likewise.
5847 (*us_sub<mode>_neon): Likewise.
5848 (neon_vadd<mode>_unspec): Likewise.
5849 (neon_vaddl<mode>): Likewise.
5850 (neon_vaddw<mode>): Likewise.
5851 (neon_vhadd<mode>): Likewise.
5852 (neon_vqadd<mode>): Likewise.
5853 (neon_vaddhn<mode>): Likewise.
5854 (neon_vmul<mode>): Likewise.
5855 (neon_vfms<VCVTF:mode>): Likewise.
5856 (neon_vmlal<mode>): Likewise.
5857 (neon_vmls<mode>): Likewise.
5858 (neon_vmlsl<mode>): Likewise.
5859 (neon_vqdmulh<mode>): Likewise.
5860 (neon_vqdmlal<mode>): Likewise.
5861 (neon_vqdmlsl<mode>): Likewise.
5862 (neon_vmull<mode>): Likewise.
5863 (neon_vqdmull<mode>): Likewise.
5864 (neon_vsub<mode>_unspec): Likewise.
5865 (neon_vsubl<mode>): Likewise.
5866 (neon_vsubw<mode>): Likewise.
5867 (neon_vqsub<mode>): Likewise.
5868 (neon_vhsub<mode>): Likewise.
5869 (neon_vsubhn<mode>): Likewise.
5870 (neon_vceq<mode>): Likewise.
5871 (neon_vcge<mode>): Likewise.
5872 (neon_vcgeu<mode>): Likewise.
5873 (neon_vcgt<mode>): Likewise.
5874 (neon_vcgtu<mode>): Likewise.
5875 (neon_vcle<mode>): Likewise.
5876 (neon_vclt<mode>): Likewise.
5877 (neon_vcage<mode>): Likewise.
5878 (neon_vcagt<mode>): Likewise.
5879 (neon_vtst<mode>): Likewise.
5880 (neon_vabd<mode>): Likewise.
5881 (neon_vabdl<mode>): Likewise.
5882 (neon_vaba<mode>): Likewise.
5883 (neon_vabal<mode>): Likewise.
5884 (neon_vmax<mode>): Likewise.
5885 (neon_vmin<mode>): Likewise.
5886 (neon_vpaddl<mode>): Likewise.
5887 (neon_vpadal<mode>): Likewise.
5888 (neon_vpmax<mode>): Likewise.
5889 (neon_vpmin<mode>): Likewise.
5890 (neon_vrecps<mode>): Likewise.
5891 (neon_vrsqrts<mode>): Likewise.
5892 (neon_vqabs<mode>): Likewise.
5893 (neon_vqneg<mode>): Likewise.
5894 (neon_vcls<mode>): Likewise.
5895 (clz<mode>2): Likewise.
5896 (popcount<mode>2): Likewise.
5897 (neon_vrecpe<mode>): Likewise.
5898 (neon_vrsqrte<mode>): Likewise.
5899 (neon_vget_lane<mode>_sext_internal): Likewise.
5900 (neon_vget_lane<mode>_zext_internal): Likewise.
5901 (neon_vdup_n<mode>): Likewise.
5902 (neon_vdup_n<mode>): Likewise.
5903 (neon_vdup_nv2di): Likewise.
5904 (neon_vdup_lane<mode>_interal): Likewise.
5905 (*neon_vswp<mode>): Likewise.
5906 (neon_vcombine<mode>): Likewise.
5907 (float<mode><V_cvtto>2): Likewise.
5908 (floatuns<mode><V_cvtto>2): Likewise.
5909 (fix_trunc<mode><V_cvtto>2): Likewise.
5910 (fixuns_trunc<mode><V_cvtto>2
5911 (neon_vcvt<mode>): Likewise.
5912 (neon_vcvt<mode>): Likewise.
5913 (neon_vcvtv4sfv4hf): Likewise.
5914 (neon_vcvtv4hfv4sf): Likewise.
5915 (neon_vcvt_n<mode>): Likewise.
5916 (neon_vcvt_n<mode>): Likewise.
5917 (neon_vmovn<mode>): Likewise.
5918 (neon_vqmovn<mode>): Likewise.
5919 (neon_vqmovun<mode>): Likewise.
5920 (neon_vmovl<mode>): Likewise.
5921 (neon_vmul_lane<mode>): Likewise.
5922 (neon_vmul_lane<mode>): Likewise.
5923 (neon_vmull_lane<mode>): Likewise.
5924 (neon_vqdmull_lane<mode>): Likewise.
5925 (neon_vqdmulh_lane<mode>): Likewise.
5926 (neon_vqdmulh_lane<mode>): Likewise.
5927 (neon_vmla_lane<mode>): Likewise.
5928 (neon_vmla_lane<mode>): Likewise.
5929 (neon_vmlal_lane<mode>): Likewise.
5930 (neon_vqdmlal_lane<mode>): Likewise.
5931 (neon_vmls_lane<mode>): Likewise.
5932 (neon_vmls_lane<mode>): Likewise.
5933 (neon_vmlsl_lane<mode>): Likewise.
5934 (neon_vqdmlsl_lane<mode>): Likewise.
5935 (neon_vext<mode>): Likewise.
5936 (neon_vrev64<mode>): Likewise.
5937 (neon_vrev32<mode>): Likewise.
5938 (neon_vrev16<mode>): Likewise.
5939 (neon_vbsl<mode>_internal): Likewise.
5940 (neon_vshl<mode>): Likewise.
5941 (neon_vqshl<mode>): Likewise.
5942 (neon_vshr_n<mode>): Likewise.
5943 (neon_vshrn_n<mode>): Likewise.
5944 (neon_vqshrn_n<mode>): Likewise.
5945 (neon_vqshrun_n<mode>): Likewise.
5946 (neon_vshl_n<mode>): Likewise.
5947 (neon_vqshl_n<mode>): Likewise.
5948 (neon_vqshlu_n<mode>): Likewise.
5949 (neon_vshll_n<mode>): Likewise.
5950 (neon_vsra_n<mode>): Likewise.
5951 (neon_vsri_n<mode>): Likewise.
5952 (neon_vsli_n<mode>): Likewise.
5953 (neon_vtbl1v8qi): Likewise.
5954 (neon_vtbl2v8qi): Likewise.
5955 (neon_vtbl3v8qi): Likewise.
5956 (neon_vtbl4v8qi): Likewise.
5957 (neon_vtbl1v16qi): Likewise.
5958 (neon_vtbl2v16qi): Likewise.
5959 (neon_vcombinev16qi): Likewise.
5960 (neon_vtbx1v8qi): Likewise.
5961 (neon_vtbx2v8qi): Likewise.
5962 (neon_vtbx3v8qi): Likewise.
5963 (neon_vtbx4v8qi): Likewise.
5964 (*neon_vtrn<mode>_insn): Likewise.
5965 (*neon_vzip<mode>_insn): Likewise.
5966 (*neon_vuzp<mode>_insn): Likewise.
5967 (neon_vld1<mode>): Likewise.
5968 (neon_vld1_lane<mode>): Likewise.
5969 (neon_vld1_lane<mode>): Likewise.
5970 (neon_vld1_dup<mode>): Likewise.
5971 (neon_vld1_dup<mode>): Likewise.
5972 (neon_vld1_dupv2di): Likewise.
5973 (neon_vst1<mode>): Likewise.
5974 (neon_vst1_lane<mode>): Likewise.
5975 (neon_vst1_lane<mode>): Likewise.
5976 (neon_vld2<mode>): Likewise.
5977 (neon_vld2<mode>): Likewise.
5978 (neon_vld2_lane<mode>): Likewise.
5979 (neon_vld2_lane<mode>): Likewise.
5980 (neon_vld2_dup<mode>): Likewise.
5981 (neon_vst2<mode>): Likewise.
5982 (neon_vst2<mode>): Likewise.
5983 (neon_vst2_lane<mode>): Likewise.
5984 (neon_vst2_lane<mode>): Likewise.
5985 (neon_vld3<mode>): Likewise.
5986 (neon_vld3qa<mode>): Likewise.
5987 (neon_vld3qb<mode>): Likewise.
5988 (neon_vld3_lane<mode>): Likewise.
5989 (neon_vld3_lane<mode>): Likewise.
5990 (neon_vld3_dup<mode>): Likewise.
5991 (neon_vst3<mode>): Likewise.
5992 (neon_vst3qa<mode>): Likewise.
5993 (neon_vst3qb<mode>): Likewise.
5994 (neon_vst3_lane<mode>): Likewise.
5995 (neon_vst3_lane<mode>): Likewise.
5996 (neon_vld4<mode>): Likewise.
5997 (neon_vld4qa<mode>): Likewise.
5998 (neon_vld4qb<mode>): Likewise.
5999 (neon_vld4_lane<mode>): Likewise.
6000 (neon_vld4_lane<mode>): Likewise.
6001 (neon_vld4_dup<mode>): Likewise.
6002 (neon_vst4<mode>): Likewise.
6003 (neon_vst4qa<mode>): Likewise.
6004 (neon_vst4qb<mode>): Likewise.
6005 (neon_vst4_lane<mode>): Likewise.
6006 (neon_vst4_lane<mode>): Likewise.
6007 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6008 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6009 (neon_vec_<US>mult_lo_<mode>): Likewise.
6010 (neon_vec_<US>mult_hi_<mode>): Likewise.
6011 (neon_vec_<US>shiftl_<mode>): Likewise.
6012 (neon_unpack<US>_<mode>): Likewise.
6013 (neon_vec_<US>mult_<mode>): Likewise.
6014 (vec_pack_trunc_<mode>): Likewise.
6015 (neon_vec_pack_trunc_<mode>): Likewise.
6016 (neon_vabd<mode>_2): Likewise.
6017 (neon_vabd<mode>_3): Likewise.
6019 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6021 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
6022 (load_pair): Update type attribute.
6023 (store_pair): Update type attribute.
6024 * config/aarch64/iterators.md (q): New.
6026 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
6028 * config/arm/types.md: Add new types for Neon insns.
6030 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6031 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6032 Sergey Lega <sergey.s.lega@intel.com>
6033 Anna Tikhonova <anna.tikhonova@intel.com>
6034 Ilya Tocar <ilya.tocar@intel.com>
6035 Andrey Turetskiy <andrey.turetskiy@intel.com>
6036 Ilya Verbin <ilya.verbin@intel.com>
6037 Kirill Yukhin <kirill.yukhin@intel.com>
6038 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6040 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
6041 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
6042 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
6044 (srcp14<mode>): Ditto.
6045 (rsqrt14<mode>): Ditto.
6046 (rsqrt14<mode>): Ditto.
6047 (avx512f_vmscalef<mode>): Ditto.
6048 (avx512f_scalef<mode>): Ditto.
6049 (avx512f_getexp<mode>): Ditto.
6050 (avx512f_sgetexp<mode>): Ditto.
6051 (avx512f_fixupimm<mode>): Ditto.
6052 (avx512f_sfixupimm<mode>): Ditto.
6053 (avx512f_rndscale<mode>): Ditto.
6054 (*avx512er_exp2<mode>): Ditto.
6055 (*avx512er_rcp28<mode>): Ditto.
6056 (avx512er_rsqrt28<mode>): Ditto.
6057 (avx512f_getmant<mode>): Ditto.
6058 (avx512f_getmant<mode>): Ditto.
6059 (avx512f_rndscale<mode>): Fix formatting.
6061 2013-10-15 Martin Jambor <mjambor@suse.cz>
6063 * ipa-utils.h (ipa_edge_within_scc): Declare.
6064 * ipa-cp.c (edge_within_scc): Moved...
6065 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
6067 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6068 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6069 Sergey Lega <sergey.s.lega@intel.com>
6070 Anna Tikhonova <anna.tikhonova@intel.com>
6071 Ilya Tocar <ilya.tocar@intel.com>
6072 Andrey Turetskiy <andrey.turetskiy@intel.com>
6073 Ilya Verbin <ilya.verbin@intel.com>
6074 Kirill Yukhin <kirill.yukhin@intel.com>
6075 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6077 * config/i386/predicates.md (const_8_to_15_operand): New.
6078 (const_16_to_31_operand): Ditto.
6079 * config/i386/sse.md (V8FI): New.
6081 (reduc_splus_v8df): Ditto.
6082 (reduc_splus_v16sf): Ditto.
6083 (avx512f_vextract<shuffletype>32x4_1): Ditto.
6084 (vec_extract_hi_<mode>): Ditto.
6085 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
6086 (vec_set_lo_<mode>): Ditto.
6087 (vec_set_hi_<mode>): Ditto.
6088 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
6089 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
6090 (avx512f_pshufd_1): Ditto.
6091 (avx512f_broadcast<mode>): Ditto.
6092 (avx512f_broadcast<mode>): Ditto.
6093 (define_split): Split vec_extract_lo into move.
6094 (ssequartermode): Ditto.
6095 (ssedoublemode): Extened with wider modes.
6096 (vec_extract_lo_<mode>): Ditto.
6098 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6099 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6100 Sergey Lega <sergey.s.lega@intel.com>
6101 Anna Tikhonova <anna.tikhonova@intel.com>
6102 Ilya Tocar <ilya.tocar@intel.com>
6103 Andrey Turetskiy <andrey.turetskiy@intel.com>
6104 Ilya Verbin <ilya.verbin@intel.com>
6105 Kirill Yukhin <kirill.yukhin@intel.com>
6106 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6108 * config/i386/predicates.md (register_or_constm1_operand): New.
6109 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
6110 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
6111 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
6112 UNSPEC_SCATTER_PREFETCH
6114 (avx512f_ucmp<mode>3): Ditto.
6115 (avx512f_vternlog<mode>): Ditto.
6116 (avx512f_align<mode>): Ditto.
6117 (<shift_insn><mode>3): Ditto.
6118 (avx512f_<rotate>v<mode>): Ditto.
6119 (avx512f_<rotate><mode>): Ditto.
6120 (avx512f_eq<mode>3): Ditto.
6121 (avx512f_eq<mode>3_1): Ditto.
6122 (avx512f_gt<mode>3): Ditto.
6123 (avx512f_testm<mode>3): Ditto.
6124 (avx512f_testnm<mode>3): Ditto.
6125 (avx512pf_gatherpf<mode>): Ditto.
6126 (*avx512pf_gatherpf<mode>_mask): Ditto.
6127 (*avx512pf_gatherpf<mode>): Ditto.
6128 (avx512pf_scatterpf<mode>): Ditto.
6129 (*avx512pf_scatterpf<mode>_mask): Ditto.
6130 (*avx512pf_scatterpf<mode>): Ditto.
6131 (avx512f_vec_dup_gpr<mode>): Ditto.
6132 (clz<mode>2): Ditto.
6133 (conflict<mode>): Ditto.
6134 (REDUC_SMINMAX_MODE): Extened with wider modes.
6135 (reduc_<code>_<mode>): Ditto.
6136 (vlshr<mode>3): Ditto.
6137 (vashl<mode>3): Ditto.
6139 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
6140 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6141 Sergey Lega <sergey.s.lega@intel.com>
6142 Anna Tikhonova <anna.tikhonova@intel.com>
6143 Ilya Tocar <ilya.tocar@intel.com>
6144 Andrey Turetskiy <andrey.turetskiy@intel.com>
6145 Ilya Verbin <ilya.verbin@intel.com>
6146 Kirill Yukhin <kirill.yukhin@intel.com>
6147 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6149 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
6152 (avx512fmaskmode): Ditto.
6153 (bcstscalarsuff): Ditto.
6154 (avx512f_blendm<mode>): Ditto.
6155 (cmp_imm_predicate): Ditto.
6156 (avx512f_cmp<mode>3): Ditto.
6157 (avx512f_vec_dup<mode>): Ditto.
6158 (avx512f_vec_dup_mem<mode>): Ditto.
6159 (avx512f_vpermi2var<mode>3): Ditto.
6160 (avx512f_vpermt2var<mode>3): Ditto.
6161 (vec_init<mode>): Ditto.
6162 (avx512f_gathersi<mode>): Ditto.
6163 (*avx512f_gathersi<mode>): Ditto.
6164 (*avx512f_gathersi<mode>_2): Ditto.
6165 (avx512f_gatherdi<mode>): Ditto.
6166 (*avx512f_gatherdi<mode>): Ditto.
6167 (*avx512f_gatherdi<mode>_2): Ditto.
6168 (avx512f_scattersi<mode>): Ditto.
6169 (*avx512f_scattersi<mode>): Ditto.
6170 (avx512f_scatterdi<mode>): Ditto.
6171 (*avx512f_scatterdi<mode>): Ditto.
6172 (sseintprefix): Extened with wider modes.
6173 (VEC_GATHER_IDXSI): Ditto.
6174 (VEC_GATHER_IDXDI): Ditto.
6175 (VEC_GATHER_SRCDI): Ditto.
6177 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
6178 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6180 * config/arm/t-aprofile: New file.
6181 * config.gcc: Handle --with-multilib-list option.
6183 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
6185 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
6188 2013-10-15 Richard Biener <rguenther@suse.de>
6190 * tree-loop-distribution.c (build_empty_rdg): Inline into
6192 (rdg_flag_vertex): Inline into single user.
6193 (rdg_flag_vertex_and_dependent): Likewise.
6194 (build_rdg_partition_for_vertex): Remove processed bitmap.
6195 (rdg_build_partitions): Simplify.
6197 2013-10-15 Richard Biener <rguenther@suse.de>
6199 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
6200 Restructure forwarding through conversions and copies to
6201 avoid performing copy-propagation the wrong way. Adjust
6202 recursion invocations.
6203 (forward_propagate_addr_expr): Add argument stating if we
6204 are recursing from a single-use.
6205 (ssa_forward_propagate_and_combine): Adjust.
6207 2013-10-14 David Malcolm <dmalcolm@redhat.com>
6209 * dumpfile.h (gcc::dump_manager): New class, to hold state
6210 relating to dumpfile management.
6211 (get_dump_file_name): Remove in favor of method of dump_manager.
6212 (dump_initialized_p): Likewise.
6213 (dump_start): Likewise.
6214 (dump_finish): Likewise.
6215 (dump_switch_p): Likewise.
6216 (dump_register): Likewise.
6217 (get_dump_file_info): Likewise.
6218 * context.c (gcc::context::context): Construct the dump_manager
6220 * context.h (gcc::context::get_dumps): New.
6221 (gcc::context::m_dumps): New.
6222 * coverage.c (coverage_init): Port to dump_manager API.
6223 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
6224 (extra_dump_files_in_use): Likewise.
6225 (extra_dump_files_alloced): Likewise.
6226 (gcc::dump_manager::dump_manager): New.
6227 (dump_register): Convert to...
6228 (gcc::dump_manager::dump_register): ...method, replacing
6229 function-static next_dump with m_next_dump field.
6230 (get_dump_file_info): Convert to...
6231 (gcc::dump_manager::get_dump_file_info): ...method.
6232 (get_dump_file_name): Convert to...
6233 (gcc::dump_manager::get_dump_file_name): ...method.
6234 (dump_start): Convert to...
6235 (gcc::dump_manager::dump_start): ...method.
6236 (dump_finish): Convert to...
6237 (gcc::dump_manager::dump_finish): ...method.
6238 (dump_begin): Replace body with...
6239 (gcc::dump_manager::dump_begin): ...new method.
6240 (dump_phase_enabled_p): Convert to...
6241 (gcc::dump_manager::dump_phase_enabled_p): ...method.
6242 (dump_phase_enabled_p): Convert to...
6243 (gcc::dump_manager::dump_phase_enabled_p): ...method.
6244 (dump_initialized_p): Convert to...
6245 (gcc::dump_manager::dump_initialized_p): ...method.
6246 (dump_flag_name): Replace body with...
6247 (gcc::dump_manager::dump_flag_name): ...new method.
6248 (dump_enable_all): Convert to...
6249 (gcc::dump_manager::dump_enable_all): ...new method.
6250 (opt_info_enable_passes): Convert to...
6251 (gcc::dump_manager::opt_info_enable_passes): ...new method.
6252 (dump_switch_p_1): Convert to...
6253 (gcc::dump_manager::dump_switch_p_1): ...new method.
6254 (dump_switch_p): Convert to...
6255 (gcc::dump_manager::dump_switch_p): ...new method.
6256 (opt_info_switch_p): Port to dump_manager API.
6257 (enable_rtl_dump_file): Likewise.
6258 * opts-global.c (handle_common_deferred_options): Port to new
6260 * passes.c (pass_manager::finish_optimization_passes): Likewise.
6261 (pass_manager::register_one_dump_file): Likewise.
6262 (pass_manager::register_pass): Likewise.
6263 (pass_init_dump_file): Likewise.
6264 (pass_fini_dump_file): Likewise.
6265 * statistics.c (statistics_early_init): Likewise.
6267 2013-10-14 Richard Biener <rguenther@suse.de>
6269 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
6270 iterative_hash_canonical_type, gimple_canonical_type_hash,
6271 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
6272 gimple_register_canonical_type, print_gimple_types_stats,
6273 free_gimple_type_tables): Move to lto/lto.c
6274 (gt-gimple.h): Do not include.
6275 * gimple.h (gimple_register_canonical_type,
6276 print_gimple_types_stats, free_gimple_type_tables): Remove.
6277 * Makefile.in (GTFILES): Remove gimple.c.
6279 2013-10-14 Travis Snoozy <quandary@remstate.com>
6282 * config/msp430/msp430.c (msp430_option_override): Correct thinko
6283 scanning for msp430x targets.
6285 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
6288 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
6289 (registers_ok_for_ldd_peep): Move around.
6290 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
6291 * config/sparc/sparc.md (widening peepholes): Use it.
6293 2013-10-14 Richard Biener <rguenther@suse.de>
6297 * gimple.c (iterative_hash_canonical_type): Make sure to
6298 record the hash into the correct hashtable slot.
6300 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
6302 PR rtl-optimization/58662
6303 * combine.c (try_combine): Take into account death nodes on I2 when
6304 splitting a PARALLEL of two independent SETs. Fix dump message.
6306 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
6309 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
6310 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
6311 * config/sh/t-sh (sh_treg_combine.o): New entry.
6312 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
6313 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
6314 (register_sh_passes): New function. Register sh_treg_combine pass.
6315 (sh_option_override): Invoke it.
6316 (sh_canonicalize_comparison): Handle op0_preserve_value.
6317 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
6318 opportunities. Canonicalize branch condition.
6319 (nott): Allow only if pseudos can be created for non-SH2A.
6321 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
6324 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
6325 (ix86_expand_movmem): Replace copy_addr_to_reg with
6326 ix86_copy_addr_to_reg.
6327 (ix86_expand_setmem): Likewise.
6329 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
6331 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
6332 provided by ix86_fp_compare_mode instead of CCFPUmode.
6334 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
6336 * config/aarch64/arm_neon.h
6337 (vtbx<1,3>_<psu>8): Fix register constriants.
6339 2013-10-11 Jeff Law <law@redhat.com>
6341 PR tree-optimization/58640
6342 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
6343 threading paths that cross over two loop entry points.
6345 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6347 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
6348 handle vector float as well.
6349 (*vsx_le_perm_load_v4si): Likewise.
6350 (*vsx_le_perm_store_v2di): Likewise.
6351 (*vsx_le_perm_store_v4si): Likewise.
6353 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6355 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
6356 directly to circumvent subtract from splat{31} workaround.
6357 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
6359 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
6360 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
6361 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
6362 (altivec_vperm_<mode>): Convert to define_insn_and_split to
6363 separate big and little endian logic.
6364 (*altivec_vperm_<mode>_internal): New define_insn.
6365 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
6366 separate big and little endian logic.
6367 (*altivec_vperm_<mode>_uns_internal): New define_insn.
6368 (vec_permv16qi): Add little endian logic.
6370 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
6372 * doc/extend.texi (returns_nonnull): Remove arguments.
6374 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6375 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6376 Sergey Lega <sergey.s.lega@intel.com>
6377 Anna Tikhonova <anna.tikhonova@intel.com>
6378 Ilya Tocar <ilya.tocar@intel.com>
6379 Andrey Turetskiy <andrey.turetskiy@intel.com>
6380 Ilya Verbin <ilya.verbin@intel.com>
6381 Kirill Yukhin <kirill.yukhin@intel.com>
6382 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6384 * config/i386/sse.md (VI48F_256_512): New.
6385 (avx2_permvar<mode>): Change to ...
6386 (<avx2_avx512f>_permvar<mode>): This.
6388 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6389 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6390 Sergey Lega <sergey.s.lega@intel.com>
6391 Anna Tikhonova <anna.tikhonova@intel.com>
6392 Ilya Tocar <ilya.tocar@intel.com>
6393 Andrey Turetskiy <andrey.turetskiy@intel.com>
6394 Ilya Verbin <ilya.verbin@intel.com>
6395 Kirill Yukhin <kirill.yukhin@intel.com>
6396 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6398 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
6399 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
6400 * config/i386/sse.md (VI4_AVX): New.
6401 (sf2simodelower): Ditto.
6402 (sse2_cvtps2dq): Change to ...
6403 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
6405 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6406 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6407 Sergey Lega <sergey.s.lega@intel.com>
6408 Anna Tikhonova <anna.tikhonova@intel.com>
6409 Ilya Tocar <ilya.tocar@intel.com>
6410 Andrey Turetskiy <andrey.turetskiy@intel.com>
6411 Ilya Verbin <ilya.verbin@intel.com>
6412 Kirill Yukhin <kirill.yukhin@intel.com>
6413 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6415 * config/i386/sse.md (V_512): New.
6417 (vcond<V_512:mode><VF_512:mode>): Ditto.
6418 (vcond<V_512:mode><VI_512:mode>): Ditto.
6419 (vcondu<V_512:mode><VI_512:mode>): Ditto.
6421 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6422 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6423 Sergey Lega <sergey.s.lega@intel.com>
6424 Anna Tikhonova <anna.tikhonova@intel.com>
6425 Ilya Tocar <ilya.tocar@intel.com>
6426 Andrey Turetskiy <andrey.turetskiy@intel.com>
6427 Ilya Verbin <ilya.verbin@intel.com>
6428 Kirill Yukhin <kirill.yukhin@intel.com>
6429 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6431 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
6432 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
6434 (fma<mode>4): Removed condition.
6435 (fms<mode>4): Ditto.
6436 (fnma<mode>4): Ditto.
6437 (fnms<mode>4): Ditto.
6438 (fma4i_fmadd_<mode>): Ditto.
6439 (*fma_fmadd_<mode>): Ditto.
6440 (*fma_fmsub_<mode>): Ditto.
6441 (*fma_fnmadd_<mode>): Ditto.
6442 (*fma_fnmsub_<mode>): Ditto.
6443 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
6444 (*fma_fmaddsub_<mode>): Ditto.
6445 (*fma_fmsubadd_<mode>): Ditto.
6446 (*fmai_fmadd_<mode>): Ditto.
6447 (*fmai_fmsub_<mode>): Ditto.
6448 (*fmai_fnmadd_<mode>): Ditto.
6449 (*fmai_fnmsub_<mode>): Ditto.
6451 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6452 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6453 Sergey Lega <sergey.s.lega@intel.com>
6454 Anna Tikhonova <anna.tikhonova@intel.com>
6455 Ilya Tocar <ilya.tocar@intel.com>
6456 Andrey Turetskiy <andrey.turetskiy@intel.com>
6457 Ilya Verbin <ilya.verbin@intel.com>
6458 Kirill Yukhin <kirill.yukhin@intel.com>
6459 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6461 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
6462 (VI124_256): Changed to ...
6463 (VI124_256_48_512): This.
6464 (ssepackmode): Extended with wider modes.
6465 (<code><mode>3): Changed iterator.
6466 (*avx2_<code><mode>3): Ditto.
6467 (vec_pack_trunc_<mode>): Ditto.
6469 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6470 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6471 Sergey Lega <sergey.s.lega@intel.com>
6472 Anna Tikhonova <anna.tikhonova@intel.com>
6473 Ilya Tocar <ilya.tocar@intel.com>
6474 Andrey Turetskiy <andrey.turetskiy@intel.com>
6475 Ilya Verbin <ilya.verbin@intel.com>
6476 Kirill Yukhin <kirill.yukhin@intel.com>
6477 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6479 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
6480 (VI8F_256_512): Ditto.
6481 (abs<mode>2): Changed iterator.
6482 (avx2_perm<mode>): Changed to ...
6483 (<avx2_avx512f>_perm<mode>): This.
6484 (avx2_perm<mode>_1): Changed to ...
6485 (<avx2_avx512f>_perm<mode>_1): This.
6487 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6488 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6489 Sergey Lega <sergey.s.lega@intel.com>
6490 Anna Tikhonova <anna.tikhonova@intel.com>
6491 Ilya Tocar <ilya.tocar@intel.com>
6492 Andrey Turetskiy <andrey.turetskiy@intel.com>
6493 Ilya Verbin <ilya.verbin@intel.com>
6494 Kirill Yukhin <kirill.yukhin@intel.com>
6495 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6497 * config/i386/sse.md (VI48_AVX512F): New.
6498 (VI48_AVX2): Changed to ...
6499 (VI48_AVX2_48_AVX512F): This.
6500 (avx2_ashrv<mode>): Changed to ...
6501 (<avx2_avx512f>_ashrv<mode>): This.
6502 (avx2_<shift_insn>v<mode>): Changed to ...
6503 (<avx2_avx512f>_<shift_insn>v<mode>): This.
6505 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6506 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6507 Sergey Lega <sergey.s.lega@intel.com>
6508 Anna Tikhonova <anna.tikhonova@intel.com>
6509 Ilya Tocar <ilya.tocar@intel.com>
6510 Andrey Turetskiy <andrey.turetskiy@intel.com>
6511 Ilya Verbin <ilya.verbin@intel.com>
6512 Kirill Yukhin <kirill.yukhin@intel.com>
6513 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6515 * config/i386/sse.md (VI4_AVX512F): New.
6516 (VI8_AVX2_AVX512F): Ditto.
6517 (mul<mode>3): Extended with wider modes.
6518 (*<sse4_1_avx2>_mul<mode>3): Ditto.
6519 (mul<mode>3): Ditto.
6520 (vec_widen_<s>mult_odd_<mode>): Ditto.
6522 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6523 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6524 Sergey Lega <sergey.s.lega@intel.com>
6525 Anna Tikhonova <anna.tikhonova@intel.com>
6526 Ilya Tocar <ilya.tocar@intel.com>
6527 Andrey Turetskiy <andrey.turetskiy@intel.com>
6528 Ilya Verbin <ilya.verbin@intel.com>
6529 Kirill Yukhin <kirill.yukhin@intel.com>
6530 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6532 * config/i386/sse.md (VI2_AVX512F): New.
6533 (VI124_AVX512F): Ditto.
6534 (sseunpackmode): Extended with wider modes.
6535 (sseunpackfltmode): Ditto.
6536 (vec_unpacks_float_hi_<mode>): Ditto.
6537 (vec_unpacks_float_lo_<mode>): Ditto.
6538 (vec_unpacku_float_hi_<mode>): Ditto.
6539 (vec_unpacku_float_lo_<mode>): Ditto.
6540 (vec_unpacks_lo_<mode>): Ditto.
6541 (vec_unpacks_hi_<mode>): Ditto.
6542 (vec_unpacku_lo_<mode>): Ditto.
6543 (vec_unpacku_hi_<mode>): Ditto.
6545 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6546 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6547 Sergey Lega <sergey.s.lega@intel.com>
6548 Anna Tikhonova <anna.tikhonova@intel.com>
6549 Ilya Tocar <ilya.tocar@intel.com>
6550 Andrey Turetskiy <andrey.turetskiy@intel.com>
6551 Ilya Verbin <ilya.verbin@intel.com>
6552 Kirill Yukhin <kirill.yukhin@intel.com>
6553 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6555 * config/i386/i386.md (multdiv): New.
6556 (multdiv_mnemonic): Ditto.
6557 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
6558 (<sse>_vm<multdiv_mnemonic><mode>3): This.
6559 (<sse>_vmdiv<mode>3): Removed.
6561 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6562 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6563 Sergey Lega <sergey.s.lega@intel.com>
6564 Anna Tikhonova <anna.tikhonova@intel.com>
6565 Ilya Tocar <ilya.tocar@intel.com>
6566 Andrey Turetskiy <andrey.turetskiy@intel.com>
6567 Ilya Verbin <ilya.verbin@intel.com>
6568 Kirill Yukhin <kirill.yukhin@intel.com>
6569 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6571 * config/i386/sse.md (V): Extended with wider modes.
6573 (ssehalfvecmode): Ditto.
6575 (ssepackfltmode): Ditto.
6576 (avx_vec_concat<mode>): Ditto.
6577 (V_256_512): New iterator.
6578 (VF2_512_256): Ditto.
6579 (si2dfmode): New attribute.
6580 (si2dfmodelower): Ditto.
6582 (concat_tg_mode): Ditto.
6583 (floatv4siv4df2): Changed to ...
6584 (float<si2dfmodelower><mode>2): This.
6585 (avx_cvtps2pd256): Changed to ...
6586 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
6587 (vec_pack_trunc_v4df): Changed to ...
6588 (vec_pack_trunc_<mode>): This.
6589 (avx_vpermil<mode>): Changed to ...
6590 (<sse2_avx_avx512f>_vpermil<mode>): This.
6591 (<fixsuffix>fix_truncv8dfv8si2): New.
6592 (vec_pack_sfix_trunc_v8df): Ditto.
6593 (avx512f_rndscale<mode>): Ditto.
6594 (avx512f_roundpd512): Ditto.
6595 (vec_pack_ufix_trunc_<mode>): Updated iterator.
6597 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6598 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6599 Sergey Lega <sergey.s.lega@intel.com>
6600 Anna Tikhonova <anna.tikhonova@intel.com>
6601 Ilya Tocar <ilya.tocar@intel.com>
6602 Andrey Turetskiy <andrey.turetskiy@intel.com>
6603 Ilya Verbin <ilya.verbin@intel.com>
6604 Kirill Yukhin <kirill.yukhin@intel.com>
6605 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6607 * config/i386/i386.md (any_fix): New iterator.
6608 (fixsuffix): New attribute.
6609 * config/i386/sse.md (VF1): Extened with wider modes.
6613 (sseintvecmodelower): Ditto.
6614 (ssescalarmode): Ditto.
6615 (ssescalarnum): Ditto.
6617 (ssexmmmode): Ditto.
6618 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
6619 (<sse>_rcp<mode>2): Change iterator.
6620 (rsqrt<mode>2): Ditto.
6621 (<sse>_rsqrt<mode>2): Ditto.
6622 (avx2_vec_dup<mode>): Ditto.
6623 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
6624 (round<mode>2_sfix): Ditto.
6625 (avx2_pbroadcast<mode>): Ditto.
6626 (*andnot<mode>3): Handle XI mode.
6627 (*<code><mode>3): Ditto.
6628 (AVXTOSSEMODE): Removed.
6629 (avx_vpermil<mode>): Changed to ...
6630 (<sse2_avx_avx512f>_vpermil<mode>): This.
6632 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6633 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6634 Sergey Lega <sergey.s.lega@intel.com>
6635 Anna Tikhonova <anna.tikhonova@intel.com>
6636 Ilya Tocar <ilya.tocar@intel.com>
6637 Andrey Turetskiy <andrey.turetskiy@intel.com>
6638 Ilya Verbin <ilya.verbin@intel.com>
6639 Kirill Yukhin <kirill.yukhin@intel.com>
6640 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6642 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
6643 (<sse>_comi): Ditto.
6644 (<sse>_ucomi): Ditto.
6645 (sse_cvtss2siq_2): Ditto.
6646 (sse2_cvtsd2si): Ditto.
6647 (sse2_cvtsd2siq): Ditto.
6648 (sse2_cvttsd2si): Ditto.
6649 (sse2_cvttsd2siq): Ditto.
6650 (<shift_insn><mode>3): Ditto.
6651 (sse2_cvtsi2sdq): Update constraint and prefix.
6652 (sse_cvtsi2ss): Update prefix.
6653 (sse_cvtsi2ssq): Ditto.
6655 2013-10-11 Jakub Jelinek <jakub@redhat.com>
6657 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
6658 ignore internal calls.
6660 2013-10-11 Richard Biener <rguenther@suse.de>
6662 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
6663 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
6664 TREE_OVERFLOW is set.
6666 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
6668 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
6670 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
6671 * gimple.h: Likewise.
6673 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
6675 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
6678 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
6680 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
6681 * configure: Regenerate.
6683 2013-10-11 Jakub Jelinek <jakub@redhat.com>
6685 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
6686 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
6687 formatting fixes, use pp_colon instead of pp_character (..., ':'),
6688 similarly pp_right_paren.
6689 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
6690 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
6691 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
6692 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
6694 (omp_declare_simd_clauses_equal,
6695 omp_remove_redundant_declare_simd_attrs): New functions.
6696 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
6697 (walk_tree_1): Handle new OpenMP 4.0 clauses.
6698 * tree.h (OMP_LOOP_CHECK): Define.
6699 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
6700 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
6701 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
6702 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
6703 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
6704 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
6705 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
6706 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
6707 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
6708 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
6709 OMP_CLAUSE_SIMDLEN_EXPR): Define.
6710 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
6711 (omp_remove_redundant_declare_simd_attrs): New prototype.
6712 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
6713 GIMPLE_OMP_TEAMS): New codes.
6714 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
6715 * omp-low.c (struct omp_context): Add cancel_label and cancellable
6717 (target_nesting_level): New variable.
6718 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
6719 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
6720 for collapse > 1 static schedule unless ordered.
6721 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
6722 (determine_parallel_type): Adjust get_ws_args_for caller.
6723 (install_var_field): Handle mask & 4 for double indirection.
6724 (scan_sharing_clauses): Ignore shared clause on teams construct.
6725 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
6726 (create_omp_child_function): If inside target or declare target
6727 constructs, set "omp declare target" attribute on the child function.
6728 (find_combined_for): New function.
6729 (scan_omp_parallel): Handle combined loops.
6730 (scan_omp_target, scan_omp_teams): New functions.
6731 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
6732 restrictions and set ctx->cancellable for cancellable constructs.
6733 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
6734 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
6735 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
6736 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
6737 (omp_clause_aligned_alignment): New function.
6738 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
6739 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
6740 on teams constructs. Handle user defined reductions and new
6742 (lower_reduction_clauses): Don't set placeholder to address of ref
6743 if it has already the right type.
6744 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
6745 (expand_parallel_call): Use the new non-_start suffixed builtins,
6746 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
6747 and GOMP_parallel_end after the call.
6748 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
6749 (expand_omp_for_init_counts): Handle combined loops.
6750 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
6752 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
6753 end of cancellable loops.
6754 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
6755 Likewise. Handle collapse > 1 loops.
6756 (expand_omp_simd): Handle combined loops.
6757 (expand_omp_for): Add inner_stmt argument, adjust callers of
6758 expand_omp_for* functions, use expand_omp_for_static*chunk even
6759 for collapse > 1 unless ordered.
6760 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
6761 of cancellable sections.
6762 (expand_omp_single): Remove need_barrier variable, just rely on
6763 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
6764 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
6765 (expand_omp_atomic_load, expand_omp_atomic_store,
6766 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
6767 (expand_omp_target): New function.
6768 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
6769 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
6770 (build_omp_regions_1): Immediately close region for
6771 GF_OMP_TARGET_KIND_UPDATE.
6772 (maybe_add_implicit_barrier_cancel): New function.
6773 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
6775 (lower_omp_single): Likewise. Add clobber after the barrier.
6776 (lower_omp_taskgroup): New function.
6777 (lower_omp_for): Handle combined loops. Adjust
6778 lower_rec_input_clauses caller. Handle cancellation.
6779 (lower_depend_clauses): New function.
6780 (lower_omp_taskreg): Lower depend clauses. Adjust
6781 lower_rec_input_clauses caller. Add clobber after the call. Handle
6783 (lower_omp_target, lower_omp_teams): New functions.
6784 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
6785 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
6786 and GOMP_cancellation_point calls.
6787 (lower_omp): Fold stmts inside of target region.
6788 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
6789 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
6790 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
6791 (BT_FN_VOID_OMPFN_PTR_UINT,
6792 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
6793 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
6794 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
6795 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
6796 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
6797 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
6798 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
6799 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
6800 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
6801 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
6802 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
6803 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
6804 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
6805 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
6806 BUILT_IN_GOMP_PARALLEL_END.
6807 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
6808 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
6809 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
6810 GF_OMP_FOR_KIND_DISTRIBUTE.
6811 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
6812 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
6813 (dump_gimple_omp_return): Print lhs if it has any.
6814 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
6815 gimple_omp_atomic_seq_cst_p.
6816 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
6817 and GIMPLE_OMP_TEAMS.
6818 * langhooks.c (lhd_omp_mappable_type): New function.
6819 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
6820 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
6822 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
6823 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
6824 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
6825 (struct gimplify_omp_ctx): Add combined_loop field.
6826 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
6827 on stmts inside of target region.
6828 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
6829 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
6830 ORT_TARGET and ORT_TARGET_DATA.
6831 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
6833 (omp_notice_threadprivate_variable): Complain about threadprivate
6834 variables in target region.
6835 (omp_notice_variable): Complain about vars with non-mappable type
6836 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
6837 (omp_check_private): Ignore ORT_TARGET* regions.
6838 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
6839 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
6840 (find_combined_omp_for): New function.
6841 (gimplify_omp_for): Handle gimplification of combined loops.
6842 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
6844 (gimplify_omp_target_update): New function.
6845 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
6846 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
6847 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
6848 (gimplify_body): If fndecl has "omp declare target" attribute, add
6849 implicit ORT_TARGET context around it.
6850 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
6851 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
6852 * tree-nested.c (convert_nonlocal_reference_stmt,
6853 convert_local_reference_stmt, convert_gimple_call): Handle
6854 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
6855 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
6856 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
6857 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
6858 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
6859 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
6860 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
6861 BUILT_IN_GOMP_LOOP_END_CANCEL,
6862 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
6863 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
6864 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
6865 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
6866 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
6867 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
6868 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
6869 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
6870 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
6871 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
6872 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
6873 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
6874 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
6875 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
6876 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
6877 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
6878 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
6879 gimple_build_omp_teams): New functions.
6880 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
6881 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
6882 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
6883 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
6884 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
6885 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
6886 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
6887 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
6888 GF_OMP_ATOMIC_SEQ_CST): New.
6889 (gimple_build_omp_taskgroup, gimple_build_omp_target,
6890 gimple_build_omp_teams): New prototypes.
6891 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
6892 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
6893 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
6894 GIMPLE_OMP_SINGLE as end of range.
6895 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
6896 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
6897 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
6898 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
6899 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
6900 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
6901 gimple_omp_target_kind, gimple_omp_target_set_kind,
6902 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
6903 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
6904 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
6905 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
6906 gimple_omp_teams_set_clauses): New inlines.
6907 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
6908 and GIMPLE_OMP_TASKGROUP.
6909 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
6910 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
6911 enum omp_clause_proc_bind_kind): New.
6912 (union omp_clause_subcode): Add depend_kind, map_kind and
6913 proc_bind_kind fields.
6914 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
6915 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
6916 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
6917 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
6918 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
6920 2013-10-10 Teresa Johnson <tejohnson@google.com>
6922 * predict.c (tree_estimate_probability): Add new parameter
6923 for estimate_bb_frequencies.
6924 (estimate_bb_frequencies): Add new parameter to force estimation.
6925 (rebuild_frequencies): When max frequency in function is small,
6926 recompute counts from frequencies.
6927 * predict.h (estimate_bb_frequencies): New parameter.
6929 2013-10-10 David Malcolm <dmalcolm@redhat.com>
6931 * ipa-inline.c (ipa_inline): Fix leak of "order" when
6932 optimizations are disabled.
6934 2013-10-10 David Malcolm <dmalcolm@redhat.com>
6936 * coverage.c (coverage_finish): Fix leak of da_file_name.
6938 2013-10-10 Jan Hubicka <jh@suse.cz>
6940 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
6941 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
6944 2013-10-10 Jakub Jelinek <jakub@redhat.com>
6947 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
6948 if any labels are in FALLTHRU_BB, use a special label emitted
6949 immediately after the asm goto insn rather than label_rtx
6951 (expand_asm_stmt): Adjust caller.
6952 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
6953 edge if the last insn in predecessor is a jump with single successor,
6954 but it isn't simplejump_p.
6956 2013-10-10 Richard Biener <rguenther@suse.de>
6958 PR tree-optimization/58656
6959 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
6961 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
6963 * gimplify.c: Include expr.h and tm_p.h for targets with special
6964 va-arg padding requirements.
6966 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
6968 * tree-flow.h: Move some prototypes to gimple.h.
6969 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
6970 * gimple.h: Relocate some prototypes from tree-flow.h
6971 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
6973 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
6974 (build_va_arg_indirect_ref): Relocate and make static.
6975 (std_gimplify_va_arg_expr): Relocate here.
6976 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
6977 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
6979 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6981 * doc/md.texi: Document the mnemonic attribute.
6983 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6986 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
6987 (set_attr_alternative x ...) when searching for user defined
6990 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
6992 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
6994 2013-10-09 Easwaran Raman <eraman@google.com>
6996 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
6997 * cfgexpand.c (defer_stack_allocation): ...use here
6998 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
7000 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
7002 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
7003 (operand_equal_for_value_replacement): New function, extracted from
7004 value_replacement and enhanced to catch more cases.
7005 (value_replacement): Use operand_equal_for_value_replacement.
7007 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7009 * loop-doloop.c (doloop_modify, doloop_optimize): Use
7010 get_max_loop_iterations.
7012 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7014 * config/arm/aarch-common.c (arm_early_load_addr_dep):
7015 Place comment above function.
7017 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7019 * tree-flow.h: Remove all remaining prototypes, enums and structs that
7020 are not related to tree-cfg.c.
7021 * tree-ssa-address.h: New file. Relocate prototypes.
7022 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
7023 (addr_for_mem_ref): New. Combine call to get_address_description and
7024 return addr_for_mem_ref.
7025 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
7026 * tree-ssa-live.h: Adjust prototypes.
7027 * passes.c: Include tree-ssa-live.h.
7028 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
7029 * graphite.c (graphite_transform_loops): Make static.
7030 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
7031 make_pass_graphite, pass_data_graphite_transforms,
7032 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
7033 * ipa-pure-const.c (warn_function_noreturn): Make static.
7034 (execute_warn_function_noreturn, gate_warn_function_noreturn,
7035 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
7036 Relocate from tree-cfg.c
7037 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
7039 (execute_warn_function_noreturn, gate_warn_function_noreturn,
7040 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
7041 Move to ipa-pure-const.c.
7042 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
7043 Relocate from tree-optimize.c.
7044 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
7045 make_pass_fixup_cfg): Move to tree-cfg.c.
7046 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
7047 Relocate some prototypes.
7048 * tree-data-ref.h (tree_check_data_deps) Add prototype.
7049 * tree-dump.c (dump_function_to_file): Remove prototype.
7050 Add tree-flow.h to the include file.
7051 * tree-dump.h: Remove prototype.
7052 * tree-parloops.h: New File. Add prototypes.
7053 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
7054 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
7055 from tree-ssa-loop.c.
7056 * tree-predcom.c (run_tree_predictive_commoning,
7057 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
7058 Relocate here from tree-ssa-loop.c.
7059 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
7060 ssa_name_values.release ().
7061 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
7062 (ssa_name_values): Relocate from tree-flow.h.
7063 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
7064 * tree-ssa-loop.c (run_tree_predictive_commoning,
7065 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
7066 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
7067 make_pass_graphite, pass_data_graphite_transforms,
7068 make_pass_graphite_transforms, gate_tree_parallelize_loops,
7069 tree_parallelize_loops, pass_data_parallelize_loops,
7070 make_pass_parallelize_loops): Move to other files.
7071 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
7073 * tree.h: Remove prototypes from tree-address.c.
7075 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7077 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
7078 (struct int_tree_map): Move to tree-hasher.h
7079 (SCALE, LABEL, PERCENT): Move to gimple.h
7080 * tree-flow-inline.h: Delete. Move functions to other files.
7081 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
7082 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
7083 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
7084 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
7085 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
7086 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
7087 * tree-hasher.h: Don't include tree-flow.h.
7088 (struct int_tree_map): Relocate from tree-flow.h.
7089 * tree-sra.c (contains_view_convert_expr_p): Relocate from
7090 tree-flow-inline.h and make static.
7091 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
7093 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
7094 tree-flow-inline.h and make static.
7095 * tree.h (is_global_var, may_be_aliased): Relocate from
7097 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
7098 * value-prof.c: No longer include tree-flow-inline.h.
7099 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
7101 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
7103 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
7104 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
7105 (enum move_pos): Move to tree-ssa-loop-im.h
7106 * cfgloop.h: Move some prototypes.
7107 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
7108 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
7109 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
7110 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
7111 (loop_containing_stmt): Relocate from tree-flow-inline.h.
7112 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
7113 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
7114 (enum move_pos): Relocate here.
7115 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
7117 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
7118 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
7119 make_pass_lim): Relocate here from tree-ssa-loop.c.
7120 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
7122 (loop_edge_to_cancel, unloop_loops): Make static.
7123 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
7124 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
7125 (tree_complete_unroll, gate_tree_complete_unroll,
7126 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
7127 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
7128 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
7129 * tree-ssa-loop-ivopts.c: Remove local prototypes.
7130 (stmt_invariant_in_loop_p): Remove unused function.
7131 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
7132 * tree-ssa-loop-manip.h: New file. Add prototypes.
7133 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
7134 (gcov_type_to_double_int): Move to cfgloop.h.
7135 (double_int_cmp, bound_index,
7136 estimate_numbers_of_iterations_loop): Make static.
7137 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
7138 (max_loop_iterations): Factor out get_max_loop_iterations.
7139 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
7141 * tree-ssa-loop-niter.h: New file. Add prototypes.
7142 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
7143 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
7144 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
7145 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
7146 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
7147 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
7148 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
7149 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
7150 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
7151 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
7152 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
7153 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
7154 pass_data_complete_unroll, make_pass_complete_unroll,
7155 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
7156 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
7157 tree-ssa-loop-ivcanon.c.
7158 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
7159 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
7160 tree-ssa-loop-prefetch.c.
7161 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
7163 (get_lsm_tmp_name): Relocate and add suffix parameter.
7164 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
7165 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
7166 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
7167 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
7168 (get_estimated_loop_iterations): Factor out accessor from
7169 estimated_loop_iterations in tree-ssa-loop-niter.c.
7170 (get_max_loop_iterations): Factor out accessor from
7171 _max_loop_iterations in tree-ssa-niter.c.
7172 * loop-unroll.c (decide_unroll_constant_iterations,
7173 decide_unroll_runtime_iterations, decide_peel_simple,
7174 decide_unroll_stupid): Use new get_* accessors.
7176 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
7178 PR tree-optimization/20318
7179 * doc/extend.texi (returns_nonnull): New function attribute.
7180 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
7182 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
7183 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
7185 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
7188 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
7189 false if both components are bitfields.
7191 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7193 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
7194 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
7195 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
7196 * config/aarch64/aarch64.h
7197 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
7198 * config/aarch64/aarch64-simd-builtins.def
7199 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
7201 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7203 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
7205 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7207 * config/aarch64/arm_neon.h (vdiv_f64): Added.
7209 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
7211 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
7212 (vneg_f64): New intrinsic.
7213 (vneg_s8): Asm replaced with C.
7214 (vneg_s16): Likewise.
7215 (vneg_s32): Likewise.
7216 (vneg_s64): New intrinsic.
7217 (vnegq_f32): Asm replaced with C.
7218 (vnegq_f64): Likewise.
7219 (vnegq_s8): Likewise.
7220 (vnegq_s16): Likewise.
7221 (vnegq_s32): Likewise.
7222 (vnegq_s64): Likewise.
7224 2013-10-09 Renlin Li <Renlin.Li@arm.com>
7226 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
7228 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7230 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
7231 packed stack special handling.
7232 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
7233 back to fixed stack slots for FPRs saved due to stdarg.
7235 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7237 * config/s390/s390.c (s390_frame_info): Restructure function.
7239 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7241 * config/s390/s390.c (struct s390_frame_layout): New field
7243 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
7244 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
7245 regs_ever_clobbered to char*.
7246 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
7247 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
7248 (s390_register_info): Enable FPR save slots. Move/Copy some
7249 functionality into ...
7250 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
7251 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
7253 (s390_frame_info): Do gpr slot allocation here now. stdarg does
7254 not imply a stack frame.
7255 (s390_init_frame_layout): Remove variable clobbered_regs.
7256 (s390_update_register_info): Remove function.
7257 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
7258 cannot be used for register renaming.
7259 (s390_hard_regno_scratch_ok): New function.
7260 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
7261 (s390_initial_elimination_offset): Change offset calculation of
7262 the return address pointer.
7263 (save_gprs): Deal with only r6 being saved from the call-saved regs.
7264 (restore_gprs): Set frame related flag.
7265 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
7266 (s390_emit_prologue): Call s390_register_info instead of
7267 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
7268 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
7269 (s390_optimize_prologue): Call s390_optimize_register_info.
7270 Try to remove also FPR slot save/restore INSNs. Remove frame
7271 related flags from restore INSNs.
7273 2013-10-08 DJ Delorie <dj@redhat.com>
7275 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
7278 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
7279 avoid conflict with the MI use of %c.
7280 * config/rl78/rl78-real.md: change %c to %C throughout.
7281 * config/rl78/rl78-virt.md: Likewise.
7283 2013-10-08 Jan Hubicka <jh@suse.cz>
7285 * config/i386/i386.c (ix86_option_override_internal): Switch
7286 to SSE math for -ffast-math when target ISA supports SSE2.
7288 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
7290 * tree-flow.h: Remove some prototypes.
7291 * tree.h: Remove some protypes, add a couple.
7292 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
7293 using_eh_for_cleanups_p): Add interface routines for front ends.
7294 * tree-eh.h: New file. Add protoptyes.
7295 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
7296 (add_stmt_to_eh_lp_fn): Make static.
7297 (lower_try_finally): Use new using_eh_for_cleanups_p.
7298 * emit-rtl.c: Include tree-eh.h.
7299 * gimple.h: Include tree-eh.h.
7301 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
7303 PR tree-optimization/58480
7304 * tree-vrp.c (infer_nonnull_range): New function.
7305 (infer_value_range): Call infer_nonnull_range.
7307 2013-10-08 Dehao Chen <dehao@google.com>
7309 PR tree-optimization/58619
7310 * tree-inline.c (copy_phis_for_bb): Combine location data
7313 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7316 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
7317 RTX_FRAME_RELATED_P on INSN.
7319 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7321 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
7322 (altivec_expand_vec_perm_const): Call it.
7324 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7326 * config/rs6000/vector.md (mov<mode>): Emit permuted move
7327 sequences for LE VSX loads and stores at expand time.
7328 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
7330 * config/rs6000/rs6000.c (rs6000_const_vec): New.
7331 (rs6000_gen_le_vsx_permute): New.
7332 (rs6000_gen_le_vsx_load): New.
7333 (rs6000_gen_le_vsx_store): New.
7334 (rs6000_gen_le_vsx_move): New.
7335 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
7336 (*vsx_le_perm_load_v4si): New.
7337 (*vsx_le_perm_load_v8hi): New.
7338 (*vsx_le_perm_load_v16qi): New.
7339 (*vsx_le_perm_store_v2di): New.
7340 (*vsx_le_perm_store_v4si): New.
7341 (*vsx_le_perm_store_v8hi): New.
7342 (*vsx_le_perm_store_v16qi): New.
7343 (*vsx_xxpermdi2_le_<mode>): New.
7344 (*vsx_xxpermdi4_le_<mode>): New.
7345 (*vsx_xxpermdi8_le_V8HI): New.
7346 (*vsx_xxpermdi16_le_V16QI): New.
7347 (*vsx_lxvd2x2_le_<mode>): New.
7348 (*vsx_lxvd2x4_le_<mode>): New.
7349 (*vsx_lxvd2x8_le_V8HI): New.
7350 (*vsx_lxvd2x16_le_V16QI): New.
7351 (*vsx_stxvd2x2_le_<mode>): New.
7352 (*vsx_stxvd2x4_le_<mode>): New.
7353 (*vsx_stxvd2x8_le_V8HI): New.
7354 (*vsx_stxvd2x16_le_V16QI): New.
7356 2013-10-07 Renlin Li <Renlin.Li@arm.com>
7358 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
7360 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7362 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
7363 loop to work also for 31bit ABI.
7364 Save the stack pointer for frame_size > 0.
7366 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7368 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
7369 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
7370 constraint letters from expanders.
7371 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
7372 retry count to general_operand.
7373 ("tabort"): Give operand 0 a mode.
7374 ("tabort_1"): Add mode and constraint letter for operand 0.
7375 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
7377 2013-10-04 Jeff Law <law@redhat.com>
7379 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
7381 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
7383 (thread_across_edge): Here. Call it.
7385 2013-10-04 Cary Coutant <ccoutant@google.com>
7387 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
7388 discarding a location list expression (or a piece of one).
7390 2013-10-03 Jan Hubicka <jh@suse.cz>
7392 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
7393 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
7394 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
7396 2013-10-03 Jan Hubicka <jh@suse.cz>
7398 * config/i386/i386.c (ix86_option_override_internal): Do not enable
7399 accumulate-outgoing-args when producing unwind info.
7401 2013-10-03 Wei Mi <wmi@google.com>
7403 * lra-constraints.c (insert_move_for_subreg): New function
7404 extracted from simplify_operand_subreg.
7405 (simplify_operand_subreg): Add reload for paradoxical subreg.
7407 2013-10-03 Rong Xu <xur@google.com>
7409 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
7410 the candidate of builtin_expect such that we should fix the
7411 size/time estimation.
7412 (estimate_function_body_sizes): Do the acutally size/time fix-up
7415 2013-10-03 Rong Xu <xur@google.com>
7417 * predict.c (tree_predict_by_opcode): Get the probability
7418 for builtin_expect from param builtin_expect_probability.
7419 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
7420 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
7421 * doc/invoke.texi: Add documentation for builtin-expect-probability.
7423 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
7426 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
7427 * calls.c (alloca_call_p): Use get_callee_fndecl.
7428 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
7429 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
7431 (vrp_visit_stmt): Remove duplicated code.
7433 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
7435 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
7436 ceildf2, btruncdf2, instead of vsx_* name.
7438 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
7439 iterators to only do V2DF and V4SF here. Move the DF code to
7440 rs6000.md where it is combined with SF mode. Replace <VSv> with
7441 just 'v' since only vector operations are handled with these insns
7442 after moving the DF support to rs6000.md.
7443 (vsx_sub<mode>3): Likewise.
7444 (vsx_mul<mode>3): Likewise.
7445 (vsx_div<mode>3): Likewise.
7446 (vsx_fre<mode>2): Likewise.
7447 (vsx_neg<mode>2): Likewise.
7448 (vsx_abs<mode>2): Likewise.
7449 (vsx_nabs<mode>2): Likewise.
7450 (vsx_smax<mode>3): Likewise.
7451 (vsx_smin<mode>3): Likewise.
7452 (vsx_sqrt<mode>2): Likewise.
7453 (vsx_rsqrte<mode>2): Likewise.
7454 (vsx_fms<mode>4): Likewise.
7455 (vsx_nfma<mode>4): Likewise.
7456 (vsx_copysign<mode>3): Likewise.
7457 (vsx_btrunc<mode>2): Likewise.
7458 (vsx_floor<mode>2): Likewise.
7459 (vsx_ceil<mode>2): Likewise.
7460 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
7461 (vsx_sminsf3): Likewise.
7462 (vsx_fmadf4): Likewise.
7463 (vsx_fmsdf4): Likewise.
7464 (vsx_nfmadf4): Likewise.
7465 (vsx_nfmsdf4): Likewise.
7466 (vsx_cmpdf_internal1): Likewise.
7468 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
7469 simpler to select whether a target has SPE or traditional floating
7470 point support in iterators.
7471 (TARGET_DF_SPE): Likewise.
7472 (TARGET_SF_FPR): Likewise.
7473 (TARGET_DF_FPR): Likewise.
7474 (TARGET_SF_INSN): Macros to say whether floating point support
7475 exists for a given operation for expanders.
7476 (TARGET_DF_INSN): Likewise.
7478 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
7479 combining of SF/DF mode operations, using both traditional and VSX
7487 (abs<mode>2): Combine SF/DF modes using traditional floating point
7488 instructions. Add support for using the upper DF registers with
7489 VSX support, and SF registers with power8-vector support. Update
7490 expanders for operations supported by both the SPE and traditional
7491 floating point units.
7492 (abs<mode>2_fpr): Likewise.
7493 (nabs<mode>2): Likewise.
7494 (nabs<mode>2_fpr): Likewise.
7495 (neg<mode>2): Likewise.
7496 (neg<mode>2_fpr): Likewise.
7497 (add<mode>3): Likewise.
7498 (add<mode>3_fpr): Likewise.
7499 (sub<mode>3): Likewise.
7500 (sub<mode>3_fpr): Likewise.
7501 (mul<mode>3): Likewise.
7502 (mul<mode>3_fpr): Likewise.
7503 (div<mode>3): Likewise.
7504 (div<mode>3_fpr): Likewise.
7505 (sqrt<mode>3): Likewise.
7506 (sqrt<mode>3_fpr): Likewise.
7507 (fre<Fs>): Likewise.
7508 (rsqrt<mode>2): Likewise.
7509 (cmp<mode>_fpr): Likewise.
7510 (smax<mode>3): Likewise.
7511 (smin<mode>3): Likewise.
7512 (smax<mode>3_vsx): Likewise.
7513 (smin<mode>3_vsx): Likewise.
7514 (negsf2): Delete SF operations that are merged with DF.
7521 (fmasf4_fpr): Likewise.
7522 (fmssf4_fpr): Likewise.
7523 (nfmasf4_fpr): Likewise.
7524 (nfmssf4_fpr): Likewise.
7525 (sqrtsf2): Likewise.
7526 (rsqrtsf_internal1): Likewise.
7527 (smaxsf3): Likewise.
7528 (sminsf3): Likewise.
7529 (cmpsf_internal1): Likewise.
7530 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
7531 (negdf2): Delete DF operations that are merged with SF.
7533 (nabsdf2): Likewise.
7539 (rsqrtdf_internal1): Likewise.
7540 (fmadf4_fpr): Likewise.
7541 (fmsdf4_fpr): Likewise.
7542 (nfmadf4_fpr): Likewise.
7543 (nfmsdf4_fpr): Likewise.
7544 (sqrtdf2): Likewise.
7545 (smaxdf3): Likewise.
7546 (smindf3): Likewise.
7547 (cmpdf_internal1): Likewise.
7548 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
7549 (btrunc<mode>2): Delete separate expander, and combine with the
7550 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
7551 (btrunc<mode>2_fpr): Likewise.
7552 (ceil<mode>2): Likewise.
7553 (ceil<mode>2_fpr): Likewise.
7554 (floor<mode>2): Likewise.
7555 (floor<mode>2_fpr): Likewise.
7556 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
7557 Add support for using the upper registers with VSX and
7558 power8-vector. Move insns to be closer to the define_expands. On
7559 VSX systems, prefer the traditional form of FMA over the VSX
7560 version, since the traditional form allows the target not to
7561 overlap with the inputs.
7562 (fms<mode>4_fpr): Likewise.
7563 (nfma<mode>4_fpr): Likewise.
7564 (nfms<mode>4_fpr): Likewise.
7566 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7567 Richard Earnshaw <richard.earnshaw@arm.com>
7569 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
7570 (struct mult_cost_table): Likewise.
7571 (struct mem_cost_table): Likewise.
7572 (struct fp_cost_table): Likewise.
7573 (struct vector_cost_table): Likewise.
7574 (cpu_cost_table): Likewise.
7575 * config/arm/arm.opt (mold-rts-costs): New option.
7576 (mnew-generic-costs): Likewise.
7577 * config/arm/arm.c (generic_extra_costs): New table.
7578 (cortexa15_extra_costs): Likewise.
7579 (arm_slowmul_tune): Use NULL as new costs.
7580 (arm_fastmul_tune): Likewise.
7581 (arm_strongarm_tune): Likewise.
7582 (arm_xscale_tune): Likewise.
7583 (arm_9e_tune): Likewise.
7584 (arm_v6t2_tune): Likewise.
7585 (arm_cortex_a5_tune): Likewise.
7586 (arm_cortex_a9_tune): Likewise.
7587 (arm_v6m_tune): Likewise.
7588 (arm_fa726te_tune): Likewise.
7589 (arm_cortex_a15_tune): Use cortex15_extra_costs.
7590 (arm_cortex_tune): Use generict_extra_costs.
7591 (shifter_op_p): New function.
7592 (arm_unspec_cost): Likewise.
7593 (LIBCALL_COST): Define.
7594 (arm_new_rtx_costs): New function.
7595 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
7596 table is available. Use old costs otherwise unless mnew-generic-costs
7598 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
7599 (cpu_cost_table): Declare.
7601 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
7604 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
7605 (*subs_mul_imm_<mode>)
7606 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
7607 (*sub_<shift>_<mode>)
7608 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
7609 Remove k constraint.
7611 2013-10-03 Ian Bolton <ian.bolton@arm.com>
7613 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
7615 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
7617 2013-10-02 Teresa Johnson <tejohnson@google.com>
7619 * predict.c (probably_never_executed): New function.
7620 (probably_never_executed_bb_p): Invoke probably_never_executed.
7621 (probably_never_executed_edge_p): Ditto.
7622 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
7623 Treat profile insanities conservatively.
7625 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
7627 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
7629 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
7631 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
7632 use it. Use smaller increase for scratch. Don't increase reject
7633 for early clobber scratch.
7634 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
7635 setting eliminated regs except setting fp from hfp.
7636 (lra_eliminate): Check lra_insn_recog_data on NULL.
7638 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7641 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
7642 setting -mvsx-timode by default until the underlying problem is fixed.
7643 (RS6000_CPU, power7 defaults): Likewise.
7645 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
7647 * config/x-linux (host-linux.o): Remove header dependencies.
7648 Use $(COMPILE) and $(POSTCOMPILE).
7649 * config/t-linux-android (linux-android.o): Ditto.
7651 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
7653 * Makefile.in (expmed.o-warn): Remove.
7655 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
7657 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
7658 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
7660 2013-10-02 Teresa Johnson <tejohnson@google.com>
7662 * dojump.c (do_jump_1): Divide probability between
7663 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
7665 2013-10-02 Tom Tromey <tromey@redhat.com>
7667 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
7669 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
7671 * tree-flow.h: Remove some prototypes.
7672 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
7673 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
7674 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
7675 mark_virtual_phi_result_for_renaming): Relocate here.
7676 * tree-into-ssa.h: Add prototypes.
7677 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
7678 single_pred_before_succ_order.
7679 (blocks_in_phiopt_order): Rename and move to cfganal.c.
7680 (nonfreeing_call_p) Move to gimple.c.
7681 * cfganal.c (single_pred_before_succ_order): Move and renamed from
7683 * basic-block.h (single_pred_before_succ_order): Add prototype.
7684 * gimple.c (nonfreeing_call_p): Relocate here.
7685 * gimple.h: Add prototype.
7686 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
7687 * tree-ssa-dom.h: New file. Relocate prototypes here.
7688 * tree-ssa.h: Include tree-ssa-dom.h.
7690 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
7692 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
7693 Use $(COMPILE) and $(POSTCOMPILE).
7695 * config/alpha/x-alpha (driver-alpha.o): Ditto.
7697 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
7699 * tree-flow.h: Remove some prototypes.
7700 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
7701 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
7702 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
7703 propagate_tree_value*): Move from here to...
7704 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
7705 propagate_tree_value*): Relocate here.
7706 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
7707 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
7708 * gimple-fold.c: Remove gimple-fold.h from include list.
7709 * tree-vrp.c: Remove gimple-fold.h from include list.
7710 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
7711 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
7712 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
7713 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
7714 * sese.c: Add tree-ssa-propagate.h to include list.
7716 2013-10-02 Richard Biener <rguenther@suse.de>
7718 * tree-loop-distribution.c: Include tree-vectorizer.h for
7720 (enum partition_kind): Remove PKIND_REDUCTION.
7721 (struct partition_s): Remove has_writes member, add reduction_p member.
7722 (partition_alloc): Adjust.
7723 (partition_builtin_p): Likewise.
7724 (partition_has_writes): Remove.
7725 (partition_reduction_p): New function.
7726 (partition_merge_into): Likewise.
7727 (generate_code_for_partition): Commonize builtin partition
7729 (rdg_cannot_recompute_vertex_p): Remove.
7730 (already_processed_vertex_p): Likewise.
7731 (rdg_flag_vertex): Do not set has_writes.
7732 (classify_partition): Adjust.
7733 (rdg_build_partitions): Do not set has_writes, treat all
7734 partitions as useful.
7735 (distribute_loop): Record number of library calls generated. Adjust.
7736 (tree_loop_distribution): Report number of loops and library
7737 calls generated as opt-info.
7739 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
7741 * tree-flow.h: Include new .h files. Move prototypes.
7742 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
7743 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
7744 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
7745 * tree-pretty-print.h: Add prototypes from tree-flow.h.
7746 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
7747 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
7748 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
7749 * tree.h (get_ref_base_and_extent): Move prototype out.
7750 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
7752 * gimple-low.h: New File. Add prototypes from tree-flow.h.
7753 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
7754 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
7755 * tree-scalar-evolution.c: Include tree.h.
7756 * sese.c: Include tree.h.
7757 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
7758 * dwarf2out.c: Include tree-dfa.h.
7759 * tree-chrec.c: Include tree.h.
7760 * tree-data-ref.c: Include tree.h.
7762 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
7764 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
7767 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7769 * config/t-sol2 (sol2-c.o): Remove header dependencies.
7770 Use $(COMPILE) and $(POSTCOMPILE).
7771 (sol2-cxx.o): Likewise.
7772 (sol2-stubs.o): Likewise.
7774 * config/x-solaris (host-solaris.o): Likewise.
7776 * config/sparc/t-sparc (sparc.o): Remove.
7777 (sparc-c.o): Remove header dependencies.
7778 Use $(COMPILE) and $(POSTCOMPILE).
7779 * config/sparc/x-sparc: Likewise.
7781 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
7783 * config/arc/arc-opts.h: Add 2013 to Copyright years.
7784 * config/arc/arc700.md: Likewise.
7785 * config/arc/arc-modes.def: Likewise.
7786 * config/arc/arc-simd.h: Likewise.
7787 * config/arc/t-arc-uClibc: Likewise.
7788 * config/arc/t-arc-newlib: Likewise.
7790 2013-10-02 Renlin Li <renlin.li@arm.com>
7792 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7794 (aarch64_expand_epilogue): Likewise.
7796 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7797 Yufeng Zhang <yufeng.zhang@arm.com>
7799 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
7801 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
7802 'base_in' represent a conversion and legal_cast_p_1 holds; set
7803 'base_in' with the returned value from get_unwidened.
7805 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7807 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
7808 plus_constant is not used.
7810 2013-10-01 Wei Mi <wmi@google.com>
7812 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
7813 * config/i386/i386.md: Add define_peephole2 to
7814 break partial reg stall for cvtss2sd/cvtsd2ss.
7816 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7818 * config/arc/arc.c (pass_arc_ifcvt::clone):
7819 Update for ctxt_ -> m_ctxt change.
7821 2013-10-01 Jeff Law <law@redhat.com>
7823 * tree-ssa-threadupdate.c (struct redirection_data): Delete
7824 outgoing_edge and intermediate_edge fields. Instead store the path.
7825 (redirection_data::hash): Hash on the last edge's destination index.
7826 (redirection_data::equal): Check the entire thread path.
7827 (lookup_redirectio_data): Corresponding changes.
7828 (create_edge_and_update_destination_phis): Likewise.
7829 (thread_single_edge): Likewise.
7831 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7832 Diego Novillo <dnovillo@google.com>
7834 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
7835 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
7836 (vld32wh_insn, vld32wl_insn): Delete commented-out old
7837 versions of these patterns.
7839 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
7840 (__builtin_arc_aligned): Likewise.
7842 * config/arc/arc.md: Expand adc_0 comment stating the intended
7843 purpose and why it isn't ready.
7844 Replace commented out call_value_via_label_mixed with a
7845 plain comment about bl_s.
7847 * config/arc/arc.c (stdio.h): Don't include directly.
7848 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
7849 Remove if (1) condition.
7850 (arc_encode_section_info): Fix comment.
7852 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7854 * config/arc/arc.c (arc_conditional_register_usage):
7855 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
7856 Also set reg_alloc_order for DMA config regs.
7858 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7859 Jeremy Bennett <jeremy.bennett@embecosm.com>
7861 * doc/install.texi (--with-cpu): Mention ARC.
7862 (arc-*-elf32): New paragraph.
7863 (arc-linux-uclibc): Likewise.
7864 * doc/md.texi (Machine Constraints): Add ARC part.
7865 * doc/invoke.texi: (menu): Add ARC Options.
7866 (Machine Dependent Options) <ARC Options>: Add synopsis.
7867 (node ARC Options): Add.
7868 * doc/extend.texi (long_call / short_call attribute): Add ARC.
7869 (ARC Built-in Functions): New section defining
7870 generic ARC built-in functions.
7871 (ARC SIMD Built-in Functions): New section defining SIMD specific
7873 (Declaring Attributes of Functions): Extended
7874 description of short_call and long_call attributes for ARC and
7875 added index entries.
7877 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
7878 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
7879 Joern Rennecke <joern.rennecke@embecosm.com>
7880 Muhammad Khurram Riaz <khurram.riaz@arc.com>
7881 Brendan Kehoe <brendan@zen.org>
7882 Michael Eager <eager@eagercon.com>
7883 Simon Cook <simon.cook@embecosm.com>
7884 Jeremy Bennett <jeremy.bennett@embecosm.com>
7886 * config/arc, common/config/arc: New directories.
7888 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7889 Brendan Kehoe <brendan@zen.org>
7890 Simon Cook <simon.cook@embecosm.com>
7892 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
7894 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
7896 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
7897 * tree-ssa-coalesce.h: New. Move prototype to here.
7898 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
7899 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
7900 (gimple_can_coalesce_p): Move to...
7901 * gimple.c (gimple_can_coalesce_p): Here.
7903 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
7905 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
7906 (dump_decl_set): Move to gimple.c.
7907 * gimple.h: Don't include tree-ssa-operands.h.
7908 (dump_decl_set): Add prototype.
7909 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
7910 Move to gimple-ssa.h.
7911 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
7912 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
7913 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
7914 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
7915 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
7917 (dump_decl_set): Relocate here.
7918 * gimple-ssa.h: New file.
7919 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
7920 Relocate from gimple.h.
7921 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
7922 * tree-ssa-operands.c (swap_ssa_operands): Rename from
7923 swap_tree_operands and remove non-ssa path.
7924 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
7925 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
7927 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
7928 vect_is_simple_reduction_1): Use swap_ssa_operands.
7929 * tree-flow.h: Move various prototypes to tree-phinodes.h.
7930 (enum need_phi_state): Move to tree-into-ssa.c.
7931 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
7932 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
7933 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
7934 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
7935 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
7936 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
7937 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
7938 num_imm_uses): Move to ssa-iterators.h.
7939 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
7940 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
7942 (op_iter_done, op_iter_next_def, op_iter_next_tree,
7943 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
7944 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
7945 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
7946 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
7947 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
7948 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
7949 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
7950 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
7951 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
7952 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
7953 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
7954 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
7955 (set_phi_nodes): Move to tree-phinodes.h.
7956 * tree-ssa-operands.h (enum ssa_op_iter_type,
7957 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
7958 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
7959 (dump_decl_set): Remove prototype.
7960 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
7961 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
7962 (set_phi_nodes): Relocate from tree-flow-inline.h.
7963 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
7965 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
7966 include list. Temporarily add gimple.h to include list.
7967 * ssa-iterators.h: New file.
7968 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
7969 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
7970 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
7971 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
7972 Relocate from tree-ssa-operands.h.
7973 (delink_imm_use, link_imm_use_to_list, link_imm_use,
7974 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
7975 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
7976 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
7977 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
7978 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
7979 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
7980 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
7981 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
7982 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
7983 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
7984 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
7985 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
7986 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
7987 Relocate from tree-flow-inline.h.
7988 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
7990 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
7993 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
7994 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
7995 Insert '\t' to output template.
7996 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
7997 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
7998 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
7999 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
8001 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
8003 * doc/install.texi (Host/target specific installation notes for GCC):
8004 Put @anchor before @heading.
8005 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
8006 Use @email for email addresses.
8008 2013-10-01 Jeff Law <law@redhat.com>
8010 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
8011 a vec. Only delete the path if we create one without successfully
8012 registering a jump thread.
8013 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
8015 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
8016 (paths): New vector of jump threading paths.
8017 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
8018 (THREAD_PATH): New accessor macro for the entire thread path.
8019 (lookup_redirection_data): Get intermediate and final outgoing edge
8020 from the thread path.
8021 (create_edge_and_update_destination_phis): Copy the threading path.
8022 (ssa_fix_duplicate_block_edges): Get edges and block types from the
8023 jump threading path.
8024 (ssa_redirect_edges): Get edges and block types from the jump threading
8025 path. Free the path vector.
8026 (thread_block): Get edges from the jump threading path. Look at the
8027 entire path to see if we thread to a loop exit. If we cancel a jump
8028 thread request, then free the path vector.
8029 (thread_single_edge): Get edges and block types from the jump threading
8030 path. Free the path vector.
8031 (thread_through_loop_header): Get edges and block types from the jump
8032 threading path. Free the path vector.
8033 (mark_threaded_blocks): Iterate over the vector of paths and store
8034 the path on the appropriate edge. Get edges and block types from the
8035 jump threading path.
8036 (mark_threaded_blocks): Get edges and block types from the jump
8037 threading path. Free the path vector.
8038 (thread_through_all_blocks): Use the vector of paths rather than
8039 a vector of 3-edge sets.
8040 (register_jump_thread): Accept pointer to a path vector rather
8041 than the path vector itself. Store the path vector for later use.
8044 2013-10-01 Jakub Jelinek <jakub@redhat.com>
8045 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8048 * config/s390/s390.c (s390_split_branches): Modify check for table
8050 (s390_chunkify_start): Rearrange table jump insn check in order to
8051 deal with compare and branch insns correctly.
8053 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
8057 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
8058 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
8059 define_insn_and_split.
8060 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
8061 (shiftsi3_compare): New pattern.
8063 * config/arm/unspecs.md (UNSPEC_RRX): New.
8065 2013-10-01 Alan Modra <amodra@gmail.com>
8067 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
8068 casing inout operands.
8070 2013-10-01 Richard Biener <rguenther@suse.de>
8072 PR tree-optimization/58553
8073 * tree-loop-distribution.c (struct partition_s): Add niter member.
8074 (classify_partition): Populate niter member for the partition
8075 and properly identify whether the relevant store happens before
8076 or after the loop exit.
8077 (generate_memset_builtin): Use niter member from the partition.
8078 (generate_memcpy_builtin): Likewise.
8080 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8082 * vec.h (vec_prefix, vec): Prefix member names with "m_".
8083 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
8085 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
8087 * basic-block.h (edge_list): Prefix member names with "m_".
8088 * context.h (context): Likewise.
8089 * domwalk.h (dom_walker): Likewise.
8090 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
8091 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
8092 * hash-table.h (hash_table): Likewise.
8093 * machmode.h (bit_field_mode_iterator): Likewise.
8094 * pass_manager.h (pass_list): Likewise.
8095 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
8096 * tree-pass.h (pass_data): Likewise.
8097 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
8098 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
8099 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
8100 * asan.c (pass_data_asan): Update accordingly.
8101 * cfganal.c (control_dependences::find_control_dependence): Likewise.
8102 (control_dependences::control_dependences): Likewise.
8103 (control_dependences::~control_dependences): Likewise.
8104 (control_dependences::~control_dependences): Likewise.
8105 (control_dependences::get_edges_dependent_on): Likewise.
8106 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
8107 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
8108 * context.c (gcc::context::context): Likewise.
8109 * cprop.c (pass_rtl_cprop::clone): Likewise.
8110 * domwalk.c (dom_walker::walk): Likewise.
8111 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
8112 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
8113 * mode-switching.c (pass_mode_switching::clone): Likewise.
8114 * passes.c (opt_pass::opt_pass): Likewise.
8115 (pass_manager::pass_manager): Likewise.
8116 * predict.c (pass_strip_predict_hints::clone): Likewise.
8117 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
8118 (pass_split_all_insns::clone): Likewise.
8119 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
8121 (bit_field_mode_iterator::next_mode): Likewise.
8122 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
8123 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
8124 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
8125 * tree-complex.c (pass_lower_complex::clone): Likewise.
8126 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
8127 * tree-object-size.c (pass_object_sizes::clone): Likewise.
8128 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
8129 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
8130 (pass_fold_builtins::clone): Likewise.
8131 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
8132 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
8133 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
8134 (pass_cd_dce::clone): Likewise.
8135 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
8136 (pass_phi_only_cprop::clone): Likewise.
8137 * tree-ssa-dse.c (pass_dse::clone): Likewise.
8138 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
8139 * tree-ssa-loop.c (pass_lim::clone): Likewise.
8140 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
8141 * tree-ssa-pre.c (pass_fre::clone): Likewise.
8142 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
8143 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
8144 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
8145 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
8146 * tree-vrp.c (pass_vrp::clone): Likewise.
8147 * tsan.c (pass_tsan::clone): Likewise.
8149 2013-09-30 Jakub Jelinek <jakub@redhat.com>
8152 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
8153 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
8156 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
8157 optimization, punt if sign_bit_p looked through any zero extension.
8159 2013-09-30 Teresa Johnson <tejohnson@google.com>
8161 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
8162 Update redirected out edge count in joiner case.
8163 (ssa_redirect_edges): Common the joiner and non-joiner cases
8164 so that joiner case gets profile updates.
8166 2013-09-30 Richard Biener <rguenther@suse.de>
8168 PR tree-optimization/58554
8169 * tree-loop-distribution.c (classify_partition): Require
8170 unconditionally executed stores for memcpy and memset recognition.
8171 (tree_loop_distribution): Calculate dominance info.
8173 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
8175 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
8176 (NO_PROFILE_COUNTERS): Likewise.
8177 (PROFILE_HOOK): Likewise.
8178 (FUNCTION_PROFILER): Likewise.
8179 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
8181 2013-09-30 Iain Sandoe <iain@codesourcery.com>
8183 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
8184 calls and defines in TARGET_MACHO conditional.
8185 (load_macho_picbase_di): Likewise.
8186 (reload_macho_picbase): Likewise.
8187 (reload_macho_picbase_si): Likewise.
8188 (reload_macho_picbase_di): Likewise.
8189 (nonlocal_goto_receiver): Likewise.
8191 2013-09-30 Nick Clifton <nickc@redhat.com>
8193 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
8194 that use the MSP430X ISA.
8195 (msp430_option_override): Scan -mmcu command line option for any
8196 MCU name that supports the MSP430X ISA.
8197 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
8198 -mmcu options which enable the MSP430X ISA.
8200 2013-09-30 Richard Biener <rguenther@suse.de>
8203 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
8204 before looking for setjmp-like calls.
8206 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8209 * config/darwin-protos.h (machopic_get_function_picbase): New.
8210 * config/darwin.c (machopic_get_function_picbase): New.
8211 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
8212 label for a new func. (load_macho_picbase_di): Likewise.
8213 (reload_macho_picbase): New expand.
8214 (reload_macho_picbase_si): New insn.
8215 (reload_macho_picbase_di): New insn.
8216 (nonlocal_goto_receiver): New define and split.
8217 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
8218 (unspecv enum): Add UNSPECV_NLGR.
8220 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8222 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
8223 that altivec registers are correctly sized on Darwin.
8225 2013-09-29 Iain Sandoe <iain@codesourcery.com>
8227 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
8228 darwin-driver.o): Use COMPILE and POSTCOMPILE.
8229 * config/x-darwin (host-darwin.o): Likewise.
8230 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
8231 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
8232 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
8234 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
8236 * doc/invoke.texi: Fix usage of @tie{} command.
8238 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
8240 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
8242 2013-09-28 Jan Hubicka <jh@suse.cz>
8244 * config/i386/x86-tune.def: Add documentation for each of the options;
8247 2013-09-28 Jan Hubicka <jh@suse.cz>
8249 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
8251 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
8252 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
8253 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
8254 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
8255 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
8257 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
8259 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
8260 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
8261 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
8262 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
8263 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
8264 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
8265 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
8266 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
8267 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
8268 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
8269 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
8270 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
8271 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
8272 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
8273 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
8274 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
8275 gimple.h, godump.c, graphite-clast-to-gimple.c,
8276 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
8277 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
8278 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
8279 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
8280 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
8281 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
8282 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
8283 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
8284 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
8285 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
8286 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
8287 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
8288 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
8289 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
8290 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
8291 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
8292 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
8293 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
8294 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
8295 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
8296 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
8297 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
8298 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
8299 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
8300 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
8301 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
8302 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
8303 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
8304 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
8305 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
8306 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
8307 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
8308 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
8309 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
8310 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8311 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
8312 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
8313 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
8314 whitespace before "(".
8316 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
8318 * expr.h (extract_bit_field): Remove packedp parameter.
8319 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
8320 from forward declaration.
8321 (store_split_bit_field): Remove packedp arg from calls to
8322 extract_fixed_bit_field.
8323 (extract_bit_field_1): Remove packedp parameter and packedp
8324 argument from recursive calls and calls to extract_fixed_bit_field.
8325 (extract_bit_field): Remove packedp parameter and corresponding
8326 arg to extract_bit_field_1.
8327 (extract_fixed_bit_field): Remove packedp parameter. Remove code
8329 (extract_split_bit_field): Remove packedp arg from call to
8330 extract_fixed_bit_field.
8331 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
8332 (copy_blkmode_from_reg): Likewise.
8333 (copy_blkmode_to_reg): Likewise.
8334 (read_complex_part): Likewise.
8335 (store_field): Likewise.
8336 (expand_expr_real_1): Likewise.
8337 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
8338 to extract_bit_field.
8339 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
8340 call to extract_bit_field.
8341 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
8342 call to extract_bit_field.
8343 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
8344 and special packedp behavior from -fstrict-volatile-bitfields
8347 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8349 * lra-eliminations.c (init_elim_table): Guard value_p.
8351 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
8353 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
8354 DFmode, DImode, and SFmode in the upper VSX registers based on the
8355 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
8356 if -mpower8-vector. Combine -mvsx-timode handling with the rest
8357 of the VSX register handling.
8359 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
8361 (zero_extendsidi2_lfiwzx): Add support for loading into the
8362 Altivec registers with -mpower8-vector. Use wu/wv constraints to
8363 only do VSX memory options on Altivec registers.
8364 (extendsidi2_lfiwax): Likewise.
8365 (extendsfdf2_fpr): Likewise.
8366 (mov<mode>_hardfloat, SF/SD modes): Likewise.
8367 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
8368 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
8369 (movdi_internal64): Likewise.
8371 2013-09-27 Xinliang David Li <davidxl@google.com>
8373 * opts.c (finish_options): Adjust parameters
8374 according to vect cost model.
8375 (common_handle_option): Set dynamic vect cost
8377 targhooks.c (default_add_stmt_cost): Compute stmt cost
8379 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
8380 Use helper function.
8381 * tree-vectorizer.h (unlimited_cost_model): New function.
8382 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
8383 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
8385 (vect_enhance_data_refs_alignment): Ditto.
8386 * flag-types.h: New enum.
8387 * common/config/i386/i386-common.c (ix86_option_init_struct):
8388 No need to initialize vect_cost_model flag.
8389 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
8392 2013-09-27 Diego Novillo <dnovillo@google.com>
8394 * gimple.h (enum ssa_mode): Remove.
8396 2013-09-27 Paulo Matos <pmatos@broadcom.com>
8398 * cfgloop.h (number_of_loops): Fix typo in check for null.
8400 2013-09-27 Jakub Jelinek <jakub@redhat.com>
8403 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
8404 are children of outermost saved_cfun's loop, and set it up to
8405 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
8406 if loop != loop0 and SESE region contains bbs that belong to loop0.
8408 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
8410 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
8411 (binary_scale_code_p, get_base_term, get_index_term): New functions.
8412 (set_address_segment, set_address_base, set_address_index)
8413 (set_address_disp): Accept the argument unconditionally.
8414 (baseness): Remove must_be_base_p and must_be_index_p checks.
8415 (decompose_normal_address): Classify as much as possible in the
8418 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
8420 * cse.c (count_reg_usage): Handle INT_LIST.
8421 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
8422 * reginfo.c (reg_scan_mark_refs): Likewise.
8423 * reload1.c (eliminate_regs_1): Likewise.
8425 2013-09-27 Iain Sandoe <iain@codesourcery.com>
8428 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
8431 2013-09-27 Richard Biener <rguenther@suse.de>
8433 PR tree-optimization/58459
8434 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
8435 restriction not propagating into loops.
8437 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
8439 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
8440 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
8441 * doc/tree-ssa.texi (Walking use-def chains): Delete.
8443 2013-09-26 Richard Biener <rguenther@suse.de>
8445 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
8447 2013-09-26 Richard Biener <rguenther@suse.de>
8449 * alias.h (component_uses_parent_alias_set): Rename to ...
8450 (component_uses_parent_alias_set_from): ... this.
8451 * alias.c (component_uses_parent_alias_set): Rename to ...
8452 (component_uses_parent_alias_set_from): ... this and return
8454 (reference_alias_ptr_type_1): Use the result from
8455 component_uses_parent_alias_set_from instead of stripping
8456 components one at a time.
8457 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
8459 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
8461 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
8462 Move prototypes to...
8463 * tree-ssa-ter.h: New File. Move prototypes here.
8464 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
8465 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
8466 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
8467 from is_replaceable_p.
8468 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
8469 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
8470 (process_replaceable): Use ter_is_replaceable_p.
8471 (find_replaceable_in_bb): Use ter_is_replaceable_p.
8472 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
8473 newly refactored ssa_is_replaceable_p.
8474 * cfgexpand.c: Include tree-outof-ssa.h.
8475 * ssaexpand.h: Delete.
8477 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
8479 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
8480 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
8482 (create_gimple_tmp): Delete.
8483 (get_expr_type, build_assign, build_type_cast): Move to...
8484 * gimple-builder.c: New File.
8485 (get_expr_type): Relocate from gimple.c.
8486 (build_assign, build_type_cast): Change to only create ssanames.
8487 * gimple.h: Move prototypes to...
8488 * gimple-builder.h: New File. Here.
8489 * tree-ssa.h: And here.
8490 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
8491 count_uses_and_derefs): Relocate from gimple.c.
8492 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
8493 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
8494 * tree-ssa-math-opts (execute_cse_reciprocals): Use
8495 gimple_replace_ssa_lhs.
8496 * asan.c: Include gimple-builder.h.
8497 * Makefile.in: Add gimple-builder.o.
8499 2013-09-26 Richard Biener <rguenther@suse.de>
8501 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
8502 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
8503 (loe_visit_block): Use gcc_checking_assert.
8504 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
8505 gimple_assign_ssa_name_copy_p.
8506 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
8509 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
8511 * config/rs6000/t-rs6000 (rs6000.o): Remove.
8512 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
8514 2013-09-26 Richard Biener <rguenther@suse.de>
8516 PR tree-optimization/58539
8517 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
8518 the fact that debug statements are not taking part in loop-closed
8521 2013-09-26 Nick Clifton <nickc@redhat.com>
8523 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
8524 time warning message.
8525 (msp430_print_operand_raw): Delete unused letter parameter.
8526 (TARGET_PRINT_OPERAND_ADDRESS): Define.
8527 (msp430_print_operand_address): New function.
8528 (msp430_print_operand): Move address printing code from here to
8530 * config/msp430/msp430.md (movsipsi2): Add comment in generated
8532 (zero_extendpsisi2): Likewise.
8533 (extendpsisi2): New pattern.
8534 (andneghi3): New pattern.
8536 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
8538 * config/aarch64/aarch64.opt (mlra): New option.
8539 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
8540 (TARGET_LRA_P): Define.
8542 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
8544 * expr.c (expand_assignment): Remove obsolete comment.
8546 2013-09-25 Jeff Law <law@redhat.com>
8548 * tree-flow.h (thread_through_all_blocks): Prototype moved into
8549 tree-ssa-threadupdate.h.
8550 (register_jump_thread): Similarly.
8551 * tree-ssa-threadupdate.h: New header file.
8552 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
8553 * tree-vrp.c: Likewise.
8554 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
8555 (thread_around_empty_blocks): Change type of path vector argument to
8556 an edge,type pair from just an edge. Initialize both elements when
8557 appending to a jump threading path. Tweak references to elements
8559 (thread_across_edge): Similarly. Release memory for the elements
8561 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
8562 (dump_jump_thread_path): New function broken out from
8563 register_jump_thread.
8564 (register_jump_thread): Use dump_jump_thread_path. Change type of
8565 path vector entries. Search the path for NULL edges and dump
8566 the path if one is found. Tweak the conversion of path to 3-edge
8567 form to use the block copy type information embedded in the path.
8569 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
8571 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
8573 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
8574 Vladimir Makarov <vmakarov@redhat.com>
8576 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
8577 from the least significant bit.
8578 (strip_address_mutations): Add bitfield operations handling.
8579 (must_be_index_p): Add shifting and rotate operations handling.
8580 (set_address_base): Use must_be_base_p predicate.
8581 (set_address_index): Use must_be_index_p predicate.
8583 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
8584 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8585 Sergey Lega <sergey.s.lega@intel.com>
8586 Anna Tikhonova <anna.tikhonova@intel.com>
8587 Ilya Tocar <ilya.tocar@intel.com>
8588 Andrey Turetskiy <andrey.turetskiy@intel.com>
8589 Ilya Verbin <ilya.verbin@intel.com>
8590 Kirill Yukhin <kirill.yukhin@intel.com>
8591 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8593 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
8595 (ix86_expand_vector_move_misalign): Support new unaligned load and
8596 stores and use new names.
8597 (CODE_FOR_sse2_storedqu): Rename to ...
8598 (CODE_FOR_sse2_storedquv16qi): ... this.
8599 (CODE_FOR_sse2_loaddqu): Rename to ...
8600 (CODE_FOR_sse2_loaddquv16qi): ... this.
8601 (CODE_FOR_avx_loaddqu256): Rename to ...
8602 (CODE_FOR_avx_loaddquv32qi): ... this.
8603 (CODE_FOR_avx_storedqu256): Rename to ...
8604 (CODE_FOR_avx_storedquv32qi): ... this.
8605 * config/i386/i386.md (fpint_logic): New.
8606 * config/i386/sse.md (VMOVE): Extend for AVX512.
8610 (VI_UNALIGNED_LOADSTORE): Ditto.
8611 (sse2_avx_avx512f): Ditto.
8612 (sse2_avx2): Extend for AVX512.
8613 (sse4_1_avx2): Ditto.
8614 (avx2_avx512f): New.
8615 (sse): Extend for AVX512.
8618 (avxsizesuffix): Ditto.
8619 (sseintvecmode): Ditto.
8621 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
8622 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
8623 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
8624 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
8625 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
8626 (<sse2_avx_avx512f>_storedqu<mode): ... this.
8627 (<sse>_movnt<mode>): Replace constraint "x" with "v".
8628 (STORENT_MODE): Extend for AVX512.
8629 (*absneg<mode>2): Replace constraint "x" with "v".
8630 (*mul<mode>3): Ditto.
8631 (*ieee_smin<mode>3): Ditto.
8632 (*ieee_smax<mode>3): Ditto.
8633 (avx_cmp<mode>3): Replace VF with VF_128_256.
8634 (*<sse>_maskcmp<mode>3_comm): Ditto.
8635 (<sse>_maskcmp<mode>3): Ditto.
8636 (<sse>_andnot<mode>3): Extend for AVX512.
8637 (<code><mode>3, anylogic): Replace VF with VF_128_256.
8638 (<code><mode>3, fpint_logic): New.
8639 (*<code><mode>3): Extend for AVX512.
8640 (avx512flogicsuff): New.
8641 (avx512f_<logic><mode>): Ditto.
8642 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
8644 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
8645 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
8646 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
8647 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
8648 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
8649 (xop_vpermil2<mode>3): Ditto.
8650 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
8651 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
8652 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
8653 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
8655 2013-09-25 Tom Tromey <tromey@redhat.com>
8657 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
8658 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
8659 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
8660 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
8661 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
8662 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
8663 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
8664 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
8665 (GRAPHITE_HTAB_H): Remove.
8667 2013-09-25 Tom Tromey <tromey@redhat.com>
8669 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
8671 2013-09-25 Tom Tromey <tromey@redhat.com>
8673 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
8675 2013-09-25 Tom Tromey <tromey@redhat.com>
8677 * config/i386/t-i386 (i386.o): Remove.
8678 (i386-c.o): Use COMPILE and POSTCOMPILE.
8680 2013-09-25 Tom Tromey <tromey@redhat.com>
8682 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
8684 2013-09-25 Tom Tromey <tromey@redhat.com>
8686 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
8687 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
8688 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
8689 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
8690 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
8691 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
8692 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
8693 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
8694 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
8695 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
8696 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
8697 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
8698 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
8699 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
8700 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
8701 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
8702 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
8703 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
8704 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
8705 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
8706 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
8707 (tree-ssa-pre.o, tree-ssa-sccvn.o)
8708 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
8709 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
8710 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
8711 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
8712 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
8713 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
8714 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
8715 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
8716 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
8717 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
8718 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
8719 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
8720 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
8721 (graphite-clast-to-gimple.o, graphite-dependences.o)
8722 (graphite-interchange.o, graphite-poly.o)
8723 (graphite-scop-detection.o, graphite-sese-to-poly.o)
8724 (graphite-optimize-isl.o, tree-vect-loop.o)
8725 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
8726 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
8727 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
8728 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
8729 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
8730 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
8731 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
8732 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
8733 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
8734 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
8735 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
8736 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
8737 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
8738 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
8739 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
8740 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
8741 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
8742 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
8743 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
8744 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
8745 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
8746 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
8747 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
8748 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
8749 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
8750 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
8751 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
8752 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
8753 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
8754 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
8755 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
8756 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
8757 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
8758 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
8759 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
8760 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
8761 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
8762 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
8763 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
8764 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
8765 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
8766 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
8767 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
8768 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
8769 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
8770 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
8771 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
8772 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
8773 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
8774 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
8775 (gcov-dump.o): Remove.
8776 (default-c.o): Use COMPILE and POSTCOMPILE.
8777 (CFLAGS-gcc.o): New variable.
8778 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
8780 2013-09-25 Tom Tromey <tromey@redhat.com>
8782 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
8783 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
8784 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
8785 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
8786 (c-family/c-pragma.o, c-family/c-pretty-print.o)
8787 (c-family/c-semantics.o, c-family/c-ada-spec.o)
8788 (c-family/array-notation-common.o, c-family/stub-objc.o)
8789 (c-family/c-ubsan.o): Remove.
8791 2013-09-25 Tom Tromey <tromey@redhat.com>
8793 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
8795 2013-09-25 Tom Tromey <tromey@redhat.com>
8797 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
8798 to add -DENABLE_SHARED_LIBGCC.
8799 (gcc.o): Don't use subshell.
8801 2013-09-25 Tom Tromey <tromey@redhat.com>
8803 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
8804 * configure.ac: Don't invoke AM_PROG_CC_C_O.
8805 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
8806 * configure, config.in: Rebuild.
8808 2013-09-25 Tom Tromey <tromey@redhat.com>
8810 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
8811 (COMPILE, POSTCOMPILE): New variables.
8812 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
8813 (DEPFILES): New variable.
8814 Include ".Po" files.
8815 * configure.ac: Add checks for dependency checking.
8816 * configure, aclocal.m4: Regenerate.
8818 2013-09-25 Tom Tromey <tromey@redhat.com>
8820 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
8821 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
8823 2013-09-25 Tom Tromey <tromey@redhat.com>
8825 * Makefile.in (generated_files): Add options.h,
8826 target-hooks-def.h, insn-opinit.h,
8827 common/common-target-hooks-def.h, pass-instances.def,
8828 c-family/c-target-hooks-def.h.
8830 2013-09-25 Jeff Law <law@redhat.com>
8832 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
8833 than foo[foo.length () - 1] to access last member in a vec.
8834 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
8836 2013-09-25 Richard Biener <rguenther@suse.de>
8839 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
8841 2013-09-25 Jan Hubicka <jh@suse.cz>
8843 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
8846 2013-09-25 Marek Polacek <polacek@redhat.com>
8849 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
8850 when determining the type name.
8852 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
8854 * config/sh/sh.md: Fix formatting.
8856 2013-09-24 Xinliang David Li <davidxl@google.com>
8858 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
8859 max peel iterations parameter.
8860 * param.def: New parameter.
8861 * doc/invoke.texi: Document New parameter.
8863 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
8865 * gimple-pretty-print.c: Various whitespace tweaks.
8866 * tree-core.h: Likewise.
8867 * tree-pretty-print.c: Likewise.
8868 * tree-ssa-alias.c: Likewise.
8869 * tree-ssa-copy.c: Likewise.
8870 * tree-ssanames.c: Likewise.
8871 * tree-ssanames.h: Likewise.
8872 * tree-vrp.c: Likewise.
8874 2013-09-24 Alan Modra <amodra@gmail.com>
8878 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
8879 for output operands that disallow regs. Don't use EXPAND_WRITE on
8882 2013-09-24 Richard Biener <rguenther@suse.de>
8885 * tree.c (reference_alias_ptr_type): Move ...
8886 * alias.c (reference_alias_ptr_type): ... here and implement
8887 in terms of the new reference_alias_ptr_type_1.
8888 (ref_all_alias_ptr_type_p): New helper.
8889 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
8890 use ref_all_alias_ptr_type_p.
8891 (get_deref_alias_set): Add flag_strict_aliasing check here.
8892 (reference_alias_ptr_type_1): New function, split out from ...
8893 (get_alias_set): ... here.
8894 (alias_ptr_types_compatible_p): New function.
8895 * alias.h (reference_alias_ptr_type): Declare.
8896 (alias_ptr_types_compatible_p): Likewise.
8897 * tree.h (reference_alias_ptr_type): Remove.
8898 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
8899 to compare MEM_REF alias types.
8901 2013-09-24 Richard Biener <rguenther@suse.de>
8903 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
8905 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8907 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
8908 reload helper function arrays into a single array reg_addr.
8909 (reload_fpr_gpr): Likewise.
8910 (reload_gpr_vsx): Likewise.
8911 (reload_vsx_gpr): Likewise.
8912 (struct rs6000_reg_addr): Likewise.
8913 (reg_addr): Likewise.
8914 (rs6000_debug_reg_global): Change rs6000_vector_reload,
8915 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
8916 (rs6000_init_hard_regno_mode_ok): Likewise.
8917 (rs6000_secondary_reload_direct_move): Likewise.
8918 (rs6000_secondary_reload): Likewise.
8920 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
8921 constraints: wu, ww, and wy. Repurpose wv constraint added during
8922 power8 changes. Put wg constraint in alphabetical order.
8924 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
8925 for future work to add ISA 2.07 VSX single precision support.
8926 (-mvsx-scalar-double): Change default from -1 to 1, update
8927 documentation comment.
8928 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
8929 (-mupper-regs-df): New debug switch to control whether DF values
8930 can go in the traditional Altivec registers.
8931 (-mupper-regs-sf): New debug switch to control whether SF values
8932 can go in the traditional Altivec registers.
8934 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
8936 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
8937 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
8938 Add new constraints, wu/ww/wy. Repurpose wv constraint.
8939 (rs6000_debug_legitimate_address_p): Print if we are running
8940 before, during, or after reload.
8941 (rs6000_secondary_reload): Add a comment.
8942 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
8944 * config/rs6000/constraints.md (wa constraint): Sort w<x>
8945 constraints. Update documentation string.
8946 (wd constraint): Likewise.
8947 (wf constraint): Likewise.
8948 (wg constraint): Likewise.
8949 (wn constraint): Likewise.
8950 (ws constraint): Likewise.
8951 (wt constraint): Likewise.
8952 (wx constraint): Likewise.
8953 (wz constraint): Likewise.
8954 (wu constraint): New constraint for ISA 2.07 SFmode scalar
8956 (ww constraint): Likewise.
8957 (wy constraint): Likewise.
8958 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
8959 in the previous submissions.
8960 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
8962 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
8964 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
8965 (REG_BR_PROB): Say that the probability is stored in an int_list.
8966 * reg-notes.def: Update commentary to mention INT_LIST.
8967 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
8968 (INT_LIST): New rtx.
8969 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
8970 * rtlanal.c (int_reg_note_p): New function.
8971 (alloc_reg_note): Assert that the note does not have an int argument.
8972 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
8973 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
8974 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
8975 rather than an INSN_LIST. Handle INT_LIST.
8976 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
8977 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
8978 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
8979 Manipulate them as ints rather than rtxes.
8980 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
8981 * regmove.c (copy_src_to_dest): Likewise.
8982 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
8984 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
8985 into the cases that need it.
8986 * config/arm/arm.c (arm_unwind_emit): Likewise.
8988 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
8989 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
8990 * loop-doloop.c (add_test, doloop_modify): Likewise.
8991 * loop-unswitch.c (compare_and_jump_seq): Likewise.
8992 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
8993 * predict.c (combine_predictions_for_insn): Likewise.
8994 * print-rtl.c (print_rtx): Handle INT_LIST.
8995 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
8996 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
8997 * config/arm/arm.c (emit_unlikely_jump): Likewise.
8998 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
8999 (ix86_split_fp_branch, predict_jump): Likewise.
9000 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
9001 * config/sh/sh.c (expand_cbranchsi4): Likewise.
9002 * config/spu/spu.c (ea_load_store_inline): Likewise.
9004 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
9005 value of a REG_BR_PROB note.
9006 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
9007 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
9008 * emit-rtl.c (try_split): Likewise.
9009 * predict.c (br_prob_note_reliable_p): Likewise.
9010 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
9011 * reorg.c (mostly_true_jump): Likewise.
9012 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
9013 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
9014 * config/i386/i386.c (ix86_print_operand): Likewise.
9015 * config/ia64/ia64.c (ia64_print_operand): Likewise.
9016 * config/mmix/mmix.c (mmix_print_operand): Likewise.
9017 * config/rs6000/rs6000.c (output_cbranch): Likewise.
9018 * config/s390/s390.c (s390_expand_tbegin): Likewise.
9019 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
9020 * config/sparc/sparc.c (output_cbranch): Likewise.
9021 * config/spu/spu.c (get_branch_target): Likewise.
9022 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
9023 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
9025 2013-09-23 Jan Hubicka <jh@suse.cz>
9027 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
9029 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
9030 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
9032 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
9033 (gimple_fold_call): Dump inconsistent devirtualizations; add
9034 sanity check for type based devirtualizations.
9035 * ipa-prop.c: Include ipa-utils.h
9036 (ipa_intraprocedural_devirtualization): Add sanity check.
9037 (try_make_edge_direct_virtual_call): Likewise.
9039 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
9041 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
9042 assignment statements.
9044 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
9046 * gimple-pretty-print.c (dump_ssaname_info): New function.
9047 (dump_gimple_phi): Call it.
9048 (pp_gimple_stmt_1): Likewise.
9049 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
9050 (range_info_def): Declare.
9051 * tree-pretty-print.c (pp_double_int): New function.
9052 (dump_generic_node): Call it.
9053 * tree-pretty-print.h (pp_double_int): Declare.
9054 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
9055 * tree-ssanames.h (range_info_def): New structure.
9056 (value_range_type): Move definition here.
9057 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
9059 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
9061 (set_range_info): New function.
9062 (get_range_info): Likewise.
9063 (duplicate_ssa_name_range_info): Likewise.
9064 (duplicate_ssa_name_fn): Check pointer type and call
9065 duplicate_ssa_name_range_info.
9066 * tree-ssa-copy.c (fini_copy_prop): Likewise.
9067 * tree-vrp.c (value_range_type): Remove definition, now in
9069 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
9070 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
9071 (SSA_NAME_RANGE_INFO): New macro.
9073 2013-09-23 Richard Biener <rguenther@suse.de>
9075 PR tree-optimization/58464
9076 * tree-ssa-pre.c (phi_trans_lookup): Remove.
9077 (phi_trans_add): Change to add conditionally on being not present.
9078 (phi_translate_1): Remove recursion detection here.
9079 (phi_translate): Pre-seed the cache with NULL to catch
9080 recursion here in a more generic way.
9081 (bitmap_find_leader): Adjust comment.
9082 (get_representative_for): Dump value-numbers.
9083 (create_expression_by_pieces): Likewise.
9084 (insert_into_preds_of_block): Likewise.
9086 2013-09-23 Christian Bruel <christian.bruel@st.com>
9089 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
9090 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
9092 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
9095 2013-09-20 Renlin Li <renlin.li@arm.com>
9097 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
9099 (aarch64_expand_epilogue): Likewise.
9100 (aarch64_legitimize_reload_address): Likewise.
9102 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
9104 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
9105 NULL_TREE before pushing them onto the vector. Likewise for labels.
9107 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
9109 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
9110 comparison operators when -fno-trapping-math is in effect.
9111 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
9112 comparison operators in TFmode and assert that unsupported operators
9114 (ia64_print_operand): Likewise.
9116 2013-09-21 Jan Hubicka <jh@suse.cz>
9118 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
9119 (sse_typeless_stores): Enable for core
9120 (sse_load0_by_pxor): Likewise.
9121 (four_jump_limit): Disable for core.
9122 (pad_returns): Likewise.
9123 (avoid_vector_decode): Likewise.
9124 (fuse_cmp_and_branch): Enable for cores.
9125 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
9127 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
9130 * config/pa/pa.c (pa_option_override): Disable auto increment and
9131 decrement instructions until reload is completed.
9133 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
9134 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
9135 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
9137 2013-09-20 DJ Delorie <dj@redhat.com>
9138 Nick Clifton <nickc@redhat.com>
9140 * config/rl78/rl78.c: Various whitespace and comment tweaks.
9141 (need_to_save): Save bank 0 on interrupts.
9142 (characterize_address): Strip far address wrappers.
9143 (rl78_as_legitimate_address): Likewise.
9144 (transcode_memory_rtx): Likewise.
9145 (rl78_peep_movhi_p): Disable this peephole after devirt.
9146 (rl78_propogate_register_origins): Forget all origins when a
9148 * config/rl78/rl78-virt.md: Various whitespace tweaks.
9149 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
9151 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
9152 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
9153 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
9155 * config/rl78/constraints.md: Various whitespace and paren tweaks.
9157 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
9159 * config/pa/pa.md: In "scc" insn patterns, change output template to
9160 handle const0_rtx in reg_or_0_operand operands.
9162 2013-09-20 Martin Husemann <martin@NetBSD.org>
9165 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
9166 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
9168 2013-09-20 Richard Biener <rguenther@suse.de>
9171 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
9172 remembering SSA name version and block index.
9173 (new_scev_info_str): Adjust.
9174 (hash_scev_info): Likewise. Also hash the block index.
9175 (eq_scev_info): Adjust.
9176 (find_var_scev_info): Likewise.
9177 (struct instantiate_cache_entry): Remove.
9178 (struct instantiate_cache_type): Use a htab to map name, block
9180 (instantiate_cache_type::~instantiate_cache_type): Adjust.
9181 (get_instantiated_value_entry): Likewise.
9182 (hash_idx_scev_info, eq_idx_scev_info): New functions.
9183 (instantiate_scev_name): Adjust.
9185 2013-09-20 Jeff Law <law@redhat.com>
9187 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
9189 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
9191 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
9192 Call aarch64_simd_expand_args to update op[argc].
9194 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
9196 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
9199 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
9201 * gengtype.c (file_rules): Added rule for *.cc files.
9202 (get_output_file_with_visibility): Give fatal message when no
9205 2013-09-20 Renlin Li <renlin.li@arm.com>
9207 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
9208 (aarch64_expand_epilogue): Likewise.
9209 (aarch64_legitimize_reload_address): Likewise.
9211 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
9214 * expr.c (expand_assignment): Remove misalignp code path.
9216 2013-09-20 Marek Polacek <polacek@redhat.com>
9219 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
9220 TYPE_PRECISION. Add asserts.
9222 2013-09-20 Richard Biener <rguenther@suse.de>
9224 PR tree-optimization/58453
9225 * tree-loop-distribution.c (distribute_loop): Apply the cost
9226 model for -ftree-loop-distribute-patterns, too.
9228 2013-09-20 Richard Biener <rguenther@suse.de>
9231 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
9232 make type comparison less strict.
9234 2013-09-20 Alan Modra <amodra@gmail.com>
9236 * configure: Regenerate.
9237 * aclocal.m4: Regenerate.
9239 2013-09-20 Marek Polacek <polacek@redhat.com>
9242 * doc/extend.texi: Document that attribute used is meant to be used
9243 on variables with static storage duration.
9245 2013-09-19 Jakub Jelinek <jakub@redhat.com>
9247 PR tree-optimization/58472
9248 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
9249 simd_lane_access set inv_p = false.
9250 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
9251 the simduid magic VAR_DECL.
9253 2013-09-19 Jan Hubicka <jh@suse.cz>
9255 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
9257 2013-09-17 Jeff Law <law@redhat.com>
9259 * tree-ssa-dom.c (record_temporary_equivalences): New function
9260 split out of dom_opt_dom_walker::after_dom_children.
9261 (dom_opt_dom_walker::thread_across_edge): Move common code
9262 in here from dom_opt_dom_walker::after_dom_children.
9263 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
9265 2013-09-19 Jan Hubicka <jh@suse.cz>
9267 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
9268 (TARGET_GENERIC): Use PROCESOR_GENERIC
9269 (enum processor_type): Unify generic32 and 64.
9270 * i386.md (cpu): Likewise.
9271 * x86-tune.def (use_leave): Enable for generic32.
9272 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
9273 * athlon.md: Change generic64 to generic in all occurences.
9274 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
9275 (ix86_target_macros_internal): Likewise.
9276 * driver-i386.c (host_detect_local_cpu): Likewise.
9277 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
9279 (generic_memcpy, generic_memset, generic_cost): This one.
9280 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
9281 (m_GENERIC32, m_GENERIC64): Remove.
9282 (m_GENERIC): Turn into one flag.
9283 (processor_target): Unify generic tunnings.
9284 (ix86_option_override_internal): Replace generic32/64 by generic.
9285 (ix86_issue_rate): Likewise.
9286 (ix86_adjust_cost): Likewise.
9288 2013-09-19 Jan Hubicka <jh@suse.cz>
9290 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
9291 of speculative flag.
9293 2013-09-19 Jakub Jelinek <jakub@redhat.com>
9295 * omp-low.c (expand_omp_sections): Always pass len - 1 to
9296 GOMP_sections_start, even if !exit_reachable.
9298 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
9300 * lra-constraints.c (need_for_all_save_p): Use macro
9301 HARD_REGNO_CALL_PART_CLOBBERED.
9302 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
9303 set up pseudo conflict hard regs.
9305 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9308 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
9311 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
9313 PR rtl-optimization/58438
9314 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
9315 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
9316 from previous subpasses.
9318 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
9320 * arm.c (arm_get_frame_offsets): Validate architecture supports
9321 LDRD/STRD before accepting the tuning preference.
9322 (arm_expand_prologue): Likewise.
9323 (arm_expand_epilogue): Likewise.
9325 2013-09-18 Richard Biener <rguenther@suse.de>
9327 PR tree-optimization/58417
9328 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
9329 have chrecs with symbols defined in the loop as operands.
9330 (chrec_fold_multiply): Likewise.
9331 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
9332 parameters before folding binary operations.
9333 (struct instantiate_cache_entry_hasher): Remove.
9334 (struct instantiate_cache_type): Use a pointer-map.
9335 (instantiate_cache_type::instantiate_cache_type): New function.
9336 (instantiate_cache_type::get): Likewise.
9337 (instantiate_cache_type::set): Likewise.
9338 (instantiate_cache_type::~instantiate_cache_type): Adjust.
9339 (get_instantiated_value_entry): Likewise.
9340 (global_cache): New global.
9341 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
9342 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
9343 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
9344 (instantiate_scev_name): Adjust.
9345 (instantiate_scev): Construct global instead of local cache.
9346 (resolve_mixers): Likewise.
9348 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
9349 Paolo Carlini <paolo.carlini@oracle.com>
9352 * doc/implement-cxx.texi: Fix references to the C++ standards.
9354 2013-09-18 Jakub Jelinek <jakub@redhat.com>
9356 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
9357 * tree-vect-data-refs.c (vect_analyze_data_refs): For
9358 simd_lane_access drs, update also DR_ALIGNED_TO.
9360 2013-09-18 Marek Polacek <polacek@redhat.com>
9363 * doc/extend.texi: Document no_sanitize_undefined attribute.
9364 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
9365 no_sanitize_undefined attribute.
9367 2013-09-18 Nick Clifton <nickc@redhat.com>
9369 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
9370 (ASM_DECLARE_FUNCTION_NAME): Define.
9372 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
9374 * compare-elim.c (find_comparison_dom_walker): New class
9375 (find_comparisons_in_bb): Rename to
9376 find_comparison_dom_walker::before_dom_children
9377 (find_comparisons): Adjust
9378 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
9380 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
9381 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
9382 (init_walk_dominator_tree): Remove declaration.
9383 (fini_walk_dominator_tree): Remove declaration.
9384 * fwprop.c (single_def_use_dom_walker): New class
9385 (single_def_use_enter_block): Convert to
9386 single_def_use_dom_walker::before_dom_children.
9387 (single_def_use_leave_block): Convert to
9388 single_def_use_dom_walker::after_dom_children.
9389 (build_single_def_use_links): Adjust.
9390 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
9392 (find_candidates_in_block): Convert to
9393 find_candidates_dom_walker::before_dom_children.
9394 (execute_strength_reduction): Adjust.
9395 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
9396 (sese_dom_walker): New class.
9397 (sese_dom_walker::sese_dom_walker): New constructor.
9398 (sese_dom_walker::~sese_dom_walker): New destructor.
9399 (build_sese_conditions_before): Convert to
9400 sese_dom_walker::before_dom_children.
9401 (build_sese_conditions_after): Convert to
9402 sese_dom_walker::after_dom_children.
9403 (build_poly_scop): Adjust
9404 * tree-into-ssa.c (rewrite_dom_walker): New class
9405 (rewrite_enter_block): Convert to
9406 rewrite_dom_walker::before_dom_children.
9407 (rewrite_leave_block): Convert to
9408 rewrite_dom_walker::after_dom_children.
9409 (rewrite_update_dom_walker): New class.
9410 (rewrite_update_enter_block): Convert to
9411 rewrite_update_dom_walker::before_dom_children.
9412 (rewrite_update_leave_block): Convert to
9413 rewrite_update_dom_walker::after_dom_children.
9414 (rewrite_blocks, rewrite_into_ssa): Adjust.
9415 (mark_def_dom_walker): New class.
9416 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
9417 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
9418 (mark_def_sites_blocks): Convert to
9419 mark_def_dom_walker::before_dom_children.
9420 (mark_def_site_blocks): Remove.
9421 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
9422 (tree_ssa_dominator_optimize): Adjust.
9423 (dom_thread_across_edge): Convert to method
9424 dom_opt_dom_walker::thread_across_edge.
9425 (dom_opt_enter_block): Convert to member function
9426 dom_opt_dom_walker::before_dom_children.
9427 (dom_opt_leave_block): Convert to member function
9428 dom_opt_dom_walker::after_dom_children.
9429 * tree-ssa-dse.c (dse_dom_walker): New class.
9430 (dse_enter_block): Convert to member function
9431 dse_dom_walker::before_dom_children.
9432 (tree_ssa_dse): Adjust.
9433 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
9434 (determine_invariantness_stmt): Convert to method
9435 invariantness_dom_walker::before_dom_children.
9436 (determine_invariantness): Remove
9437 (move_computations_dom_walker): New class.
9438 (move_computations_stmt): Convert to method
9439 move_computations_dom_walker::before_dom_children.
9440 (move_computations, tree_ssa_lim): Adjust.
9441 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
9442 (nt_init_block): Convert to method
9443 notrappping_dom_walker::before_dom_children.
9444 (nt_fini_block): Convert to method
9445 method nontrapping_dom_walker::after_dom_children.
9446 (get_non_trapping): Adjust.
9447 * tree-ssa-pre.c (eliminate_dom_walker): New class.
9448 (eliminate_bb): Convert to method
9449 eliminate_dom_walker::before_dom_children.
9450 (eliminate_leave_block): Convert to method
9451 eliminate_dom_walker::after_dom_children.
9452 (eliminate): Adjust.
9453 * tree-ssa-strlen.c (strlen_dom_walker): New class.
9454 (strlen_enter_block): Convert to method
9455 strlen_dom_walker::before_dom_children.
9456 (strlen_leave_block): Convert to method
9457 method strlen_dom_walker::after_dom_children.
9458 (tree_ssa_strlen): Adjust.
9459 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
9460 (tree_ssa_uncprop): Adjust.
9461 (uncprop_leave_block): Convert to method
9462 uncprop_dom_walker::after_dom_children.
9463 (uncprop_leave_block): Convert to method
9464 uncprop_dom_walker::before_dom_children.
9466 2013-09-18 Bin Cheng <bin.cheng@arm.com>
9468 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
9469 branch in same basic block. Check both src and dest of the move insn.
9471 2013-09-17 Nick Clifton <nickc@redhat.com>
9473 * config/rl78/rl78-real.md (bf): New pattern.
9475 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
9476 (rl78_print_operand): Do not put a # before a %B.
9477 * config/rl78/rl78.opt: Tweak doc strings.
9479 2013-09-17 DJ Delorie <dj@redhat.com>
9481 * config/rl78/constraints.md (Wcv): Allow up to $r31.
9482 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
9483 (rl78_option_override): Likewise, if -mallregs.
9484 (is_virtual_register): Likewise.
9485 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
9486 (REGNO_OK_FOR_BASE_P): Likewise.
9487 * config/rl78/rl78.opt (-mallregs): New.
9489 2013-09-17 Nick Clifton <nickc@redhat.com>
9491 * config/rl78/rl78.c (need_to_save): Change return type to bool.
9492 For interrupt functions: save all call clobbered registers if the
9493 interrupt handler is not a leaf function.
9494 (rl78_expand_prologue): Always recompute the frame information.
9495 For interrupt functions: only select bank 0 if one of the bank 0
9496 registers is going to be psuhed.
9498 2013-09-17 DJ Delorie <dj@redhat.com>
9500 * config/rl78/constraints.md: For each W* constraint, rename to C*
9501 and create a W* constraint that checks for an optional ES: prefix
9503 * config/rl78/rl78.md (UNS_ES_ADDR): New.
9504 (es_addr): New. Used to wrap far addresses.
9505 * config/rl78/rl78-protos.h (rl78_es_addr): New.
9506 (rl78_es_base): New.
9507 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
9508 wrapped far addresses.
9509 (rl78_print_operand_1): Unwrap far addresses before processing.
9510 (rl78_lo16): Wrap far addresses in unspecs.
9511 (rl78_es_addr): New.
9512 (rl78_es_base): New.
9513 (insn_ok_now): Check for not-yet-wrapped far addresses.
9514 (transcode_memory_rtx): Properly re-wrap far addresses.
9516 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
9518 * config/sparc/t-rtems: Add leon3 multilibs.
9520 2013-09-17 Cong Hou <congh@google.com>
9522 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
9523 when checking the dot production pattern. The type of rhs operand
9524 of multiply is now checked correctly.
9526 2013-09-17 Jeff Law <law@redhat.com>
9528 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9529 edge implied equivalences into successor phis.
9530 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
9531 here from tree-ssa-threadedge.c.
9532 (mark_threaded_blocks): When threading through a joiner, if both
9533 successors of the joiner's clone reach the same block, verify the
9534 PHI arguments are equal. If not, cancel the jump threading request.
9535 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
9536 tree-ssa-threadupdate.c
9537 (thread_across_edge): Don't check PHI argument equality when
9538 threading through joiner block here.
9540 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
9542 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
9543 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
9544 (warn_uninit, warn_uninitialized_vars,
9545 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
9546 Move to tree-ssa-uninit.c.
9547 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
9548 (has_undefined_value_p): New. Pass dependant parts of
9549 ssa_undefined_value_p.
9550 (uninit_undefined_value_p): Use has_undefined_value_p.
9551 (warn_uninit, warn_uninitialized_vars,
9552 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
9553 Move from tree-ssa.c.
9554 * tree-ssa.h: Adjust prototypes.
9556 2013-09-17 Jan Hubicka <jh@suse.cz>
9559 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
9560 * ipa-inline.c (can_inline_edge_p): Do not downgrade
9561 FUNCTION_NOT_OPTIMIZED.
9562 * ipa-inline-analysis.c (compute_inline_parameters): Function
9563 not optimized is not inlinable unless it is alwaysinline.
9564 (inline_analyze_function): Force calls in not optimized
9565 function not inlinable.
9567 2013-09-17 Jan Hubicka <jh@suse.cz>
9570 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
9572 * ipa.c (function_and_variable_visibility): Likewise.
9573 * ipa-profile.c (ipa_profile): Likewise.
9575 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
9578 * cgraph.c (cgraph_function_body_availability): Check for ifunc
9579 attribute, and don't inline the resolver in this case.
9581 2013-09-17 Teresa Johnson <tejohnson@google.com>
9583 * coverage.c (get_coverage_counts): Add missing newline.
9585 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9587 PR tree-optimization/58088
9588 * fold-const.c (mask_with_trailing_zeros): New function.
9589 (fold_binary_loc): Make sure we don't recurse infinitely
9590 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
9591 Use mask_with_trailing_zeros where appropriate.
9593 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
9595 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
9596 of 'prev' var to get better distance estimation.
9598 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
9600 * tree-inline.h (struct copy_body_data): Add transform_parameter.
9601 * tree-inline.c (is_parameter_of): New predicate.
9602 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
9603 a parameter has been remapped.
9604 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
9605 (optimize_inline_calls): Initialize transform_parameter.
9606 (copy_gimple_seq_and_replace_locals): Likewise.
9607 (tree_function_versioning): Likewise.
9608 (maybe_inline_call_in_expr): Likewise.
9610 2013-09-17 Nick Clifton <nickc@redhat.com>
9612 * config/msp430/msp430-protos.h: Add prototypes for new functions.
9613 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
9615 (is_attr_func): New function.
9616 (msp430_is_interrupt_func): New function.
9617 (is_naked_func): New function.
9618 (is_reentrant_func): New function.
9619 (is_critical_func): New function.
9620 (msp430_start_function): Add annotations for function attributes.
9621 (msp430_attr): New function.
9622 (msp430_attribute_table): New.
9623 (msp430_function_section): New function.
9624 (TARGET_ASM_FUNCTION_SECTION): Define.
9625 (msp430_builtin): New enum.
9626 (msp430_init_builtins): New function.
9627 (msp430_builtin_devl): New function.
9628 (msp430_expand_builtin): New function.
9629 (TARGET_INIT_BUILTINS): Define.
9630 (TARGET_EXPAND_BUILTINS): Define.
9631 (TARGET_BUILTIN_DECL): Define.
9632 (msp430_expand_prologue): Add support for naked, interrupt,
9633 critical and reentrant functions.
9634 (msp430_expand_epilogue): Likewise.
9635 (msp430_print_operand): Handle 'O' character.
9636 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
9638 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
9639 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
9640 (pushm): Use a 'n' rather than an 'i' constraint.
9641 (msp_return): Add generation of the interrupt return instruction.
9642 (disable_interrupts): New pattern.
9643 (enable_interrupts): New pattern.
9644 (push_intr_state): New pattern.
9645 (pop_intr_state): New pattern.
9646 (bic_SR): New pattern.
9647 (bis_SR): New pattern.
9648 * doc/extend.texi: Document MSP430 function attributes and builtin
9651 2013-09-17 Richard Biener <rguenther@suse.de>
9653 PR tree-optimization/58432
9654 * tree-loop-distribution.c (tree_loop_distribution): Also
9655 scan PHIs for outside loop uses and seed a partition from them.
9657 2013-09-17 Bin Cheng <bin.cheng@arm.com>
9659 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
9660 (restructure_reference): Call backtrace_base_for_ref.
9662 2013-09-17 Alan Modra <amodra@gmail.com>
9665 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
9668 2013-09-16 DJ Delorie <dj@redhat.com>
9670 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
9671 vregs location for RL78/G10.
9672 (rl78_expand_prologue): Avoid SEL on G10.
9673 (rl78_expand_epilogue): Likewise.
9674 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
9675 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
9676 __RL78_G10__ when appropriate.
9677 (ASM_SPEC): Pass -mg10 along to the assembler.
9678 * config/rl78/rl78.md (sel_rb): Disable for G10.
9679 * config/rl78/rl78.opt: Add -mg10 option.
9680 * config/rl78/t-rl78: Add -mg10 multilib.
9682 2013-09-16 Xinliang David Li <davidxl@google.com>
9684 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
9685 * omp-low.c (omp_max_vf): Ditto.
9686 (expand_omp_simd): Ditto.
9687 * tree-vectorizer.c (vectorize_loops): Ditto.
9688 (gate_vect_slp): Ditto.
9689 (gate_increase_alignment): Ditto.
9690 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
9691 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
9692 (gate_tree_vectorize): Name change.
9693 (tree_vectorize): Ditto.
9694 (pass_vectorize::gate): Call new function.
9695 (pass_vectorize::execute): Ditto.
9696 * opts.c: O3 default setting change.
9697 (finish_options): Check new flag.
9698 * doc/invoke.texi: Document new flags.
9699 * common.opt: New flags.
9701 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
9703 * doc/tm.texi.in (Cond Exec Macros): Remove node.
9704 (Condition Code): Don't reference it.
9705 * doc/tm.texi: Regenerate.
9707 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
9710 * lra-constraints.c (undo_optional_reloads): Consider all optional
9711 reload even if it did not get a hard reg.
9713 2013-09-16 Teresa Johnson <tejohnson@google.com>
9715 * dumpfile.c (dump_loc): Remove newline emission.
9716 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
9717 emission to dump_printf_loc calls where missing.
9718 (vect_mark_for_runtime_alias_test): Ditto.
9719 (vect_analyze_data_ref_dependence): Ditto.
9720 (vect_analyze_data_ref_dependences): Ditto.
9721 (vect_slp_analyze_data_ref_dependence): Ditto.
9722 (vect_slp_analyze_data_ref_dependences): Ditto.
9723 (vect_compute_data_ref_alignment): Ditto.
9724 (vect_update_misalignment_for_peel): Ditto.
9725 (vect_verify_datarefs_alignment): Ditto.
9726 (vector_alignment_reachable_p): Ditto.
9727 (vect_get_data_access_cost): Ditto.
9728 (vect_enhance_data_refs_alignment): Ditto.
9729 (vect_find_same_alignment_drs): Ditto.
9730 (vect_analyze_data_refs_alignment): Ditto.
9731 (vect_analyze_group_access): Ditto.
9732 (vect_analyze_data_ref_access): Ditto.
9733 (vect_analyze_data_ref_accesses): Ditto.
9734 (vect_prune_runtime_alias_test_list): Ditto.
9735 (vect_analyze_data_refs): Ditto.
9736 (vect_create_addr_base_for_vector_ref): Ditto.
9737 (vect_create_data_ref_ptr): Ditto.
9738 (vect_grouped_store_supported): Ditto.
9739 (vect_grouped_load_supported): Ditto.
9740 * value-prof.c (check_counter): Ditto.
9741 (check_ic_target): Ditto.
9742 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
9743 (vect_recog_widen_mult_pattern): Ditto.
9744 (vect_recog_widen_sum_pattern): Ditto.
9745 (vect_recog_over_widening_pattern): Ditto.
9746 (vect_recog_widen_shift_pattern): Ditto.
9747 (vect_recog_rotate_pattern): Ditto.
9748 (vect_recog_vector_vector_shift_pattern): Ditto.
9749 (vect_recog_divmod_pattern): Ditto.
9750 (vect_recog_mixed_size_cond_pattern): Ditto.
9751 (vect_recog_bool_pattern): Ditto.
9752 (vect_pattern_recog_1): Ditto.
9753 (vect_pattern_recog): Ditto.
9754 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
9755 (vect_is_simple_iv_evolution): Ditto.
9756 (vect_analyze_scalar_cycles_1): Ditto.
9757 (vect_get_loop_niters): Ditto.
9758 (vect_analyze_loop_1): Ditto.
9759 (vect_analyze_loop_form): Ditto.
9760 (vect_analyze_loop_operations): Ditto.
9761 (vect_analyze_loop_2): Ditto.
9762 (vect_analyze_loop): Ditto.
9763 (report_vect_op): Ditto.
9764 (vect_is_slp_reduction): Ditto.
9765 (vect_is_simple_reduction_1): Ditto.
9766 (vect_get_known_peeling_cost): Ditto.
9767 (vect_estimate_min_profitable_iters): Ditto.
9768 (vect_model_reduction_cost): Ditto.
9769 (vect_model_induction_cost): Ditto.
9770 (get_initial_def_for_induction): Ditto.
9771 (vect_create_epilog_for_reduction): Ditto.
9772 (vectorizable_reduction): Ditto.
9773 (vectorizable_induction): Ditto.
9774 (vectorizable_live_operation): Ditto.
9775 (vect_loop_kill_debug_uses): Ditto.
9776 (vect_transform_loop): Ditto.
9777 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
9778 (vect_stmt_relevant_p): Ditto.
9779 (process_use): Ditto.
9780 (vect_mark_stmts_to_be_vectorized): Ditto.
9781 (vect_model_simple_cost): Ditto.
9782 (vect_model_promotion_demotion_cost): Ditto.
9783 (vect_model_store_cost): Ditto.
9784 (vect_get_store_cost): Ditto.
9785 (vect_model_load_cost): Ditto.
9786 (vect_get_load_cost): Ditto.
9787 (vect_init_vector_1): Ditto.
9788 (vect_get_vec_def_for_operand): Ditto.
9789 (vect_finish_stmt_generation): Ditto.
9790 (vectorizable_call): Ditto.
9791 (vectorizable_conversion): Ditto.
9792 (vectorizable_assignment): Ditto.
9793 (vectorizable_shift): Ditto.
9794 (vectorizable_operation): Ditto.
9795 (vectorizable_store): Ditto.
9796 (vectorizable_load): Ditto.
9797 (vectorizable_condition): Ditto.
9798 (vect_analyze_stmt): Ditto.
9799 (vect_transform_stmt): Ditto.
9800 (vect_is_simple_use): Ditto.
9801 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
9802 (vect_can_advance_ivs_p): Ditto.
9803 (vect_update_ivs_after_vectorizer): Ditto.
9804 (vect_do_peeling_for_loop_bound): Ditto.
9805 (vect_gen_niters_for_prolog_loop): Ditto.
9806 (vect_update_inits_of_drs): Ditto.
9807 (vect_create_cond_for_alias_checks): Ditto.
9808 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
9809 (vect_build_slp_tree_1): Ditto.
9810 (vect_supported_load_permutation_p): Ditto.
9811 (vect_analyze_slp_instance): Ditto.
9812 (vect_analyze_slp): Ditto.
9813 (vect_make_slp_decision): Ditto.
9814 (vect_detect_hybrid_slp): Ditto.
9815 (vect_bb_vectorization_profitable_p): Ditto.
9816 (vect_slp_analyze_bb_1): Ditto.
9817 (vect_update_slp_costs_according_to_vf): Ditto.
9818 (vect_get_mask_element): Ditto.
9819 (vect_transform_slp_perm_load): Ditto.
9820 (vect_schedule_slp_instance): Ditto.
9821 (vect_schedule_slp): Ditto.
9822 (vect_slp_transform_bb): Ditto.
9823 * profile.c (read_profile_edge_counts): Ditto.
9824 (compute_branch_probabilities): Ditto.
9825 * coverage.c (get_coverage_counts): Ditto.
9827 2013-09-16 Diego Novillo <dnovillo@google.com>
9829 * tree-core.h: Add missing comment lines from refactoring of tree.h.
9831 2013-09-16 Jan Hubicka <jh@suse.cz>
9833 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
9834 abstract functions; for static functions check the presence of body.
9836 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
9838 * config/aarch64/aarch64-simd-builtins.def (fma): New.
9839 * config/aarch64/aarch64-simd.md
9840 (aarch64_mla_elt<mode>): New.
9841 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9842 (aarch64_mls_elt<mode>): Likewise.
9843 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9844 (aarch64_fma4_elt<mode>): Likewise.
9845 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9846 (aarch64_fma4_elt_to_128v2df): Likewise.
9847 (aarch64_fma4_elt_to_64df): Likewise.
9848 (fnma<mode>4): Likewise.
9849 (aarch64_fnma4_elt<mode>): Likewise.
9850 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9851 (aarch64_fnma4_elt_to_128v2df): Likewise.
9852 (aarch64_fnma4_elt_to_64df): Likewise.
9853 * config/aarch64/iterators.md (VDQSF): New.
9854 * config/aarch64/arm_neon.h
9855 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
9856 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
9858 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
9860 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
9861 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
9862 (aarch64_mul3_elt_to_128df): Likewise.
9863 (aarch64_mul3_elt_to_64v2df): Likewise.
9864 * config/aarch64/iterators.md (VEL): Also handle DFmode.
9866 (VMUL_CHANGE_NLANES) Likewise.
9869 * config/aarch64/arm_neon.h
9870 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
9872 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
9874 * config/aarch64/arm_neon.h
9875 (vcvtx_high_f32_f64): Fix parameters.
9877 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9878 Uros Bizjak <ubizjak@gmail.com>
9880 * config/alpha.c: Include tree-ssanames.h.
9882 2013-09-16 Richard Biener <rguenther@suse.de>
9884 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
9885 (dot_rdg_1): Handle control_dd.
9886 (create_edge_for_control_dependence): New function.
9887 (create_rdg_edges): Add control dependences if asked for.
9888 (build_rdg): Likewise.
9889 (generate_loops_for_partition): If there are not necessary
9890 control stmts remove all their dependencies.
9891 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
9892 (distribute_loop): Pass on control dependences.
9893 (tree_loop_distribution): Compute control dependences and remove
9894 restriction on number of loop nodes.
9896 2013-09-16 Jakub Jelinek <jakub@redhat.com>
9898 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
9901 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
9903 * cse.c (try_const_anchors): Punt on CC modes.
9905 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9907 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
9909 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
9912 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
9915 2013-09-14 Iain Sandoe <iain@codesourcery.com>
9918 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
9920 (darwin_objc1_section): Likewise.
9921 (darwin_file_end): Emit Image Info section when required.
9923 2013-09-14 Jan Hubicka <jh@suse.cz>
9925 * tree-into-ssa.c (gate_into_ssa): New.
9926 (pass_data_build_ssa): Use it.
9927 * cgraph.h (expand_thunk): Update prototype.
9928 * cgraphunit.c (analyze_function): Expand thunks early.
9929 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
9930 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
9931 set lowered flag; do not add new function.
9932 (assemble_thunks_and_aliases): Update.
9933 * tree-ssa.c (gate_init_datastructures): New gate.
9934 (pass_data_init_datastructures): Use it.
9936 2013-09-14 Iain Sandoe <iain@codesourcery.com>
9939 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
9940 xmm register set described in the psABI.
9942 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
9944 * dwarf2out.c (should_emit_struct_debug): Add check
9945 for type_decl variable is not NULL.
9947 2013-09-13 Jacek Caban <jacek@codeweavers.com>
9949 * config.gcc: Use new winnt-c.c target hooks
9950 * config/t-winnt: New file
9951 * config/winnt-c.c: New file
9952 * doc/tm.texi.in: Document new hook
9953 * doc/tm.texi: Regenerated
9955 2013-09-13 Jan Hubicka <jh@suse.cz>
9958 * ipa-inline.c (check_callers): New function.
9959 (check_caller_edge): Remove.
9960 (want_inline_function_to_all_callers_p): Also permit alises that are
9962 (inline_to_all_callers): Terminate the walk when devirtualization
9963 introduce new calls.
9965 2013-09-13 Jan Hubicka <jh@suse.cz>
9967 * ipa-inline-analysis.c (struct growth_data): Add node.
9968 (do_estimate_growth_1): Fix detection of recursion.
9970 2013-09-13 Jakub Jelinek <jakub@redhat.com>
9972 PR tree-optimization/58392
9973 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
9974 to avoid shadowing of outer loop variable. If
9975 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
9976 replace_by_duplicate_decl simduid of loops that have it set and
9977 set dest_cfun->has_simduid_loops and/or
9978 dest_cfun->has_force_vect_loops.
9979 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
9980 instead of maybe_lookup_decl.
9981 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
9982 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
9983 call. Copy over force_vect and copy and remap simduid. Set
9984 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
9985 (copy_cfg_body): Remove blocks_to_copy argument. Use
9986 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
9987 caller. Don't set cfun->has_simduid_loops and/or
9988 cfun->has_force_vect_loops here.
9989 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
9991 (expand_call_inline, tree_function_versioning): Adjust copy_body
9994 2013-09-13 Martin Jambor <mjambor@suse.cz>
9997 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
9998 the assert if the edge was a speculative one.
10000 2013-09-13 Richard Biener <rguenther@suse.de>
10002 * tree-data-ref.h (known_dependences_p): Move here ...
10003 * tree-loop-distribution.c (known_dependences_p): ... from here.
10004 (dump_rdg_component, debug_rdg_component): Remove.
10005 (dump_rdg): Adjust.
10006 (generate_loops_for_partition): Use gimple_uid instead of
10007 relying on matching stmt visit order.
10008 (rdg_build_partitions): Take starting stmt vector.
10009 (ldist_gen): Merge into ...
10010 (distribute_loop): ... this function. Do not compute starting
10012 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
10014 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10016 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
10017 Set type attribute correctly. Set predicable_short_it attribute.
10018 (cmpsi_shiftsi): Remove %? from output template.
10020 2013-09-13 Richard Biener <rguenther@suse.de>
10022 * tree-loop-distribution.c (struct rdg_component,
10023 rdg_defs_used_in_other_loops_p, free_rdg_components,
10024 rdg_build_components): Remove.
10025 (stmts_from_loop): Do not record virtual PHIs.
10026 (generate_loops_for_partition): Skip virtual PHIs.
10027 (build_rdg_partition_for_component): Rename to ...
10028 (build_rdg_partition_for_vertex): ... this and adjust.
10029 (rdg_build_partitions): Take a vector of starting vertices
10030 instead of components. Remove unnecessary leftover handling.
10031 (ldist_gen): Do not build components or record other stores.
10032 (distribute_loop): Do not distribute loops containing stmts
10035 2013-09-13 Christian Bruel <christian.bruel@st.com>
10038 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
10040 2013-09-13 Kai Tietz <ktietz@redhat.com>
10042 * config.gcc: Separate cases for mingw and cygwin targets,
10043 and add 64-bit cygwin target case.
10045 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
10046 dll-export inline-functions.
10047 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
10049 2013-09-13 Jeff Law <law@redhat.com>
10051 PR middle-end/58387
10053 2013-09-06 Jeff Law <law@redhat.com>
10055 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
10056 edge implied equivalences into successor phis.
10058 2013-09-12 DJ Delorie <dj@redhat.com>
10060 * config/rl78/rl78-virt.md: Change from | to \; for asm line
10063 2013-09-12 Brooks Moses <bmoses@google.com>
10066 * Makefile.in: Do not install driver binaries in $(target)/bin.
10068 2013-09-12 DJ Delorie <dj@redhat.com>
10070 * config/rl78/rl78.opt (mrelax): New.
10071 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
10072 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
10074 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
10076 (rl78_expand_epilogue): Likewise.
10078 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
10080 PR middle-end/58335
10081 * lra-eliminations.c (remove_reg_equal_offset_note): New.
10082 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
10083 pointer elimination with using remove_reg_equal_offset_note.
10085 2013-09-12 DJ Delorie <dj@redhat.com>
10087 * config/msp430/: New port.
10088 * config.gcc (msp430): Added.
10089 * doc/invoke.texi: Document MSP430 options.
10090 * doc/install.texi: Document msp430-elf
10091 * doc/md.texi: Document msp430-elf
10092 * doc/contrib.texi: Document msp430-elf
10094 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
10097 2013-09-12 Martin Jambor <mjambor@suse.cz>
10100 * ipa-prop.c (remove_described_reference): Give up if the edge in the
10101 reference descriptor is NULL.
10102 (ipa_edge_removal_hook): If owning a reference descriptor, set its
10105 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
10107 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
10108 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
10109 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
10110 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
10111 make_temp_ssa_name): move to tree-ssanames.h
10112 * tree-ssa-alias.h: Move prototype.
10113 * tree-ssa.h: Include tree-ssanames.h.
10114 * tree-ssanames.c (FREE_SSANAMES): Move to here.
10115 * tree-ssanames.h: New. Move items from tree-flow*.h
10116 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
10118 2013-09-12 Richard Biener <rguenther@suse.de>
10120 PR tree-optimization/58404
10121 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
10122 propagate non-invariant addresses into dereferences wrapped
10123 in component references.
10125 2013-09-12 Richard Biener <rguenther@suse.de>
10127 PR tree-optimization/58402
10128 * passes.def: Move pass_late_warn_uninitialized later.
10130 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
10132 * tree-ssa.h: New. Move content from tree-flow.h and
10133 tree-flow-inline.h.
10134 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
10135 Move prototypes belonging to tree-ssa.c.
10136 * tree-flow-inline.h (redirect_edge_var_map_def,
10137 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
10139 * gimple.h: Adjust prototypes.
10140 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
10142 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
10143 * tree.h: Move prototype to tree-ssa.h.
10144 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
10145 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
10146 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
10147 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
10148 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
10149 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
10150 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
10151 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
10152 graphite-blocking.c, graphite-clast-to-gimple.c,
10153 graphite-dependences.c, graphite-interchange.c,
10154 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
10155 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
10156 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
10157 ipa-reference.c, ipa-split.c, ipa-utils.c,
10158 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
10159 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
10160 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
10161 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
10162 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
10163 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
10164 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
10165 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
10166 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
10167 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
10168 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
10169 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
10170 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
10171 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
10172 value-prof.c, var-tracking.c,
10173 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
10175 2013-09-12 Richard Biener <rguenther@suse.de>
10177 PR tree-optimization/58396
10178 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
10179 (build_rdg): Take a loop-nest parameter, fix memleaks.
10180 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
10182 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
10184 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
10187 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
10189 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
10190 The FMA instruction names should have a 'v' prefix.
10192 2013-09-12 Richard Biener <rguenther@suse.de>
10194 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
10195 (dot_rdg): Use popen instead of system in optional code.
10196 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
10197 (already_processed_vertex_p): Adjust.
10198 (has_anti_or_output_dependence, predecessor_has_mem_write,
10199 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
10200 rdg_flag_uses): Remove.
10201 (rdg_flag_vertex): Simplify.
10202 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
10204 (build_rdg_partition_for_component): Process the first vertex
10205 of a component only.
10206 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
10208 2013-09-12 Alan Modra <amodra@gmail.com>
10210 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
10212 2013-09-11 DJ Delorie <dj@redhat.com>
10213 Nick Clifton <nickc@redhat.com>
10215 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
10216 (rl78_stack_based_mem): New.
10217 * config/rl78/constraints.md (Iv08): New.
10226 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
10228 (movsi): Change from expand to insn-and-split.
10229 (ashrsi3): Clobber AX.
10233 * config/rl78/rl78.md (CC_REG): Fix.
10234 (addsi3): Allow memory and immediate operands.
10235 (addsi3_internal): Split into...
10236 (addsi3_internal_virt): ...new, and ...
10237 (addsi3_internal_real): ...new.
10239 (subsi3_internal_virt): New.
10240 (subsi3_internal_real): New.
10241 (mulsi3): Add memory operand.
10242 (mulsi3_rl78): Likewise.
10243 (mulsi3_g13): Likewise.
10244 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
10245 (cbranchqi4_real): Add more constraint options.
10246 (cbranchhi4_real): Expand pattern.
10247 (cbranchhi4_real_signed): New.
10248 (cbranchhi4_real_inverted): New.
10249 (cbranchsi4_real_lt): New.
10250 (cbranchsi4_real_ge): New.
10251 (cbranchsi4_real_signed): New.
10252 (cbranchsi4_real): New.
10254 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
10256 (lshrsi3_virt): Likewise.
10257 (ashlsi3_virt): Likewise.
10258 (cbranchqi4_virt_signed): New.
10259 (cbranchhi4_virt_signed): New.
10260 (cbranchsi4_virt): New.
10261 * config/rl78/rl78.c: Whitespace fixes throughout.
10262 (move_elim_pass): New.
10263 (pass_data_rl78_move_elim): New.
10264 (pass_rl78_move_elim): New.
10265 (make_pass_rl78_move_elim): New.
10266 (rl78_devirt_info): Run devirt earlier.
10267 (rl78_move_elim_info): New.
10268 (rl78_asm_file_start): Register it.
10269 (rl78_split_movsi): New.
10270 (rl78_as_legitimate_address): Allow virtual base registers when
10272 (rl78_addr_space_convert): Remove spurious debug stuff.
10273 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
10274 (rl78_print_operand): More cases for not printing '#'.
10275 (rl78_expand_compare): Remove most of the logic.
10276 (content_memory): New.
10277 (clear_content_memory): New.
10278 (get_content_index): New.
10279 (get_content_name): New.
10280 (display_content_memory): New.
10281 (update_content): New.
10282 (record_content): New.
10283 (already_contains): New.
10284 (insn_ok_now): Re-recog insns with virtual registers.
10285 (add_postponed_content_update): New.
10286 (process_postponed_content_update): New.
10287 (gen_and_emit_move): New.
10288 (transcode_memory_rtx): Record new location content.
10289 Use gen_and_emit_move.
10290 (force_into_acc): New.
10291 (move_to_acc): Use gen_and_emit_move.
10292 (move_from_acc): Likewise.
10293 (move_acc_to_reg): Likewise.
10294 (move_to_x): Likewise.
10295 (move_to_hl): Likewise.
10296 (move_to_de): Likewise.
10297 (rl78_alloc_physical_registers_op1): Record location content.
10298 (has_constraint): New.
10299 (rl78_alloc_physical_registers_op2): Record location content.
10300 Optimize use of HL.
10301 (rl78_alloc_physical_registers_ro1): Likewise.
10302 (rl78_alloc_physical_registers_cmp): Likewise.
10303 (rl78_alloc_physical_registers_umul): Likewise.
10304 (rl78_alloc_address_registers_macax): New.
10305 (rl78_alloc_physical_registers): Initialize and set location
10306 content memory as needed.
10307 (rl78_reorg): Make sure split2 is called.
10308 (rl78_rtx_costs): New.
10310 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
10312 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
10313 for (not (neg ...)) and (neg (not ...)) cases.
10315 2013-09-11 Richard Biener <rguenther@suse.de>
10317 PR middle-end/58377
10318 * passes.def: Split critical edges before late uninit warning passes.
10319 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
10321 2013-09-11 Jakub Jelinek <jakub@redhat.com>
10323 PR tree-optimization/58385
10324 * fold-const.c (build_range_check): If both low and high are NULL,
10325 use omit_one_operand_loc to preserve exp side-effects.
10327 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10329 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
10331 2013-09-11 Richard Biener <rguenther@suse.de>
10333 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
10334 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
10335 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
10336 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
10337 stmts_from_loop, known_dependences_p, build_empty_rdg,
10338 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
10339 * tree-loop-distribution.c: ... here.
10340 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
10341 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
10342 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
10343 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
10344 * tree-loop-distribution.c: ... here.
10345 * tree-loop-distribution.c: Include gimple-pretty-print.h.
10346 (struct partition_s): Add loops member.
10347 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
10348 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
10349 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
10351 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
10352 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
10353 Sergey Lega <sergey.s.lega@intel.com>
10354 Anna Tikhonova <anna.tikhonova@intel.com>
10355 Ilya Tocar <ilya.tocar@intel.com>
10356 Andrey Turetskiy <andrey.turetskiy@intel.com>
10357 Ilya Verbin <ilya.verbin@intel.com>
10358 Kirill Yukhin <kirill.yukhin@intel.com>
10359 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10361 * config/i386/constraints.md (k): New.
10363 * config/i386/i386.c (const regclass_map): Add new mask registers.
10364 (dbx_register_map): Ditto.
10365 (dbx64_register_map): Ditto.
10366 (svr4_dbx_register_map): Ditto.
10367 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
10369 (ix86_preferred_reload_class): Disable constants for mask registers.
10370 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
10371 (ix86_hard_regno_mode_ok): Support new mask registers.
10372 (x86_order_regs_for_local_alloc): Ditto.
10373 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
10374 (FIXED_REGISTERS): Add new mask registers.
10375 (CALL_USED_REGISTERS): Ditto.
10376 (REG_ALLOC_ORDER): Ditto.
10377 (VALID_MASK_REG_MODE): New.
10378 (FIRST_MASK_REG): Ditto.
10379 (LAST_MASK_REG): Ditto.
10380 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
10381 (MAYBE_MASK_CLASS_P): New.
10382 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
10383 (REG_CLASS_CONTENTS): Ditto.
10384 (MASK_REGNO_P): New.
10385 (ANY_MASK_REG_P): Ditto.
10386 (HI_REGISTER_NAMES): Add new mask registers.
10387 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
10388 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
10390 (attribute "type"): Add mskmov, msklog.
10391 (attribute "length_immediate"): Support them.
10392 (attribute "memory"): Ditto.
10393 (attribute "prefix_0f"): Ditto.
10394 (*movhi_internal): Support new mask registers.
10395 (*movqi_internal): Ditto.
10396 (define_split): Split out clobber pattern is a logic
10397 insn on mask registers.
10398 (*k<logic><mode>): New.
10399 (*andhi_1): Extend to support mask regs.
10400 (*andqi_1): Extend to support mask regs.
10401 (kandn<mode>): New.
10402 (define_split): Split and-not to and and not if operands
10404 (*<code><mode>_1): Separate HI mode to new pattern...
10405 (*<code>hi_1): This.
10406 (*<code>qi_1): Extend to support mask regs.
10407 (kxnor<mode>): New.
10408 (kortestzhi): Ditto.
10409 (kortestchi): Ditto.
10411 (*one_cmpl<mode>2_1): Remove HImode and handle it...
10412 (*one_cmplhi2_1): ...Here, now with mask registers support.
10413 (*one_cmplqi2_1): Support new mask registers.
10414 (HI/QImode arithmetics splitter): Don't split if mask registers
10416 (HI/QImode not splitter): Ditto.
10417 * config/i386/predicated.md (mask_reg_operand): New.
10418 (general_reg_operand): Ditto.
10420 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
10422 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
10423 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
10425 2013-09-10 Jeff Law <law@redhat.com>
10427 PR tree-optimization/58380
10428 * tree-ssa-threadupdate.c (thread_block): Recognize another case
10429 of threading through a buried loop header.
10431 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
10432 return value for single successor case.
10434 2013-09-10 Jan Hubicka <jh@suse.cz>
10436 * ipa-devirt.c (ipa_devirt): Enable with LTO.
10438 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
10441 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
10442 support conditional execution.
10443 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
10445 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
10447 * lra.c (lra): Clear lra_optional_reload_pseudos before every
10449 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
10450 Check destination too to check move insn.
10451 (undo_optional_reloads): Add check that the original peudo did not
10452 changed its allocation and the optional reload was inherited on last
10453 inheritance pass. Break loop after deciding to keep optional reload.
10454 (lra_undo_inheritance): Add check that inherited pseudo still in
10457 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
10459 * config/aarch64/aarch64.md (generic_sched): New.
10460 * config/aarch64/aarch64-generic.md (load): Make conditional
10461 on generic_sched attribute.
10462 (nonload): Likewise.
10464 2013-09-10 Jan Hubicka <jh@suse.cz>
10466 * lto-cgraph.c: Include ipa-utils.h.
10467 (compute_ltrans_boundary): Also add possible targets into the boundary.
10469 2013-09-10 Jan Hubicka <jh@suse.cz>
10471 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
10472 VAR_DECL of vtable rather than full expression.
10474 2013-09-10 Jan Hubicka <jh@suse.cz>
10475 Paolo Carlini <paolo.carlini@oracle.com>
10477 * cgraphunit.c (analyze_functions): Save input_location, set it
10478 to UNKNOWN_LOCATION and restore it at the end.
10480 2013-09-10 Martin Jambor <mjambor@suse.cz>
10482 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
10483 represented by a thunk.
10485 2013-09-10 Jeff Law <law@redhat.com>
10487 PR tree-optimization/58343
10488 * tree-ssa-threadupdate.c (thread_block): Identify and disable
10489 jump threading requests through loop headers buried in the middle
10490 of a jump threading path.
10492 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
10493 in return value/type.
10495 2013-09-10 Jakub Jelinek <jakub@redhat.com>
10497 PR rtl-optimization/58365
10498 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
10499 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
10502 2013-09-10 Richard Biener <rguenther@suse.de>
10504 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
10505 * tree-data-ref.c (create_rdg_vertices): Collect all data
10506 references, signal failure to the caller, use data-ref API.
10507 (build_rdg): Compute data references only once. Maintain lifetime
10508 of data references and data dependences from within RDG.
10509 (free_rdg): Free dependence relations.
10510 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
10511 inventing extra dependences.
10512 (distribute_loop): Update for RDG API changes.
10514 2013-09-10 Kai Tietz <ktietz@redhat.com>
10516 * doc/invoke.texi (fms-extensions): Document changed
10517 behavior for ms-abi targets.
10518 * config/i386/i386.c (ix86_option_override_internal):
10519 Set default value of option -fms-extension for ms-abi targets.
10521 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
10523 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
10525 2013-09-10 Alan Modra <amodra@gmail.com>
10528 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
10530 2013-09-10 Alan Modra <amodra@gmail.com>
10532 * config/rs6000/predicates.md (add_cint_operand): New.
10533 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
10534 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
10535 using add_cint_operand.
10536 (largetoc_high_plus_aix): Likewise.
10538 2013-09-09 Jakub Jelinek <jakub@redhat.com>
10540 PR tree-optimization/58364
10541 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
10542 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
10543 the current range can't be an unconditional true or false.
10545 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
10547 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
10549 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
10551 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
10553 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
10556 * doc/invoke.texi (-Wdelete-incomplete): Document it.
10558 2013-09-09 Ian Bolton <ian.bolton@arm.com>
10560 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
10561 NO_REGS for immediate that can't be moved directly into FP_REGS.
10563 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10565 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
10566 comparison with negated operand.
10567 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
10570 2013-09-09 Richard Biener <rguenther@suse.de>
10572 PR middle-end/58326
10573 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
10574 of a subloop record all its block as affecting loop-closed SSA form.
10576 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10578 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
10579 of an rtx/bitpos pair.
10580 (store_fixed_bit_field): Update accordingly.
10582 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10584 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
10586 * builtins.c (expand_errno_check): Likewise.
10587 * dwarf2cfi.c (init_return_column_size): Likewise.
10588 * except.c (sjlj_mark_call_sites): Likewise.
10589 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
10590 * lra-constraints.c (emit_inc): Likewise.
10591 * ree.c (combine_set_extension): Likewise.
10592 * regmove.c (fixup_match_2): Likewise.
10593 * reload1.c (inc_for_reload): Likewise.
10595 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10597 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
10598 (force_to_mode, simplify_shift_const_1, simplify_comparison):
10599 Use gen_int_mode with the mode of the associated simplify_* call.
10600 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
10601 * expmed.c (expand_shift_1): Likewise.
10602 * function.c (instantiate_virtual_regs_in_insn): Likewise.
10603 * loop-iv.c (iv_number_of_iterations): Likewise.
10604 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
10605 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
10607 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10609 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
10610 of the associated expand_* call.
10611 (asan_emit_stack_protection): Likewise.
10612 * builtins.c (round_trampoline_addr): Likewise.
10613 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
10614 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
10615 (emit_store_flag): Likewise.
10616 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
10617 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
10619 * function.c (instantiate_virtual_regs_in_insn): Likewise.
10620 * ifcvt.c (noce_try_store_flag_constants): Likewise.
10621 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
10622 * modulo-sched.c (generate_prolog_epilog): Likewise.
10623 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
10624 (expand_ctz, expand_ffs, expand_unop): Likewise.
10626 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10628 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
10629 of the associated gen_rtx_* call.
10630 * caller-save.c (init_caller_save): Likewise.
10631 * combine.c (find_split_point, make_extraction): Likewise.
10632 (make_compound_operation): Likewise.
10633 * dwarf2out.c (mem_loc_descriptor): Likewise.
10634 * explow.c (plus_constant, probe_stack_range): Likewise.
10635 * expmed.c (expand_mult_const): Likewise.
10636 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
10637 * reload1.c (init_reload): Likewise.
10638 * valtrack.c (cleanup_auto_inc_dec): Likewise.
10639 * var-tracking.c (adjust_mems): Likewise.
10640 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
10641 rather than gen_rtx_fmt_ee.
10643 2013-09-09 Jan Hubicka <jh@suse.cz>
10645 PR middle-end/58294
10646 * value-prof.c (gimple_ic): Copy also abnormal edges.
10648 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
10650 * asan.c (asan_shadow_cst): Use gen_int_mode.
10652 2013-09-08 Jan Hubicka <jh@suse.cz>
10654 * ipa-profile.c: Add toplevel comment.
10655 (ipa_propagate_frequency_1): Be more conservative when profile is read.
10656 (contains_hot_call_p): New function.
10657 (ipa_propagate_frequency): Set frequencies based on counts when
10659 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
10660 profile; do not tamper with profile after inlining if it is read.
10662 2013-09-08 Jan Hubicka <jh@suse.cz>
10664 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
10667 2013-09-08 Jan Hubicka <jh@suse.cz>
10669 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
10670 summary generation.
10672 2013-09-08 Jeff Law <law@redhat.com>
10675 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
10678 2013-09-08 Andi Kleen <ak@linux.intel.com>
10680 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
10682 2013-09-08 Jan Hubicka <jh@suse.cz>
10684 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
10686 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
10688 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
10689 for non-debug insns.
10690 * lra.c (new_insn_reg): Take the containing insn as a parameter.
10691 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
10692 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
10695 2013-09-08 Jan Hubicka <jh@suse.cz>
10697 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
10698 targets and devirtualize to BUILT_IN_UNREACHABLE.
10699 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
10700 * ipa.c (walk_polymorphic_call_targets): New function.
10701 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
10702 functions; use the new timevar.
10703 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
10704 was removed from the program.
10705 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
10706 not consider it in the walk when its vtable is dead.
10707 (possible_polymorphic_call_targets_1): Pass anonymous flag to
10709 (devirt_variable_node_removal_hook): New function.
10710 (possible_polymorphic_call_targets): Also register
10711 devirt_variable_node_removal_hook.
10712 (ipa_devirt): Do not do non-speculative devirtualization.
10713 (gate_ipa_devirt): One execute if devirtualizing speculatively.
10715 2013-09-08 Jan Hubicka <jh@suse.cz>
10717 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
10718 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
10719 varpool_remove_variable_insertion_hook): Declare.
10720 * varpool.c (varpool_node_hook_list): New structure.
10721 (first_varpool_node_removal_hook,
10722 first_varpool_variable_insertion_hook): New variables.
10723 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
10724 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
10725 varpool_remove_variable_insertion_hook,
10726 varpool_call_variable_insertion_hooks): New functions.
10727 (varpool_remove_node): Use it.
10729 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
10732 * diagnostic.c (diagnostic_build_prefix): When s.file is
10733 "<built-in>" don't output line and column numbers.
10735 2013-09-06 Jan Hubicka <jh@suse.cz>
10737 * cgraphunit.c (expand_thunk): Get body before touching arguments.
10738 * lto-streamer-out.c: Stream thunks, too.
10739 * lto-streamer-in.c (input_function): Pop cfun here
10740 (lto_read_body): Instead of here.
10742 2013-09-06 Caroline Tice <cmtice@google.com>
10744 * doc/install.texi: Add documentation for the --enable-vtable-verify
10745 and the --disable-libvtv configure options.
10747 2013-09-06 Jeff Law <law@redhat.com>
10749 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
10750 edge implied equivalences into successor phis.
10752 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
10754 * resource.c (mark_referenced_resources): Handle COND_EXEC.
10756 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
10758 * resource.c (mark_target_live_regs): Compute resources taking
10759 into account if a call is predicated or not.
10761 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
10763 * toplev.c (output_stack_usage): Be prepared for suffixes created by
10764 the compiler in the function names.
10766 2013-09-06 Jan Hubicka <jh@suse.cz>
10768 PR middle-end/58094
10769 * ipa-inline.c (has_caller_p): New function.
10770 (want_inline_function_to_all_callers_p): Use it.
10771 (sum_callers, inline_to_all_callers): Break out from ...
10772 (ipa_inline): ... here.
10774 2013-09-06 Jan Hubicka <jh@suse.cz>
10776 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
10777 only when AVX is enabled.
10779 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10781 * config/aarch64/aarch64.md
10782 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
10783 is fpsimd_<load/store>2.
10784 (load_pair<mode>): Likewise.
10785 (store_pair<mode>): Likewise.
10787 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10789 * config/arm/types.md (type): Add "mrs" type.
10790 * config/aarch64/aarch64.md
10791 (aarch64_load_tp_hard): Make type "mrs".
10792 * config/arm/arm.md
10793 (load_tp_hard): Make type "mrs".
10794 * config/arm/cortex-a15.md: Update with new attributes.
10795 * config/arm/cortex-a5.md: Update with new attributes.
10796 * config/arm/cortex-a53.md: Update with new attributes.
10797 * config/arm/cortex-a7.md: Update with new attributes.
10798 * config/arm/cortex-a8.md: Update with new attributes.
10799 * config/arm/cortex-a9.md: Update with new attributes.
10800 * config/arm/cortex-m4.md: Update with new attributes.
10801 * config/arm/cortex-r4.md: Update with new attributes.
10802 * config/arm/fa526.md: Update with new attributes.
10803 * config/arm/fa606te.md: Update with new attributes.
10804 * config/arm/fa626te.md: Update with new attributes.
10805 * config/arm/fa726te.md: Update with new attributes.
10807 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10809 * config/aarch64/aarch64.md
10810 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
10811 (*movtf_aarch64): Likewise.
10812 * config/arm/arm.md
10813 (thumb1_movdi_insn): Use "multiple" for type where more than one
10814 instruction is used for a move.
10815 (*arm32_movhf): Likewise.
10816 (*thumb_movdf_insn): Likewise.
10818 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10820 * config/arm/types.md (type): Rename fcpys to fmov.
10821 * config/arm/vfp.md
10822 (*arm_movsi_vfp): Rename type fcpys as fmov.
10823 (*thumb2_movsi_vfp): Likewise
10824 (*movhf_vfp_neon): Likewise
10825 (*movhf_vfp): Likewise
10826 (*movsf_vfp): Likewise
10827 (*thumb2_movsf_vfp): Likewise
10828 (*movsfcc_vfp): Likewise
10829 (*thumb2_movsfcc_vfp): Likewise
10830 * config/aarch64/aarch64-simd.md
10831 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
10832 * config/aarch64/aarch64.md
10833 (*movsi_aarch64): Replace type mov_reg with fmovs.
10834 (*movdi_aarch64): Likewise
10835 (*movsf_aarch64): Likewise
10836 (*movdf_aarch64): Likewise
10838 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
10839 * config/arm/iwmmxt.md
10840 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
10841 * config/arm/arm1020e.md: Update with new attributes.
10842 * config/arm/cortex-a15-neon.md: Update with new attributes.
10843 * config/arm/cortex-a5.md: Update with new attributes.
10844 * config/arm/cortex-a53.md: Update with new attributes.
10845 * config/arm/cortex-a7.md: Update with new attributes.
10846 * config/arm/cortex-a8-neon.md: Update with new attributes.
10847 * config/arm/cortex-a9.md: Update with new attributes.
10848 * config/arm/cortex-m4-fpu.md: Update with new attributes.
10849 * config/arm/cortex-r4f.md: Update with new attributes.
10850 * config/arm/marvell-pj4.md: Update with new attributes.
10851 * config/arm/vfp11.md: Update with new attributes.
10853 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10855 * config/aarch64/aarch64.md
10856 (*madd<mode>): Fix type attribute.
10857 (*maddsi_uxtw): Likewise.
10858 (*msub<mode>): Likewise.
10859 (*msubsi_uxtw): Likewise.
10860 (<su_optab>maddsidi4): Likewise.
10861 (<su_optab>msubsidi4): Likewise.
10863 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10865 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
10866 * config/arm/arm.md (core_cycles): Remove fdiv.
10867 * config/arm/vfp.md:
10868 (*sqrtsf2_vfp): Update for attribute changes.
10869 (*sqrtdf2_vfp): Likewise.
10870 * config/aarch64/aarch64.md:
10871 (sqrt<mode>2): Update for attribute changes.
10872 * config/arm/arm1020e.md: Update with new attributes.
10873 * config/arm/cortex-a15-neon.md: Update with new attributes.
10874 * config/arm/cortex-a5.md: Update with new attributes.
10875 * config/arm/cortex-a53.md: Update with new attributes.
10876 * config/arm/cortex-a7.md: Update with new attributes.
10877 * config/arm/cortex-a8-neon.md: Update with new attributes.
10878 * config/arm/cortex-a9.md: Update with new attributes.
10879 * config/arm/cortex-m4-fpu.md: Update with new attributes.
10880 * config/arm/cortex-r4f.md: Update with new attributes.
10881 * config/arm/marvell-pj4.md: Update with new attributes.
10882 * config/arm/vfp11.md: Update with new attributes.
10884 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10886 * config/arm/types.md
10887 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
10888 * config/aarch64/aarch64.md
10889 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
10891 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
10892 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
10893 (float<GPI:mode><GPF:mode>2): Likewise.
10894 * config/arm/vfp.md
10895 (*truncsisf2_vfp): Update with new attributes.
10896 (*truncsidf2_vfp): Likewise.
10897 (fixuns_truncsfsi2): Likewise.
10898 (fixuns_truncdfsi2): Likewise.
10899 (*floatsisf2_vfp): Likewise.
10900 (*floatsidf2_vfp): Likewise.
10901 (floatunssisf2): Likewise.
10902 (floatunssidf2): Likewise.
10903 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
10904 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
10905 * config/arm/arm1020e.md: Update with new attributes.
10906 * config/arm/cortex-a15-neon.md: Update with new attributes.
10907 * config/arm/cortex-a5.md: Update with new attributes.
10908 * config/arm/cortex-a53.md: Update with new attributes.
10909 * config/arm/cortex-a7.md: Update with new attributes.
10910 * config/arm/cortex-a8-neon.md: Update with new attributes.
10911 * config/arm/cortex-a9.md: Update with new attributes.
10912 * config/arm/cortex-m4-fpu.md: Update with new attributes.
10913 * config/arm/cortex-r4f.md: Update with new attributes.
10914 * config/arm/marvell-pj4.md: Update with new attributes.
10915 * config/arm/vfp11.md: Update with new attributes.
10917 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10919 * config/aarch64/arm_neon.h
10920 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
10921 (vqtbx<1,2,3,4><q>_s8): Likewise.
10923 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10925 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
10926 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
10927 (add<mode>3): Add type attribute.
10928 (add<mode>3): Likewise.
10929 (usadd<mode>3): Likewise.
10930 (ssadd<mode>3): Likewise.
10931 (sub<mode>3): Likewise.
10932 (sub<mode>3): Likewise.
10933 (ussub<mode>3): Likewise.
10934 (sssub<mode>3): Likewise.
10935 (ssmulsa3): Likewise.
10936 (usmulusa3): Likewise.
10937 (arm_usatsihi): Likewise.
10938 * config/arm/vfp.md
10939 (*movdi_vfp): Add types for all instructions.
10940 (*movdi_vfp_cortexa8): Likewise.
10941 (*movhf_vfp_neon): Likewise.
10942 (*movhf_vfp): Likewise.
10943 (*movdf_vfp): Likewise.
10944 (*thumb2_movdf_vfp): Likewise.
10945 (*thumb2_movdfcc_vfp): Likewise.
10946 * config/arm/arm.md: Add type attribute to all insn patterns.
10947 (*thumb1_adddi3): Add type attribute.
10948 (*arm_adddi3): Likewise.
10949 (*adddi_sesidi_di): Likewise.
10950 (*adddi_zesidi_di): Likewise.
10951 (*thumb1_addsi3): Likewise.
10952 (addsi3_compare0): Likewise.
10953 (*addsi3_compare0_scratch): Likewise.
10954 (*compare_negsi_si): Likewise.
10955 (cmpsi2_addneg): Likewise.
10956 (*addsi3_carryin_<optab>): Likewise.
10957 (*addsi3_carryin_alt2_<optab>): Likewise.
10958 (*addsi3_carryin_clobercc_<optab>): Likewise.
10959 (*subsi3_carryin): Likewise.
10960 (*subsi3_carryin_const): Likewise.
10961 (*subsi3_carryin_compare): Likewise.
10962 (*subsi3_carryin_compare_const): Likewise.
10963 (*arm_subdi3): Likewise.
10964 (*thumb_subdi3): Likewise.
10965 (*subdi_di_zesidi): Likewise.
10966 (*subdi_di_sesidi): Likewise.
10967 (*subdi_zesidi_di): Likewise.
10968 (*subdi_sesidi_di): Likewise.
10969 (*subdi_zesidi_ze): Likewise.
10970 (thumb1_subsi3_insn): Likewise.
10971 (*arm_subsi3_insn): Likewise.
10972 (*anddi3_insn): Likewise.
10973 (*anddi_zesidi_di): Likewise.
10974 (*anddi_sesdi_di): Likewise.
10975 (*ne_zeroextracts): Likewise.
10976 (*ne_zeroextracts): Likewise.
10977 (*ite_ne_zeroextr): Likewise.
10978 (*ite_ne_zeroextr): Likewise.
10979 (*anddi_notdi_di): Likewise.
10980 (*anddi_notzesidi): Likewise.
10981 (*anddi_notsesidi): Likewise.
10982 (andsi_notsi_si): Likewise.
10983 (thumb1_bicsi3): Likewise.
10984 (*iordi3_insn): Likewise.
10985 (*iordi_zesidi_di): Likewise.
10986 (*iordi_sesidi_di): Likewise.
10987 (*thumb1_iorsi3_insn): Likewise.
10988 (*xordi3_insn): Likewise.
10989 (*xordi_zesidi_di): Likewise.
10990 (*xordi_sesidi_di): Likewise.
10991 (*arm_xorsi3): Likewise.
10992 (*andsi_iorsi3_no): Likewise.
10993 (*smax_0): Likewise.
10994 (*smax_m1): Likewise.
10995 (*arm_smax_insn): Likewise.
10996 (*smin_0): Likewise.
10997 (*arm_smin_insn): Likewise.
10998 (*arm_umaxsi3): Likewise.
10999 (*arm_uminsi3): Likewise.
11000 (*minmax_arithsi): Likewise.
11001 (*minmax_arithsi_): Likewise.
11002 (*satsi_<SAT:code>): Likewise.
11003 (arm_ashldi3_1bit): Likewise.
11004 (arm_ashrdi3_1bit): Likewise.
11005 (arm_lshrdi3_1bit): Likewise.
11006 (*arm_negdi2): Likewise.
11007 (*thumb1_negdi2): Likewise.
11008 (*arm_negsi2): Likewise.
11009 (*thumb1_negsi2): Likewise.
11010 (*negdi_extendsid): Likewise.
11011 (*negdi_zero_extend): Likewise.
11012 (*arm_abssi2): Likewise.
11013 (*thumb1_abssi2): Likewise.
11014 (*arm_neg_abssi2): Likewise.
11015 (*thumb1_neg_abss): Likewise.
11016 (one_cmpldi2): Likewise.
11017 (extend<mode>di2): Likewise.
11018 (*compareqi_eq0): Likewise.
11019 (*arm_extendhisi2addsi): Likewise.
11020 (*arm_movdi): Likewise.
11021 (*thumb1_movdi_insn): Likewise.
11022 (*arm_movt): Likewise.
11023 (*thumb1_movsi_insn): Likewise.
11024 (pic_add_dot_plus_four): Likewise.
11025 (pic_add_dot_plus_eight): Likewise.
11026 (tls_load_dot_plus_eight): Likewise.
11027 (*thumb1_movhi_insn): Likewise.
11028 (*thumb1_movsf_insn): Likewise.
11029 (*movdf_soft_insn): Likewise.
11030 (*thumb_movdf_insn): Likewise.
11031 (cbranchsi4_insn): Likewise.
11032 (cbranchsi4_scratch): Likewise.
11033 (*negated_cbranchsi4): Likewise.
11034 (*tbit_cbranch): Likewise.
11035 (*tlobits_cbranch): Likewise.
11036 (*tstsi3_cbranch): Likewise.
11037 (*cbranchne_decr1): Likewise.
11038 (*addsi3_cbranch): Likewise.
11039 (*addsi3_cbranch_scratch): Likewise.
11040 (*arm_cmpdi_insn): Likewise.
11041 (*arm_cmpdi_unsig): Likewise.
11042 (*arm_cmpdi_zero): Likewise.
11043 (*thumb_cmpdi_zero): Likewise.
11044 (*deleted_compare): Likewise.
11045 (*mov_scc): Likewise.
11046 (*mov_negscc): Likewise.
11047 (*mov_notscc): Likewise.
11048 (*cstoresi_eq0_thumb1_insn): Likewise.
11049 (cstoresi_nltu_thumb1): Likewise.
11050 (cstoresi_ltu_thu): Likewise.
11051 (thumb1_addsi3_addgeu): Likewise.
11052 (*arm_jump): Likewise.
11053 (*thumb_jump): Likewise.
11054 (*check_arch2): Likewise.
11055 (arm_casesi_internal): Likewise.
11056 (thumb1_casesi_dispatch): Likewise.
11057 (*arm_indirect_jump): Likewise.
11058 (*thumb1_indirect_jump): Likewise.
11060 (*and_scc): Likewise.
11061 (*ior_scc): Likewise.
11062 (*compare_scc): Likewise.
11063 (*cond_move): Likewise.
11064 (*cond_arith): Likewise.
11065 (*cond_sub): Likewise.
11066 (*cmp_ite0): Likewise.
11067 (*cmp_ite1): Likewise.
11068 (*cmp_and): Likewise.
11069 (*cmp_ior): Likewise.
11070 (*ior_scc_scc): Likewise.
11071 (*ior_scc_scc_cmp): Likewise.
11072 (*and_scc_scc): Likewise.
11073 (*and_scc_scc_cmp): Likewise.
11074 (*and_scc_scc_nod): Likewise.
11075 (*negscc): Likewise.
11076 (movcond_addsi): Likewise.
11077 (movcond): Likewise.
11078 (*ifcompare_plus_move): Likewise.
11079 (*if_plus_move): Likewise.
11080 (*ifcompare_move_plus): Likewise.
11081 (*if_move_plus): Likewise.
11082 (*ifcompare_arith_arith): Likewise.
11083 (*if_arith_arith): Likewise.
11084 (*ifcompare_arith_move): Likewise.
11085 (*if_arith_move): Likewise.
11086 (*ifcompare_move_arith): Likewise.
11087 (*if_move_arith): Likewise.
11088 (*ifcompare_move_not): Likewise.
11089 (*if_move_not): Likewise.
11090 (*ifcompare_not_move): Likewise.
11091 (*if_not_move): Likewise.
11092 (*ifcompare_shift_move): Likewise.
11093 (*if_shift_move): Likewise.
11094 (*ifcompare_move_shift): Likewise.
11095 (*if_move_shift): Likewise.
11096 (*ifcompare_shift_shift): Likewise.
11097 (*ifcompare_not_arith): Likewise.
11098 (*ifcompare_arith_not): Likewise.
11099 (*if_arith_not): Likewise.
11100 (*ifcompare_neg_move): Likewise.
11101 (*if_neg_move): Likewise.
11102 (*ifcompare_move_neg): Likewise.
11103 (*if_move_neg): Likewise.
11104 (prologue_thumb1_interwork): Likewise.
11105 (*cond_move_not): Likewise.
11106 (*sign_extract_onebit): Likewise.
11107 (*not_signextract_onebit): Likewise.
11108 (stack_tie): Likewise.
11109 (align_4): Likewise.
11110 (align_8): Likewise.
11111 (consttable_end): Likewise.
11112 (consttable_1): Likewise.
11113 (consttable_2): Likewise.
11114 (consttable_4): Likewise.
11115 (consttable_8): Likewise.
11116 (consttable_16): Likewise.
11117 (*thumb1_tablejump): Likewise.
11118 (prefetch): Likewise.
11119 (force_register_use): Likewise.
11120 (thumb_eh_return): Likewise.
11121 (load_tp_hard): Likewise.
11122 (load_tp_soft): Likewise.
11123 (tlscall): Likewise.
11124 (*arm_movtas_ze): Likewise.
11125 (*arm_rev): Likewise.
11126 (*arm_revsh): Likewise.
11127 (*arm_rev16): Likewise.
11128 * config/arm/thumb2.md
11129 (*thumb2_smaxsi3): Likewise.
11130 (*thumb2_sminsi3): Likewise.
11131 (*thumb32_umaxsi3): Likewise.
11132 (*thumb2_uminsi3): Likewise.
11133 (*thumb2_negdi2): Likewise.
11134 (*thumb2_abssi2): Likewise.
11135 (*thumb2_neg_abss): Likewise.
11136 (*thumb2_movsi_insn): Likewise.
11137 (tls_load_dot_plus_four): Likewise.
11138 (*thumb2_movhi_insn): Likewise.
11139 (*thumb2_mov_scc): Likewise.
11140 (*thumb2_mov_negs): Likewise.
11141 (*thumb2_mov_negs): Likewise.
11142 (*thumb2_mov_nots): Likewise.
11143 (*thumb2_mov_nots): Likewise.
11144 (*thumb2_movsicc_): Likewise.
11145 (*thumb2_movsfcc_soft_insn): Likewise.
11146 (*thumb2_indirect_jump): Likewise.
11147 (*thumb2_and_scc): Likewise.
11148 (*thumb2_ior_scc): Likewise.
11149 (*thumb2_ior_scc_strict_it): Likewise.
11150 (*thumb2_cond_move): Likewise.
11151 (*thumb2_cond_arith): Likewise.
11152 (*thumb2_cond_ari): Likewise.
11153 (*thumb2_cond_sub): Likewise.
11154 (*thumb2_negscc): Likewise.
11155 (*thumb2_movcond): Likewise.
11156 (thumb2_casesi_internal): Likewise.
11157 (thumb2_casesi_internal_pic): Likewise.
11158 (*thumb2_alusi3_short): Likewise.
11159 (*thumb2_mov<mode>_shortim): Likewise.
11160 (*thumb2_addsi_short): Likewise.
11161 (*thumb2_subsi_short): Likewise.
11162 (thumb2_addsi3_compare0): Likewise.
11163 (*thumb2_cbz): Likewise.
11164 (*thumb2_cbnz): Likewise.
11165 (*thumb2_one_cmplsi2_short): Likewise.
11166 (*thumb2_negsi2_short): Likewise.
11167 (*orsi_notsi_si): Likewise.
11168 * config/arm/arm1020e.md: Update with new attributes.
11169 * config/arm/arm1026ejs.md: Update with new attributes.
11170 * config/arm/arm1136jfs.md: Update with new attributes.
11171 * config/arm/arm926ejs.md: Update with new attributes.
11172 * config/arm/cortex-a15.md: Update with new attributes.
11173 * config/arm/cortex-a5.md: Update with new attributes.
11174 * config/arm/cortex-a53.md: Update with new attributes.
11175 * config/arm/cortex-a7.md: Update with new attributes.
11176 * config/arm/cortex-a8.md: Update with new attributes.
11177 * config/arm/cortex-a9.md: Update with new attributes.
11178 * config/arm/cortex-m4.md: Update with new attributes.
11179 * config/arm/cortex-r4.md: Update with new attributes.
11180 * config/arm/fa526.md: Update with new attributes.
11181 * config/arm/fa606te.md: Update with new attributes.
11182 * config/arm/fa626te.md: Update with new attributes.
11183 * config/arm/fa726te.md: Update with new attributes.
11185 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
11187 * config/aarch64/aarch64-simd.md
11188 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
11189 <vwx> iterator to ensure correct register choice.
11190 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
11191 (aarch64_sqdmull_n<mode>): Likewise.
11192 (aarch64_sqdmull2_n<mode>_internal): Likewise.
11193 * config/aarch64/arm_neon.h
11194 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
11195 (vml<as><q>_n_<su>16): Likewise.
11196 (vml<as>l_high_lane<q>_<su>16): Likewise.
11197 (vml<as>l_high_n_<su>16): Likewise.
11198 (vml<as>l_lane<q>_<su>16): Likewise.
11199 (vml<as>l_n_<su>16): Likewise.
11200 (vmul<q>_lane<q>_<su>16): Likewise.
11201 (vmul<q>_n_<su>16): Likewise.
11202 (vmull_lane<q>_<su>16): Likewise.
11203 (vmull_n_<su>16): Likewise.
11204 (vmull_high_lane<q>_<su>16): Likewise.
11205 (vmull_high_n_<su>16): Likewise.
11206 (vqrdmulh<q>_n_s16): Likewise.
11208 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
11210 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
11211 have the correct lane parameter.
11213 2013-09-06 Richard Biener <rguenther@suse.de>
11215 * cfganal.c (control_dependences::~control_dependences):
11216 Properly free all of the vector.
11218 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
11221 * config/i386/i386.c (ix86_conditional_register_usage):
11222 Proper initialize extended SSE registers.
11224 2013-09-06 Jan Hubicka <jh@suse.cz>
11226 PR tree-optimization/58311
11227 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
11229 2013-09-06 Jan Hubicka <jh@suse.cz>
11231 * Makefile.in (tree-sra.o): Update dependencies.
11232 * tree-sra.c: Include ipa-utils.h
11233 (scan_function): Use recursive_call_p.
11234 (has_caller_p): New function.
11235 (cgraph_for_node_and_aliases): Count also callers of aliases.
11237 2013-09-06 Jan Hubicka <jh@suse.cz>
11239 PR middle-end/58094
11240 * cgraph.h (symtab_semantically_equivalent_p): Declare.
11241 * tree-tailcall.c: Include ipa-utils.h.
11242 (find_tail_calls): Use it.
11243 * ipa-pure-const.c (check_call): Likewise.
11244 * ipa-utils.c (recursive_call_p): New function.
11245 * ipa-utils.h (recursive_call_p): Dclare.
11246 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
11247 (symtab_semantically_equivalent_p): New function.
11248 * Makefile.in (tree-tailcall.o): Update dependencies.
11250 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
11252 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
11253 non-inlinable part.
11255 2013-09-06 Richard Biener <rguenther@suse.de>
11257 * lto-streamer.h (lto_global_var_decls): Remove.
11258 * Makefile.in (OBJS): Remove lto-symtab.o.
11259 (lto-symtab.o): Remove.
11260 (GTFILES): Remove lto-symtab.c
11261 * lto-symtab.c: Move to lto/
11263 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11265 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
11266 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
11267 (UNSPEC_FPINT_RINT): New constant definitions.
11268 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
11269 definition with 2 attributes.
11270 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
11271 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
11274 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11276 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
11277 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
11278 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
11279 attribute to "z196_alone".
11280 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
11283 2013-09-06 Richard Biener <rguenther@suse.de>
11285 * basic-block.h (class control_dependences): New.
11286 * tree-ssa-dce.c (control_dependence_map): Remove.
11288 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
11289 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
11290 find_pdom, find_control_dependence, find_all_control_dependences):
11292 (mark_control_dependent_edges_necessary,
11293 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
11294 tree_dce_done, perform_tree_ssa_dce): Adjust.
11295 * cfganal.c (set_control_dependence_map_bit,
11296 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
11297 find_all_control_dependences): Move from tree-ssa-dce.c and
11298 implement as methods of control_dependences class.
11299 (control_dependences::control_dependences): New.
11300 (control_dependences::~control_dependences): Likewise.
11301 (control_dependences::get_edges_dependent_on): Likewise.
11302 (control_dependences::get_edge): Likewise.
11304 2013-09-04 Jan Hubicka <jh@suse.cz>
11306 * tree.c (types_same_for_odr): Drop overactive check.
11307 * ipa-devirt.c (hash_type_name): Likewise.
11309 2013-09-04 Jan Hubicka <jh@suse.cz>
11311 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
11312 (analyze_functions): ... here.
11314 2013-09-04 Jan Hubicka <jh@suse.cz>
11316 PR middle-end/58201
11317 * cgraphunit.c (analyze_functions): Clear AUX fields
11318 after processing; initialize assembler name has.
11320 2013-09-05 Jeff Law <law@redhat.com>
11322 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
11323 from thread_around_empty_block. Record threading path into PATH.
11324 Recurse if threading through the initial block is successful.
11325 (thread_across_edge): Corresponding changes to slightly simplify.
11327 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11329 * config/aarch64/aarch64.md
11330 (type): Remove frecpe, frecps, frecpx.
11331 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
11332 fix to be a TARGET_SIMD instruction.
11333 (aarch64_frecps): Remove.
11334 * config/aarch64/aarch64-simd.md
11335 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
11336 (aarch64_frecps<mode>): Handle all float/vector of float modes.
11338 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11339 Sofiane Naci <sofiane.naci@arm.com>
11341 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
11342 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
11343 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
11344 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
11345 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
11346 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
11347 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
11348 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
11349 into "clz, "rbit". Rename "shift" to "shift_imm".
11350 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
11351 changes. Update for attribute changes all occurrences of arlo_* and
11353 * config/arm/arm-fixed.md: Update for attribute changes
11354 all occurrences of arlo_* types.
11355 * config/arm/thumb2.md: Update for attribute changes all occurrences
11357 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
11358 (cortexa7_older_only): Likewise.
11359 (cortexa7_younger): Likewise.
11360 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
11361 (1020alu_shift_op): Likewise.
11362 (1020alu_shift_reg_op): Likewise.
11363 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
11364 (alu_shift_op): Likewise.
11365 (alu_shift_reg_op): Likewise.
11366 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
11367 (11_alu_shift_op): Likewise.
11368 (11_alu_shift_reg_op): Likewise.
11369 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
11370 (9_alu_shift_reg_op): Likewise.
11371 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
11373 (cortex_a15_alu_shift): Likewise.
11374 (cortex_a15_alu_shift_reg): Likewise.
11375 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
11377 (cortex_a5_alu_shift): Likewise.
11378 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
11380 (cortex_a53_alu_shift): Likewise.
11381 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
11383 (cortex_a7_alu_reg): Likewise.
11384 (cortex_a7_alu_shift): Likewise.
11385 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
11387 (cortex_a8_alu_shift): Likewise.
11388 (cortex_a8_alu_shift_reg): Likewise.
11389 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
11390 (cortex_a9_dp_shift): Likewise.
11391 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
11393 * config/arm/cortex-r4.md
11394 (cortex_r4_alu): Update for attribute changes.
11395 (cortex_r4_mov): Likewise.
11396 (cortex_r4_alu_shift_reg): Likewise.
11397 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
11398 (526_alu_shift_op): Likewise.
11399 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
11400 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
11401 (626te_alu_shift_op): Likewise.
11402 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
11403 (726te_alu_shift_op): Likewise.
11404 (726te_alu_shift_reg_op): Likewise.
11405 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
11406 (mp626_alu_shift_op): Likewise.
11407 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
11408 (pj4_alu_conds): Likewise.
11409 (pj4_shift): Likewise.
11410 (pj4_shift_conds): Likewise.
11411 (pj4_alu_shift): Likewise.
11412 (pj4_alu_shift_conds): Likewise.
11413 * config/aarch64/aarch64.md: Update for attribute change
11414 all occurrences of arlo_* and shift* types.
11416 2013-09-05 Mike Stump <mikestump@comcast.net>
11418 * tree.h: Move documentation for tree_function_decl to tree-core.h
11419 with the declaration.
11421 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
11424 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
11425 looking for widest mode.
11427 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
11429 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
11431 2013-09-05 Richard Biener <rguenther@suse.de>
11433 PR tree-optimization/58137
11434 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
11435 Do not create vectors of pointers.
11436 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
11437 types for the components of the vector initializer.
11438 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
11439 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
11441 2013-09-05 Martin Jambor <mjambor@suse.cz>
11443 * ipa-prop.c (remove_described_reference): Accept missing references,
11444 return false if that hppens, otherwise return true.
11445 (cgraph_node_for_jfunc): New function.
11446 (try_decrement_rdesc_refcount): Likewise.
11447 (try_make_edge_direct_simple_call): Use them.
11448 (ipa_edge_removal_hook): Remove references from rdescs.
11449 (ipa_edge_duplication_hook): Clone rdescs and their references
11450 when the new edge has the same caller as the old one.
11451 * cgraph.c (cgraph_resolve_speculation): Remove speculative
11452 reference before removing any edges.
11454 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
11456 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
11458 * thumb2.md (thumb2_storewb_parisi): New pattern.
11460 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
11462 * config/aarch64/aarch64-option-extensions.def: Add
11463 AARCH64_OPT_EXTENSION of 'crc'.
11464 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
11465 (AARCH64_ISA_CRC): Ditto.
11466 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
11467 description of the CRC extension.
11469 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
11471 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
11472 * config/rs6000/linux.h: Ditto.
11473 * alpha/linux.h: Ditto.
11474 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
11475 no_c99_libc_has_function.
11476 * config/c6x/uclinux-elf.h: Ditto.
11477 * config/lm32/uclinux-elf.h: Ditto.
11478 * config/m68k/uclinux.h: Ditto.
11479 * config/moxie/uclinux.h: Ditto.
11480 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
11481 (crisv32-*-linux*, cris-*-linux*): Ditto.
11482 * config/bfin/bfin.c: Include "tm_p.h".
11484 2013-09-05 Richard Biener <rguenther@suse.de>
11486 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
11487 check for a definition without a basic-block.
11489 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
11490 Sofiane Naci <sofiane.naci@arm.com>
11492 * config/aarch64/aarch64.md
11493 (*movti_aarch64): Rename r_2_f and f_2_r.
11494 (*movsf_aarch64): Likewise.
11495 (*movdf_aarch64): Likewise.
11496 (*movtf_aarch64): Likewise.
11497 (aarch64_movdi_<mode>low): Likewise.
11498 (aarch64_movdi_<mode>high): Likewise.
11499 (aarch64_mov<mode>high_di): Likewise.
11500 (aarch64_mov<mode>low_di): Likewise.
11501 (aarch64_movtilow_tilow): Likewise.
11502 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
11503 values to config/arm/types.md
11504 (attribute "conds"): Update for attribute change.
11505 (anddi3_insn): Likewise.
11506 (iordi3_insn): Likewise.
11507 (xordi3_insn): Likewise.
11508 (one_cmpldi2): Likewise.
11509 * config/arm/types.md (type): Add Neon types.
11510 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
11511 use "type" attribute.
11512 (movmisalign<mode>_neon_store): Likewise.
11513 (movmisalign<mode>_neon_load): Likewise.
11514 (vec_set<mode>_internal): Likewise.
11515 (vec_setv2di_internal): Likewise.
11516 (vec_extract<mode>): Likewise.
11517 (vec_extractv2di): Likewise.
11518 (add<mode>3_neon): Likewise.
11519 (adddi3_neon): Likewise.
11520 (sub<mode>3_neon): Likewise.
11521 (subdi3_neon): Likewise.
11522 (mul<mode>3_neon): Likewise.
11523 (mul<mode>3add<mode>_neon): Likewise.
11524 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
11525 (fma<VCVTF:mode>4)): Likewise.
11526 (fma<VCVTF:mode>4_intrinsic): Likewise.
11527 (fmsub<VCVTF:mode>4)): Likewise.
11528 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
11529 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
11530 (ior<mode>3): Likewise.
11531 (and<mode>3): Likewise.
11532 (anddi3_neon): Likewise.
11533 (orn<mode>3_neon): Likewise.
11534 (orndi3_neon): Likewise.
11535 (bic<mode>3_neon): Likewise.
11536 (bicdi3_neon): Likewise.
11537 (xor<mode>3): Likewise.
11538 (one_cmpl<mode>2): Likewise.
11539 (abs<mode>2): Likewise.
11540 (neg<mode>2): Likewise.
11541 (umin<mode>3_neon): Likewise.
11542 (umax<mode>3_neon): Likewise.
11543 (smin<mode>3_neon): Likewise.
11544 (smax<mode>3_neon): Likewise.
11545 (vashl<mode>3): Likewise.
11546 (vashr<mode>3_imm): Likewise.
11547 (vlshr<mode>3_imm): Likewise.
11548 (ashl<mode>3_signed): Likewise.
11549 (ashl<mode>3_unsigned): Likewise.
11550 (neon_load_count): Likewise.
11551 (ashldi3_neon_noclobber): Likewise.
11552 (signed_shift_di3_neon): Likewise.
11553 (unsigned_shift_di3_neon): Likewise.
11554 (ashrdi3_neon_imm_noclobber): Likewise.
11555 (lshrdi3_neon_imm_noclobber): Likewise.
11556 (widen_ssum<mode>3): Likewise.
11557 (widen_usum<mode>3): Likewise.
11558 (quad_halves_<code>v4si): Likewise.
11559 (quad_halves_<code>v4sf): Likewise.
11560 (quad_halves_<code>v8hi): Likewise.
11561 (quad_halves_<code>v16qi): Likewise.
11562 (reduc_splus_v2di): Likewise.
11563 (neon_vpadd_internal<mode>): Likewise.
11564 (neon_vpsmin<mode>): Likewise.
11565 (neon_vpsmax<mode>): Likewise.
11566 (neon_vpumin<mode>): Likewise.
11567 (neon_vpumax<mode>): Likewise.
11568 (ss_add<mode>_neon): Likewise.
11569 (us_add<mode>_neon): Likewise.
11570 (ss_sub<mode>_neon): Likewise.
11571 (us_sub<mode>_neon): Likewise.
11572 (neon_vadd<mode>_unspec): Likewise.
11573 (neon_vaddl<mode>): Likewise.
11574 (neon_vaddw<mode>): Likewise.
11575 (neon_vhadd<mode>): Likewise.
11576 (neon_vqadd<mode>): Likewise.
11577 (neon_vaddhn<mode>): Likewise.
11578 (neon_vmul<mode>): Likewise.
11579 (neon_vmla<mode>): Likewise.
11580 (neon_vmlal<mode>): Likewise.
11581 (neon_vmls<mode>): Likewise.
11582 (neon_vmlsl<mode>): Likewise.
11583 (neon_vqdmulh<mode>): Likewise.
11584 (neon_vqdmlal<mode>): Likewise.
11585 (neon_vqdmlsl<mode>): Likewise.
11586 (neon_vmull<mode>): Likewise.
11587 (neon_vqdmull<mode>): Likewise.
11588 (neon_vsub<mode>_unspec): Likewise.
11589 (neon_vsubl<mode>): Likewise.
11590 (neon_vsubw<mode>): Likewise.
11591 (neon_vqsub<mode>): Likewise.
11592 (neon_vhsub<mode>): Likewise.
11593 (neon_vsubhn<mode>): Likewise.
11594 (neon_vceq<mode>): Likewise.
11595 (neon_vcge<mode>): Likewise.
11596 (neon_vcgeu<mode>): Likewise.
11597 (neon_vcgt<mode>): Likewise.
11598 (neon_vcgtu<mode>): Likewise.
11599 (neon_vcle<mode>): Likewise.
11600 (neon_vclt<mode>): Likewise.
11601 (neon_vcage<mode>): Likewise.
11602 (neon_vcagt<mode>): Likewise.
11603 (neon_vtst<mode>): Likewise.
11604 (neon_vabd<mode>): Likewise.
11605 (neon_vabdl<mode>): Likewise.
11606 (neon_vaba<mode>): Likewise.
11607 (neon_vabal<mode>): Likewise.
11608 (neon_vmax<mode>): Likewise.
11609 (neon_vmin<mode>): Likewise.
11610 (neon_vpaddl<mode>): Likewise.
11611 (neon_vpadal<mode>): Likewise.
11612 (neon_vpmax<mode>): Likewise.
11613 (neon_vpmin<mode>): Likewise.
11614 (neon_vrecps<mode>): Likewise.
11615 (neon_vrsqrts<mode>): Likewise.
11616 (neon_vqabs<mode>): Likewise.
11617 (neon_vqneg<mode>): Likewise.
11618 (neon_vcls<mode>): Likewise.
11619 (clz<mode>2): Likewise.
11620 (popcount<mode>2): Likewise.
11621 (neon_vrecpe): Likewise.
11622 (neon_vrsqrte): Likewise.
11623 (neon_vget_lane<mode>_sext_internal): Likewise.
11624 (neon_vget_lane<mode>_zext_internal): Likewise.
11625 (neon_vdup_n<mode>): Likewise.
11626 (neon_vdup_nv2di): Likewise.
11627 (neon_vdpu_lane<mode>_internal): Likewise.
11628 (neon_vswp<mode>): Likewise.
11629 (float<mode><V_cvtto>2): Likewise.
11630 (floatuns<mode><V_cvtto>2): Likewise.
11631 (fix_trunc<mode><V_cvtto>)2): Likewise
11632 (fixuns_trunc<mode><V_cvtto)2): Likewise.
11633 (neon_vcvt<mode>): Likewise.
11634 (neon_vcvtv4sfv4hf): Likewise.
11635 (neon_vcvtv4hfv4sf): Likewise.
11636 (neon_vcvt_n<mode>): Likewise.
11637 (neon_vmovn<mode>): Likewise.
11638 (neon_vqmovn<mode>): Likewise.
11639 (neon_vqmovun<mode>): Likewise.
11640 (neon_vmovl<mode>): Likewise.
11641 (neon_vmul_lane<mode>): Likewise.
11642 (neon_vmull_lane<mode>): Likewise.
11643 (neon_vqdmull_lane<mode>): Likewise.
11644 (neon_vqdmulh_lane<mode>): Likewise.
11645 (neon_vmla_lane<mode>): Likewise.
11646 (neon_vmlal_lane<mode>): Likewise.
11647 (neon_vqdmlal_lane<mode>): Likewise.
11648 (neon_vmls_lane<mode>): Likewise.
11649 (neon_vmlsl_lane<mode>): Likewise.
11650 (neon_vqdmlsl_lane<mode>): Likewise.
11651 (neon_vext<mode>): Likewise.
11652 (neon_vrev64<mode>): Likewise.
11653 (neon_vrev32<mode>): Likewise.
11654 (neon_vrev16<mode>): Likewise.
11655 (neon_vbsl<mode>_internal): Likewise.
11656 (neon_vshl<mode>): Likewise.
11657 (neon_vqshl<mode>): Likewise.
11658 (neon_vshr_n<mode>): Likewise.
11659 (neon_vshrn_n<mode>): Likewise.
11660 (neon_vqshrn_n<mode>): Likewise.
11661 (neon_vqshrun_n<mode>): Likewise.
11662 (neon_vshl_n<mode>): Likewise.
11663 (neon_vqshl_n<mode>): Likewise.
11664 (neon_vqshlu_n<mode>): Likewise.
11665 (neon_vshll_n<mode>): Likewise.
11666 (neon_vsra_n<mode>): Likewise.
11667 (neon_vsri_n<mode>): Likewise.
11668 (neon_vsli_n<mode>): Likewise.
11669 (neon_vtbl1v8qi): Likewise.
11670 (neon_vtbl2v8qi): Likewise.
11671 (neon_vtbl3v8qi): Likewise.
11672 (neon_vtbl4v8qi): Likewise.
11673 (neon_vtbx1v8qi): Likewise.
11674 (neon_vtbx2v8qi): Likewise.
11675 (neon_vtbx3v8qi): Likewise.
11676 (neon_vtbx4v8qi): Likewise.
11677 (neon_vtrn<mode>_internal): Likewise.
11678 (neon_vzip<mode>_internal): Likewise.
11679 (neon_vuzp<mode>_internal): Likewise.
11680 (neon_vld1<mode>): Likewise.
11681 (neon_vld1_lane<mode>): Likewise.
11682 (neon_vld1_dup<mode>): Likewise.
11683 (neon_vld1_dupv2di): Likewise.
11684 (neon_vst1<mode>): Likewise.
11685 (neon_vst1_lane<mode>): Likewise.
11686 (neon_vld2<mode>): Likewise.
11687 (neon_vld2_lane<mode>): Likewise.
11688 (neon_vld2_dup<mode>): Likewise.
11689 (neon_vst2<mode>): Likewise.
11690 (neon_vst2_lane<mode>): Likewise.
11691 (neon_vld3<mode>): Likewise.
11692 (neon_vld3qa<mode>): Likewise.
11693 (neon_vld3qb<mode>): Likewise.
11694 (neon_vld3_lane<mode>): Likewise.
11695 (neon_vld3_dup<mode>): Likewise.
11696 (neon_vst3<mode>): Likewise.
11697 (neon_vst3qa<mode>): Likewise.
11698 (neon_vst3qb<mode>): Likewise.
11699 (neon_vst3_lane<mode>): Likewise.
11700 (neon_vld4<mode>): Likewise.
11701 (neon_vld4qa<mode>): Likewise.
11702 (neon_vld4qb<mode>): Likewise.
11703 (neon_vld4_lane<mode>): Likewise.
11704 (neon_vld4_dup<mode>): Likewise.
11705 (neon_vst4<mode>): Likewise.
11706 (neon_vst4qa<mode>): Likewise.
11707 (neon_vst4qb<mode>): Likewise.
11708 (neon_vst4_lane<mode>): Likewise.
11709 (neon_vec_unpack<US>_lo_<mode>): Likewise.
11710 (neon_vec_unpack<US>_hi_<mode>): Likewise.
11711 (neon_vec_<US>mult_lo_<mode>): Likewise.
11712 (neon_vec_<US>mult_hi_<mode>): Likewise.
11713 (neon_vec_<US>shiftl_<mode>): Likewise.
11714 (neon_unpack<US>_<mode>): Likewise.
11715 (neon_vec_<US>mult_<mode>): Likewise.
11716 (vec_pack_trunc_<mode>): Likewise.
11717 (neon_vec_pack_trunk_<mode>): Likewise.
11718 (neon_vabd<mode>_2): Likewise.
11719 (neon_vabd<mode>_3): Likewise.
11720 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
11721 (thumb2_movsi_vfp): Likewise.
11722 (movdi_vfp): Likewise.
11723 (movdi_vfp_cortexa8): Likewise.
11724 (movhf_vfp_neon): Likewise.
11725 (movhf_vfp): Likewiwse.
11726 (movsf_vfp): Likewiwse.
11727 (thumb2_movsf_vfp): Likewiwse.
11728 (movdf_vfp): Likewise.
11729 (thumb2_movdf_vfp): Likewise.
11730 (movsfcc_vfp): Likewise.
11731 (thumb2_movsfcc_vfp): Likewise.
11732 (movdfcc_vfp): Likewise.
11733 (thumb2_movdfcc_vfp): Likewise.
11734 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
11735 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
11736 (v10_v2c): Likewise.
11737 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
11739 (cortex_a15_neon_int_2): Likewise.
11740 (cortex_a15_neon_int_3): Likewise.
11741 (cortex_a15_neon_int_4): Likewise.
11742 (cortex_a15_neon_int_5): Likewise.
11743 (cortex_a15_neon_vqneg_vqabs): Likewise.
11744 (cortex_a15_neon_vmov): Likewise.
11745 (cortex_a15_neon_vaba): Likewise.
11746 (cortex_a15_neon_vaba_qqq): Likewise.
11747 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11748 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
11749 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
11750 scalar_64_32_long_scalar): Likewise.
11751 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11752 (cortex_a15_neon_mla_qqq_8_16): Likewise.
11753 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
11754 lotype_qdd_64_32_long): Likewise.
11755 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
11756 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
11757 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
11758 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
11759 (cortex_a15_neon_shift_1): Likewise.
11760 (cortex_a15_neon_shift_2): Likewise.
11761 (cortex_a15_neon_shift_3): Likewise.
11762 (cortex_a15_neon_vshl_ddd): Likewise.
11763 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
11764 (cortex_a15_neon_vsra_vrsra): Likewise.
11765 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
11766 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
11767 (cortex_a15_neon_fp_vmul_ddd): Likewise.
11768 (cortex_a15_neon_fp_vmul_qqd): Likewise.
11769 (cortex_a15_neon_fp_vmla_ddd): Likewise.
11770 (cortex_a15_neon_fp_vmla_qqq): Likewise.
11771 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
11772 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
11773 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
11774 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
11775 (cortex_a15_neon_bp_simple): Likewise.
11776 (cortex_a15_neon_bp_2cycle): Likewise.
11777 (cortex_a15_neon_bp_3cycle): Likewise.
11778 (cortex_a15_neon_vld1_1_2_regs): Likewise.
11779 (cortex_a15_neon_vld1_3_4_regs): Likewise.
11780 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
11781 (cortex_a15_neon_vld2_4_regs): Likewise.
11782 (cortex_a15_neon_vld3_vld4): Likewise.
11783 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
11784 (cortex_a15_neon_vst1_3_4_regs): Likewise.
11785 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
11786 (cortex_a15_neon_vst3_vst4): Likewise.
11787 (cortex_a15_neon_vld1_vld2_lane): Likewise.
11788 (cortex_a15_neon_vld3_vld4_lane" 10
11789 (cortex_a15_neon_vst1_vst2_lane): Likewise.
11790 (cortex_a15_neon_vst3_vst4_lane): Likewise.
11791 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
11792 (cortex_a15_neon_ldm_2): Likewise.0
11793 (cortex_a15_neon_stm_2): Likewise.
11794 (cortex_a15_neon_mcr): Likewise.
11795 (cortex_a15_neon_mcr_2_mcrr): Likewise.
11796 (cortex_a15_neon_mrc): Likewise.
11797 (cortex_a15_neon_mrrc): Likewise.
11798 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11800 (cortex_a15_alu_shift): Likewise.
11801 (cortex_a15_alu_shift_reg): Likewise.
11802 (cortex_a15_mult32): Likewise.
11803 (cortex_a15_mult64): Likewise.
11804 (cortex_a15_block): Likewise.
11805 (cortex_a15_branch): Likewise.
11806 (cortex_a15_load1): Likewise.
11807 (cortex_a15_load3): Likewise.
11808 (cortex_a15_store1): Likewise.
11809 (cortex_a15_store3): Likewise.
11810 (cortex_a15_call): Likewise.
11811 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
11812 (cortex_a5_f2r): Likewise.
11813 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
11815 (cortex_a53_f2r): Likewise.
11816 * config/arm/cortex-a7.md
11817 (cortex_a7_branch): Update for attribute change.
11818 (cortex_a7_call): Likewise.
11819 (cortex_a7_alu_imm): Likewise.
11820 (cortex_a7_alu_reg): Likewise.
11821 (cortex_a7_alu_shift): Likewise.
11822 (cortex_a7_mul): Likewise.
11823 (cortex_a7_load1): Likewise.
11824 (cortex_a7_store1): Likewise.
11825 (cortex_a7_load2): Likewise.
11826 (cortex_a7_store2): Likewise.
11827 (cortex_a7_load3): Likewise.
11828 (cortex_a7_store3): Likewise.
11829 (cortex_a7_load4): Likewise.
11830 (cortex_a7_store4): Likewise.
11831 (cortex_a7_fpalu): Likewise.
11832 (cortex_a7_fconst): Likewise.
11833 (cortex_a7_fpmuls): Likewise.
11834 (cortex_a7_neon_mul): Likewise.
11835 (cortex_a7_fpmacs): Likewise.
11836 (cortex_a7_neon_mla: Likewise.
11837 (cortex_a7_fpmuld: Likewise.
11838 (cortex_a7_fpmacd: Likewise.
11839 (cortex_a7_fpfmad: Likewise.
11840 (cortex_a7_fdivs: Likewise.
11841 (cortex_a7_fdivd: Likewise.
11842 (cortex_a7_r2f: Likewise.
11843 (cortex_a7_f2r: Likewise.
11844 (cortex_a7_f_flags: Likewise.
11845 (cortex_a7_f_loads: Likewise.
11846 (cortex_a7_f_loadd: Likewise.
11847 (cortex_a7_f_stores: Likewise.
11848 (cortex_a7_f_stored: Likewise.
11849 (cortex_a7_neon): Likewise.
11850 * config/arm/cortex-a8-neon.md
11851 (cortex_a8_neon_mrc): Update for attribute change.
11852 (cortex_a8_neon_mrrc): Likewise.
11853 (cortex_a8_neon_int_1): Likewise.
11854 (cortex_a8_neon_int_2): Likewise.
11855 (cortex_a8_neon_int_3): Likewise.
11856 (cortex_a8_neon_int_4): Likewise.
11857 (cortex_a8_neon_int_5): Likewise.
11858 (cortex_a8_neon_vqneg_vqabs): Likewise.
11859 (cortex_a8_neon_vmov): Likewise.
11860 (cortex_a8_neon_vaba): Likewise.
11861 (cortex_a8_neon_vaba_qqq): Likewise.
11862 (cortex_a8_neon_vsma): Likewise.
11863 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11864 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
11865 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
11866 long_scalar): Likewise.
11867 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11868 (cortex_a8_neon_mla_qqq_8_16): Likewise.
11869 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
11870 long_scalar_qdd_64_32_long): Likewise.
11871 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
11872 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
11873 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
11874 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
11875 (cortex_a8_neon_shift_1): Likewise.
11876 (cortex_a8_neon_shift_2): Likewise.
11877 (cortex_a8_neon_shift_3): Likewise.
11878 (cortex_a8_neon_vshl_ddd): Likewise.
11879 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
11880 (cortex_a8_neon_vsra_vrsra): Likewise.
11881 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
11882 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
11883 (cortex_a8_neon_fp_vsum): Likewise.
11884 (cortex_a8_neon_fp_vmul_ddd): Likewise.
11885 (cortex_a8_neon_fp_vmul_qqd): Likewise.
11886 (cortex_a8_neon_fp_vmla_ddd): Likewise.
11887 (cortex_a8_neon_fp_vmla_qqq): Likewise.
11888 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
11889 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
11890 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
11891 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
11892 (cortex_a8_neon_bp_simple): Likewise.
11893 (cortex_a8_neon_bp_2cycle): Likewise.
11894 (cortex_a8_neon_bp_3cycle): Likewise.
11895 (cortex_a8_neon_ldr): Likewise.
11896 (cortex_a8_neon_str): Likewise.
11897 (cortex_a8_neon_vld1_1_2_regs): Likewise.
11898 (cortex_a8_neon_vld1_3_4_regs): Likewise.
11899 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
11900 (cortex_a8_neon_vld2_4_regs): Likewise.
11901 (cortex_a8_neon_vld3_vld4): Likewise.
11902 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
11903 (cortex_a8_neon_vst1_3_4_regs): Likewise.
11904 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
11905 (cortex_a8_neon_vst3_vst4): Likewise.
11906 (cortex_a8_neon_vld1_vld2_lane): Likewise.
11907 (cortex_a8_neon_vld3_vld4_lane): Likewise.
11908 (cortex_a8_neon_vst1_vst2_lane): Likewise.
11909 (cortex_a8_neon_vst3_vst4_lane): Likewise.
11910 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
11911 (cortex_a8_neon_mcr): Likewise.
11912 (cortex_a8_neon_mcr_2_mcrr): Likewise.
11913 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11914 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
11916 (ca9_neon_mrrc): Likewise.
11917 (cortex_a9_neon_int_1): Likewise.
11918 (cortex_a9_neon_int_2): Likewise.
11919 (cortex_a9_neon_int_3): Likewise.
11920 (cortex_a9_neon_int_4): Likewise.
11921 (cortex_a9_neon_int_5): Likewise.
11922 (cortex_a9_neon_vqneg_vqabs): Likewise.
11923 (cortex_a9_neon_vmov): Likewise.
11924 (cortex_a9_neon_vaba): Likewise.
11925 (cortex_a9_neon_vaba_qqq): Likewise.
11926 (cortex_a9_neon_vsma): Likewise.
11927 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11928 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
11929 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
11930 long_scalar): Likewise.
11931 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11932 (cortex_a9_neon_mla_qqq_8_16): Likewise.
11933 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
11934 long_scalar_qdd_64_32_long): Likewise.
11935 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
11936 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
11937 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
11938 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
11939 (cortex_a9_neon_shift_1): Likewise.
11940 (cortex_a9_neon_shift_2): Likewise.
11941 (cortex_a9_neon_shift_3): Likewise.
11942 (cortex_a9_neon_vshl_ddd): Likewise.
11943 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
11944 (cortex_a9_neon_vsra_vrsra): Likewise.
11945 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
11946 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
11947 (cortex_a9_neon_fp_vsum): Likewise.
11948 (cortex_a9_neon_fp_vmul_ddd): Likewise.
11949 (cortex_a9_neon_fp_vmul_qqd): Likewise.
11950 (cortex_a9_neon_fp_vmla_ddd): Likewise.
11951 (cortex_a9_neon_fp_vmla_qqq): Likewise.
11952 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
11953 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
11954 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
11955 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
11956 (cortex_a9_neon_bp_simple): Likewise.
11957 (cortex_a9_neon_bp_2cycle): Likewise.
11958 (cortex_a9_neon_bp_3cycle): Likewise.
11959 (cortex_a9_neon_ldr): Likewise.
11960 (cortex_a9_neon_str): Likewise.
11961 (cortex_a9_neon_vld1_1_2_regs): Likewise.
11962 (cortex_a9_neon_vld1_3_4_regs): Likewise.
11963 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
11964 (cortex_a9_neon_vld2_4_regs): Likewise.
11965 (cortex_a9_neon_vld3_vld4): Likewise.
11966 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
11967 (cortex_a9_neon_vst1_3_4_regs): Likewise.
11968 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
11969 (cortex_a9_neon_vst3_vst4): Likewise.
11970 (cortex_a9_neon_vld1_vld2_lane): Likewise.
11971 (cortex_a9_neon_vld3_vld4_lane): Likewise.
11972 (cortex_a9_neon_vst1_vst2_lane): Likewise.
11973 (cortex_a9_neon_vst3_vst4_lane): Likewise.
11974 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
11975 (cortex_a9_neon_mcr): Likewise.
11976 (cortex_a9_neon_mcr_2_mcrr): Likewise.
11977 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
11978 (cortex_a9_fps): Likewise.
11979 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
11981 (cortex_m4_fmuls): Likewise.
11982 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
11984 (cortex_r4_mrc): Likewise.
11985 * config/arm/iterators.md: Update comment referring to neon_type.
11986 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
11987 (iwmmxt_movsi_insn): Likewise.
11988 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
11990 (pj4_core_to_vfp): Likewise.
11991 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
11993 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
11994 (vfp_fstore): Likewise.
11995 * doc/md.texi: Change references to neon_type to refer to type.
11997 2013-09-04 Dodji Seketeli <dodji@redhat.com>
11999 * tree.h (DECL_BUILT_IN): Fix typo in comment.
12001 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
12003 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
12004 lglobl if not weak.
12006 2013-09-04 Easwaran Raman <eraman@google.com>
12008 PR middle-end/57370
12009 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
12010 (build_and_add_sum): Use it.
12011 (appears_later_in_bb): Simplify code.
12013 2013-09-04 Teresa Johnson <tejohnson@google.com>
12015 * dumpfile.c (dump_finish): Don't close stderr/stdout.
12017 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
12019 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
12021 2013-09-04 Jan Hubicka <jh@suse.cz>
12023 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
12024 * ipa-devirt.c: Include diganostic.h
12025 (odr_type_d): Add types and types_set.
12026 (hash_type_name): Work for types with vtables during LTO.
12027 (odr_hasher::remove): Fix comment; destroy types_set.
12028 (add_type_duplicate): New function,
12029 (get_odr_type): Use it.
12030 (dump_type_inheritance_graph): Dump type duplicates.
12031 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
12033 * tree.c (types_same_for_odr): Give exact answers on types with
12036 2013-09-04 Dodji Seketeli <dodji@redhat.com>
12038 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
12039 explaining their differences.
12041 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
12043 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
12045 2013-09-03 Jeff Law <law@redhat.com>
12047 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
12048 when not threading through a joiner block. Pass joiner/no joiner
12049 state to register_jump_thread.
12050 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
12051 state from argument rather than implying on path length.
12052 Dump the entire jump thread path into debugging dump.
12053 * tree-flow.h (register_jump_thread): Update prototype.
12055 2013-08-29 Xinliang David Li <davidxl@google.com>
12057 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
12058 Remove a trivial gcc_assert.
12060 2013-08-29 Xinliang David Li <davidxl@google.com>
12062 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
12063 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
12064 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
12065 Delay base decl alignment adjustment.
12066 * tree-vectorizer.c (vect_destroy_datarefs): New function.
12067 * tree-vectorizer.h: New data structure.
12068 (set_dr_misalignment): New function.
12069 (dr_misalignment): Ditto.
12070 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
12071 (vectorizable_load): Ditto.
12072 (ensure_base_align): New function.
12073 (vectorize_loops): Add dbg_cnt support.
12074 (execute_vect_slp): Ditto.
12075 * dbgcnt.def: New debug counter.
12076 * Makefile: New dependency.
12078 2013-09-03 Meador Inge <meadori@codesourcery.com>
12082 2013-08-30 Meador Inge <meadori@codesourcery.com>
12084 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
12086 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
12088 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
12089 function descriptor.
12091 2013-09-03 Richard Biener <rguenther@suse.de>
12093 * tree-affine.c (add_elt_to_tree): Fix association issue,
12094 avoid useless converts and make sure to always return a
12095 properly typed result.
12097 2013-09-03 Richard Biener <rguenther@suse.de>
12099 PR middle-end/57656
12100 * fold-const.c (negate_expr_p): Fix division case.
12101 (negate_expr): Likewise.
12103 2013-09-03 Richard Biener <rguenther@suse.de>
12106 * tree-streamer-out.c: Include tm.h.
12107 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
12109 2013-09-03 Jan Hubicka <jh@suse.cz>
12111 * tree-profile.c (tree_profiling): Cleanup CFG when done.
12113 2013-09-03 Alan Modra <amodra@gmail.com>
12115 * config.gcc (powerpc*-*-linux*): Add support for little-endian
12116 multilibs to big-endian target and vice versa.
12117 * config/rs6000/t-linux64: Use := assignment on all vars.
12118 (MULTILIB_EXTRA_OPTS): Remove fPIC.
12119 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
12120 * config/rs6000/t-linux64le: New file.
12121 * config/rs6000/t-linux64bele: New file.
12122 * config/rs6000/t-linux64lebe: New file.
12124 2013-09-02 Jan Hubicka <jh@suse.cz>
12126 * ipa-inline-transform.c (inline_transform): Do not
12127 optimize_inline_calls when not optimizing.
12129 2013-09-02 Jan Hubicka <jh@suse.cz>
12131 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
12132 duplicated nodes for assembler names.
12133 * symtab.c (symtab_unregister_node): Do not attempt to unlink
12134 hard registers from assembler name hash.
12136 2013-09-02 Jan Hubicka <jh@suse.cz>
12138 * ipa-split.c (execute_split_functions): Split externally visible
12139 functions called once.
12141 2013-09-02 Martin Jambor <mjambor@suse.cz>
12144 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
12145 linked list. When finding the correct duplicate, also consider also
12146 the caller in additon to its inlined_to node.
12148 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
12150 * config/aarch64/aarch64-simd-builtins.def
12151 (dup_lane_scalar): Remove.
12152 * config/aarch64/aarch64-simd.md
12153 (aarch64_simd_dup): Add 'w->w' alternative.
12154 (aarch64_dup_lane<mode>): Allow for VALL.
12155 (aarch64_dup_lane_scalar<mode>): Remove.
12156 (aarch64_dup_lane_<vswap_width_name><mode>): New.
12157 (aarch64_get_lane_signed<mode>): Add w->w altenative.
12158 (aarch64_get_lane_unsigned<mode>): Likewise.
12159 (aarch64_get_lane<mode>): Likewise.
12160 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
12161 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
12162 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
12163 (VCON): Change container of V2SF.
12164 (vswap_width_name): Likewise.
12165 * config/aarch64/arm_neon.h
12166 (__aarch64_vdup_lane_any): New.
12167 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
12168 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
12169 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
12171 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
12173 PR middle-end/56382
12174 * expr.c (emit_move_complex): Do not move complex FP values as parts if
12175 the source or the destination is a single hard register.
12177 2013-09-02 Richard Biener <rguenther@suse.de>
12179 PR middle-end/57511
12180 * tree-scalar-evolution.c (instantiate_scev_name): Allow
12183 2013-09-02 Richard Biener <rguenther@suse.de>
12185 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
12186 arithmetic to sizetype.
12188 2013-09-02 Bin Cheng <bin.cheng@arm.com>
12190 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
12191 Find auto-increment use both before and after candidate.
12193 2013-09-02 Marek Polacek <polacek@redhat.com>
12195 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
12197 2013-09-01 Jan Hubicka <jh@suse.cz>
12199 * Makefile.in: Add ipa-profile.o
12200 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
12201 * cgraph.c (struct cgraph_propagate_frequency_data,
12202 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
12203 ipa-profile.c; replace cgraph_ by ipa_ prefix.
12204 * cgraph.h (cgraph_propagate_frequency): Remove.
12205 * ipa-inline-analysis.c: Include ipa-utils.h;
12206 drop duplicated cfgloop.h.
12207 (inline_update_callee_summaries): Update.
12208 * ipa-profile.c: New file.
12209 * ipa-utils.h (ipa_propagate_frequency): Declare.
12210 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
12211 data-streamer.h, value-prof.h.
12212 (symtab_remove_unreachable_nodes): Update profile.
12213 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
12214 account_time_size, cmp_counts, dump_histogram,
12215 ipa_profile_generate_summary, ipa_profile_write_summary,
12216 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
12217 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
12218 Move to ipa-profile.c.
12220 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
12222 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
12224 2013-09-01 Jan Hubicka <jh@suse.cz>
12226 * common.opt (fdevirtualize-speculatively): New function.
12227 * invoke.texi (fdevirtualize-speculatively): Document.
12228 * ipa-devirt.c: Include ipa-inline.h
12229 (likely_target_p): New function.
12230 (ipa_devirt): New function.
12231 (gate_ipa_devirt): New function.
12232 (pass_data_ipa_devirt): New static var.
12233 (pass_ipa_devirt): Likewise.
12234 (make_pass_ipa_devirt): New function.
12235 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
12236 (common_handle_option): Disable devirtualization when
12237 value range profiling is available.
12238 * passes.def (pass_ipa_devirt): Add.
12239 * timever.def (TV_IPA_DEVIRT): New timevar.
12240 * tree-pass.h (make_pass_ipa_devirt):
12242 2013-09-01 Iain Sandoe <iain@codesourcery.com>
12244 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
12245 include sanitize(undefined).
12247 2013-08-31 Diego Novillo <dnovillo@google.com>
12249 * Makefile.in (TREE_CORE_H): Define.
12251 (GTFILES): Add tree-core.h.
12252 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
12254 * tree-core.h: New file.
12255 Move all data structures, enum, typedefs, global
12256 declarations and constants from ...
12257 * tree.h: ... here.
12259 2013-08-31 Jan Hubicka <jh@suse.cz>
12261 * bulitins.c (expand_builtin): Do not early exit for gcov
12262 instrumented functions.
12264 2013-08-31 Marek Polacek <polacek@redhat.com>
12266 * ubsan.c: Include tm_p.h.
12268 2013-08-31 Jan Hubicka <jh@suse.cz>
12270 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
12273 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
12274 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
12275 * tree-cfg.c (verify_gimple_label): ... here.
12276 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
12277 (ipa_merge_profiles): New function.
12278 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
12279 (lto_input_function_body): Likewise.
12280 * ipa-utils.h (ipa_merge_profiles): Declare.
12281 * lto-streamer.h (lto_input_function_body): Update prototype.
12282 (emit_label_in_global_context_p): Remove.
12283 * lto-symtab.c: Include ipa-utils.h
12284 (lto_cgraph_replace_node): Use ipa_merge_profiles.
12286 2013-08-31 Jan Hubicka <jh@suse.cz>
12288 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
12290 2013-08-31 Jan Hubicka <jh@suse.cz>
12292 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
12294 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
12296 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
12297 "cmp" RTX before signed_comparison_operator check to account
12298 for "code" changes.
12300 2013-08-30 Jan Hubicka <jh@suse.cz>
12302 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
12303 (detect_type_change_1): Rename to ...
12304 (detect_type_change): ... this one; early return on non-polymorphic
12306 (detect_type_change_ssa): Add comp_type parameter; update
12307 use of detect_type_change.
12308 (compute_complex_assign_jump_func): Add param_type parameter;
12309 update use of detect_type_change_ssa.
12310 (compute_complex_ancestor_jump_func): Likewise.
12311 (ipa_get_callee_param_type): New function.
12312 (ipa_compute_jump_functions_for_edge): Compute parameter type;
12313 update calls to the jump function computation functions.
12315 2013-08-30 Teresa Johnson <tejohnson@google.com>
12316 Steven Bosscher <steven@gcc.gnu.org>
12318 * cfgrtl.c (fixup_new_cold_bb): New routine.
12319 (commit_edge_insertions): Invoke fixup_partitions.
12320 (find_partition_fixes): New routine.
12321 (fixup_partitions): Ditto.
12322 (verify_hot_cold_block_grouping): Update comments.
12323 (rtl_verify_edges): Invoke find_partition_fixes.
12324 (rtl_verify_bb_pointers): Update comments.
12325 (rtl_verify_bb_layout): Ditto.
12326 * basic-block.h (probably_never_executed_edge_p): Declare.
12327 (fixup_partitions): Ditto.
12328 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
12329 * bb-reorder.c (sanitize_hot_paths): New function.
12330 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
12331 sanitize_hot_paths.
12332 * predict.c (probably_never_executed_edge_p): New routine.
12333 * cfg.c (check_bb_profile): Add partition insanity warnings.
12335 2013-08-30 Meador Inge <meadori@codesourcery.com>
12337 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
12339 2013-08-30 Marek Polacek <polacek@redhat.com>
12341 * Makefile.in (ubsan.o): Add.
12342 (c-family/c-ubsan.o): Add.
12343 (builtins.o): Add ubsan.h dependency.
12344 * ubsan.h: New file.
12345 * ubsan.c: New file.
12346 * common.opt: Add -fsanitize=undefined option.
12347 (flag_sanitize): Add variable.
12348 (fsanitize=): Add option. Add Driver.
12349 (fsanitize=thread): Remove option.
12350 (fsanitize=address): Likewise.
12351 (static-libubsan): New option.
12352 * doc/invoke.texi: Document the new flag and -static-libubsan.
12353 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
12354 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
12355 * builtin-attrs.def (ATTR_COLD): Define.
12356 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
12357 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
12358 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
12359 * flag-types.h (sanitize_code): New enum.
12360 * opts.c (common_handle_option): Parse command line arguments
12361 of -fsanitize=. Add -fsanitize=unreachable option.
12362 * varasm.c (get_variable_section): Adjust.
12363 (assemble_noswitch_variable): Likewise.
12364 (assemble_variable): Likewise.
12365 (output_constant_def_contents): Likewise.
12366 (categorize_decl_for_section): Likewise.
12367 (place_block_symbol): Likewise.
12368 (output_object_block): Likewise.
12369 * builtins.def: Likewise.
12370 * toplev.c (compile_file): Likewise.
12371 (process_options): Likewise.
12372 * cppbuiltin.c: Likewise.
12373 * tsan.c (tsan_pass): Likewise.
12374 (tsan_gate): Likewise.
12375 (tsan_gate_O0): Likewise.
12376 * cfgexpand.c (partition_stack_vars): Likewise.
12377 (expand_stack_vars): Likewise.
12378 (defer_stack_allocation): Likewise.
12379 (expand_used_vars): Likewise.
12380 * cfgcleanup.c (old_insns_match_p): Likewise.
12381 * asan.c (asan_finish_file): Likewise.
12382 (asan_instrument): Likewise.
12383 (gate_asan): Likewise.
12384 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
12385 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
12386 (asan_global_struct): Use pointer_sized_int_node instead
12387 calling build_nonstandard_integer_type.
12388 (initialize_sanitizer_builtins): Likewise.
12389 (asan_finish_file): Likewise.
12390 * gcc.c: Document %{%:function(args):X}.
12391 (static_spec_functions): Add sanitize.
12392 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
12393 store funcval != NULL there.
12394 (do_spec_1): Adjust handle_spec_function caller.
12395 (handle_braces): Allow %:function(args) as condition.
12396 (sanitize_spec_function): New function.
12397 (ADD_STATIC_LIBUBSAN_LIBS): Define.
12398 (LIBUBSAN_SPEC): Likewise.
12399 (LIBUBSAN_EARLY_SPEC): Likewise.
12400 (SANITIZER_SPEC): Handle libubsan.
12401 (SANITIZER_EARLY_SPEC): Likewise.
12402 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
12403 instead of fsanitize=address.
12404 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
12405 instead of fsanitize=address*.
12406 * builtins.c: Include ubsan.h.
12407 (fold_builtin_0): Instrument __builtin_unreachable.
12408 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
12409 instead of flag_asan.
12410 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
12411 (pointer_sized_int_node): Define.
12412 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
12414 2013-08-30 Mike Stump <mikestump@comcast.net>
12416 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
12419 2013-08-29 Jan Hubicka <jh@suse.cz>
12421 * cgraph.c (cgraph_function_body_availability): Handle weakref
12423 * passes.def: Remove pass_fixup_cfg.
12424 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
12425 track when we need to remove functions.
12426 (gate_ipa_inline): Execute inlining always; add comment why.
12427 (pass_data_ipa_inline): Remove TODO_remove_functions.
12428 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
12429 do not produce summaries.
12430 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
12431 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
12432 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
12433 constant pool and vtable.
12435 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
12437 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
12438 New arm_neon.h's internal macros to specify 64-bit constants.
12439 Avoid using stdint.h's macros.
12441 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
12443 * recog.c (verify_changes): Verify that changes[i].old is non-zero
12444 before applying REG_P.
12446 2013-08-30 Jakub Jelinek <jakub@redhat.com>
12448 PR tree-optimization/58277
12449 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
12450 after seeing too many stmts with vdef in between dombb and current
12451 bb, invalidate everything.
12453 2013-08-30 Richard Biener <rguenther@suse.de>
12455 * fold-const.c (fold_single_bit_test): Fix overflow test.
12457 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
12459 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
12460 and which can live in a register, always retrieve the value on entry.
12461 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
12462 passed by invisible reference specially.
12463 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
12464 (vt_add_function_parameter): Correctly deal with a parameter passed by
12465 invisible reference.
12467 2013-08-30 Jan Hubicka <jh@suse.cz>
12469 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
12471 2013-08-30 Richard Biener <rguenther@suse.de>
12473 PR tree-optimization/58228
12474 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
12475 allow invariant loads in nested loop vectorization.
12477 2013-08-30 Richard Biener <rguenther@suse.de>
12479 PR tree-optimization/58223
12480 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
12481 (has_anti_or_output_dependence): ... this and adjust to also
12482 look for output dependences.
12483 (mark_nodes_having_upstream_mem_writes): Adjust.
12484 (rdg_flag_uses): Likewise.
12486 2013-08-30 Richard Biener <rguenther@suse.de>
12488 PR tree-optimization/58010
12489 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
12490 assert that we have a loop-closed PHI.
12492 2013-08-29 Jan Hubicka <jh@suse.cz>
12494 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
12495 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
12496 * lto-section-in.c (lto_free_function_in_decl_state): New function.
12497 (lto_free_function_in_decl_state_for_node): New function.
12499 2013-08-29 Xinliang David Li <davidxl@google.com>
12501 * loop-unroll.c (report_unroll_peel): Minor message change.
12502 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
12503 Emit alignment peeling message with default -fopt-info.
12504 (vect_loop_versioning): Emit loop version info message.
12505 * tree-vectorizer.c (vectorize_loops): Minor message change.
12506 (execute_vect_slp): Ditto.
12508 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
12510 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
12511 of the clone from the DECL_NAME of the original function.
12513 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
12515 * passes.c (register_pass): Add overload.
12516 * tree-pass.h (register_pass): Forward declare it. Add comment.
12518 2013-08-29 Jan Hubicka <jh@suse.cz>
12520 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
12521 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
12522 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
12523 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
12524 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
12525 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
12526 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
12527 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
12529 2013-08-29 Teresa Johnson <tejohnson@google.com>
12531 * dumpfile.c (dump_loc): Output column number.
12532 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
12533 * doc/invoke.texi: Document optall -fopt-info flag.
12534 * profile.c (read_profile_edge_counts): Use new dump framework.
12535 (compute_branch_probabilities): Ditto.
12536 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
12537 when pass not in any opt group.
12538 * pass_manager.h (pass_manager::get_pass_profile): New method.
12539 * value-prof.c (check_counter): Use new dump framework.
12540 (check_ic_target): Ditto.
12541 * coverage.c (get_coverage_counts): Ditto.
12542 (coverage_init): Setup new dump framework.
12544 2013-08-29 Richard Biener <rguenther@suse.de>
12546 PR tree-optimization/58246
12547 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
12548 handle the dominance check inside a basic-block.
12550 2013-08-29 Richard Biener <rguenther@suse.de>
12552 PR middle-end/57287
12553 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
12554 of default defs that appear in abnormal PHI nodes.
12556 2013-08-29 Richard Biener <rguenther@suse.de>
12558 PR tree-optimization/57685
12559 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
12560 single-use operands to avoid exponential complexity.
12562 2013-08-28 Dehao Chen <dehao@google.com>
12564 * ipa-inline.c (edge_badness): Fix integer underflow.
12566 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
12568 * gtm-builtins.def (_ITM_free): Declare leaf.
12570 2013-08-28 Jakub Jelinek <jakub@redhat.com>
12573 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
12574 (*tls_local_dynamic_base_64_largepic): Likewise.
12575 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
12576 Remove predicate from call operand.
12577 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
12578 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
12580 2013-08-28 Jeff Law <law@redhat.com>
12582 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
12583 checks for the number of predecessors and successors allowed.
12584 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
12585 which require copying a joiner block if there is a request which
12586 is a subpath that requires no joiner block copying.
12588 2013-08-28 Jan Hubicka <jh@suse.cz>
12590 * lto-streamer-out.c (DFS_write_tree_body): Drop
12591 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
12592 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
12593 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
12594 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
12595 Do not read DECL_ERROR_ISSUED.
12596 (unpack_ts_decl_with_vis_value_fields): Do not read
12598 (lto_input_ts_binfo_tree_pointers): Do not read
12599 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
12600 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
12601 write DECL_ERROR_ISSUED..
12602 (pack_ts_decl_with_vis_value_fields): Do not write
12604 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
12605 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
12606 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
12607 * tree.h (tree_decl_common): Update comment.
12608 (DECL_ERROR_ISSUED): Remove.
12610 2013-08-28 Jakub Jelinek <jakub@redhat.com>
12612 PR middle-end/58257
12613 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
12615 2013-08-28 Jan Hubicka <jh@suse.cz>
12617 * builtins.def (free): Declare leaf.
12619 2013-08-27 David Malcolm <dmalcolm@redhat.com>
12621 * gdbhooks.py: New.
12622 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
12623 * configure: Regenerate.
12625 2013-08-27 Martin Jambor <mjambor@suse.cz>
12627 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
12628 (ipa_ancestor_jf_data): Likewise.
12629 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
12630 (ipa_get_jf_pass_through_type_preserved): New function.
12631 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
12632 (ipa_get_jf_ancestor_type_preserved): New function.
12633 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
12634 (ipa_get_jf_ancestor_result): Likewise.
12635 (propagate_vals_accross_pass_through): Use
12636 ipa_get_jf_pass_through_result to do all the value mappings.
12637 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
12638 type_preserved flag.
12639 (ipa_set_jf_cst_copy): New function.
12640 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
12641 (ipa_set_jf_arith_pass_through): Likewise.
12642 (ipa_set_ancestor_jf): Likewise.
12643 (compute_complex_assign_jump_func): Set type_preserved instead of
12645 (ipa_compute_jump_functions_for_edge): Likewise.
12646 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
12647 (update_jump_functions_after_inlining): Update type_preserved.
12648 Explicitely create jump functions when combining one with pass_through.
12649 (ipa_write_jump_function): Stream the type_preserved flags.
12650 (ipa_read_jump_function): Likewise.
12652 2013-08-27 Jakub Jelinek <jakub@redhat.com>
12653 Aldy Hernandez <aldyh@redhat.com>
12655 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
12656 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
12657 * function.h (struct function): Add has_force_vect_loops and
12659 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
12660 * gimple.c (gimple_build_omp_critical): Add KIND argument and
12662 * gimple.def: Update CLAUSES comments.
12663 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
12664 (gimple_build_omp_for): Add argument to prototype.
12665 (gimple_omp_for_kind): New.
12666 (gimple_omp_for_set_kind): New.
12667 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
12668 GOVD_DATA_SHARE_CLASS.
12669 (enum omp_region_type): Add ORT_SIMD.
12670 (gimple_add_tmp_var): Handle ORT_SIMD.
12671 (gimplify_var_or_parm_decl): Same.
12672 (is_gimple_stmt): Same.
12673 (omp_firstprivatize_variable): Same.
12674 (omp_add_variable): Only use splay_tree_insert if lookup failed.
12675 (omp_notice_variable): Handle ORT_SIMD.
12676 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
12677 (omp_check_private): Handle ORT_SIMD.
12678 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
12679 OMP_CLAUSE_SAFELEN.
12680 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
12681 Handle OMP_CLAUSE_LASTPRIVATE.
12682 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
12683 OMP_CLAUSE_SAFELEN.
12684 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
12685 (gimplify_expr): Handle OMP_SIMD.
12686 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
12687 (expand_GOMP_SIMD_VF): New.
12688 (expand_GOMP_SIMD_LAST_LANE): New.
12689 * internal-fn.def (GOMP_SIMD_LANE): New.
12690 (GOMP_SIMD_VF): New.
12691 (GOMP_SIMD_LAST_LANE): New.
12692 * omp-low.c: Include target.h.
12693 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
12694 OMP_CLAUSE_SAFELEN.
12695 (check_omp_nesting_restrictions): Same.
12697 (lower_rec_simd_input_clauses): New.
12698 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
12700 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
12701 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
12702 (expand_omp_build_assign): New.
12703 (expand_omp_for_init_counts): New.
12704 (expand_omp_for_init_vars): New.
12705 (extract_omp_for_update_vars): New.
12706 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
12707 and rewrite accordingly.
12708 (expand_omp_simd): New.
12709 (expand_omp_for): Use expand_omp_simd.
12710 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
12711 (lower_omp_for): Do not lower the body.
12712 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
12713 in their own loops.
12714 * tree-flow.h (find_omp_clause): Remove prototype.
12715 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
12716 forcing vectorization of the loop, or if flag_tree_vectorize.
12717 (gate_tree_if_conversion): Similarly.
12718 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
12719 gimple_build_omp_for.
12720 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
12721 * tree-parloops (create_parallel_loop): Pass kind argument to
12722 gimple_build_omp_for.
12723 * tree-pretty-print.c (dump_omp_clause): Add cases for
12724 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
12725 OMP_CLAUSE__SIMDUID_.
12726 (dump_generic_node): Handle OMP_SIMD.
12727 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
12728 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
12729 unroll OMP_SIMD loops here.
12730 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
12731 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12733 (vect_analyze_data_refs): Handle simd loops.
12734 * tree-vect-loop.c (vectorizable_live_operation): Handle
12736 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
12737 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
12738 (vectorizable_load): Same.
12739 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
12740 (struct simduid_to_vf): New.
12741 (simduid_to_vf::hash): New.
12742 (simduid_to-vf::equal): New.
12743 (struct simd_array_to_simduid): New.
12744 (simd_array_to_simduid::hash): New.
12745 (simd_array_to_simduid::equal): New.
12746 (adjust_simduid_builtins): New.
12747 (struct note_simd_array_uses_struct): New.
12748 (note_simd_array_uses_cb): New.
12749 (note_simd_array_uses): New.
12750 (vectorize_loops): Handle simd hints and adjust simd builtins
12752 * tree-vectorizer.h (struct _stmt_vec_info): Add
12753 simd_lane_access_p field.
12754 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
12755 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
12756 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
12757 (omp_clause_code_name): Same.
12758 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
12759 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
12760 * tree.def (OMP_SIMD): New entry.
12761 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
12762 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
12763 (OMP_CLAUSE_DECL): Adjust range for new clauses.
12764 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
12765 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
12766 (OMP_CLAUSE_LINEAR_STEP): New.
12767 (OMP_CLAUSE_SAFELEN_EXPR): New.
12768 (OMP_CLAUSE__SIMDUID__DECL): New.
12769 (find_omp_clause): New prototype.
12771 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
12773 * config/i386/driver-i386.c (host_detect_local_cpu): Update
12774 Haswell processor detection.
12776 2013-08-27 Christian Widmer <shadow@umbrox.de>
12779 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
12780 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
12781 AVX2 capable processors.
12783 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
12785 * config/aarch64/arm_neon.h: Replace all inline asm implementations
12786 of vget_low_* with implementations in terms of other intrinsics.
12788 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
12790 PR middle-end/57219
12791 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
12792 values to -1, 0 and 1.
12794 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
12796 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
12797 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
12798 (<optab><mode>3_insn): Remove.
12799 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
12800 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
12801 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
12802 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
12803 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
12804 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
12805 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
12806 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
12807 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
12808 (ror<mode>3_insn): Likewise.
12809 * config/aarch64/predicates.md (aarch64_simd_register): New.
12811 2013-08-27 Richard Biener <rguenther@suse.de>
12813 PR tree-optimization/57521
12814 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
12815 one edge is non-critical.
12816 (find_phi_replacement_condition): Make sure to use a non-critical
12817 edge. Cleanup and remove old bug workarounds.
12818 (bb_postdominates_preds): Remove.
12819 (if_convertible_loop_p_1): Do not compute post-dominators.
12820 (combine_blocks): Do not free post-dominators.
12821 (main_tree_if_conversion): Likewise.
12822 (pass_data_if_conversion): Add TODO_verify_ssa.
12824 2013-08-27 DJ Delorie <dj@redhat.com>
12826 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
12828 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
12830 * function.c (assign_parm_find_data_types): Set passed_mode and
12831 nominal_mode to the TYPE_MODE of nominal_type for the built
12832 pointer type in case of the struct-pass-by-reference.
12834 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
12836 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
12837 (UINT16_TYPE): Change default to "unsigned int".
12839 * config/avr/avr.opt (mfract-convert-truncate): New option.
12840 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
12841 is set, round negative fractional integers according to n1169
12842 when converting to integer types.
12844 2013-08-26 Jan Hubicka <jh@suse.cz>
12846 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
12847 methods can not be called indirectly when their address is not taken.
12849 2013-08-26 Jan Hubicka <jh@suse.cz>
12851 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
12854 2013-08-26 Jan Hubicka <jh@suse.cz>
12856 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
12859 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
12861 * reload.c (find_valid_class): Allow classes that do not include
12862 FIRST_PSEUDO_REGISTER - 1.
12864 2013-08-26 Jan Hubicka <jh@suse.cz>
12866 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
12867 fix edge count/frequency when speculation failed; fix type check
12868 for the direct call.
12870 2013-08-26 Jan Hubicka <jh@suse.cz>
12872 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
12874 2013-08-26 Jan Hubicka <jh@suse.cz>
12876 * ipa-inline-transform.c (inline_transform): Be ready for basic block
12877 to be changed by edge redirection.
12879 2013-08-26 Jan Hubicka <jh@suse.cz>
12881 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
12882 formating; add sanity check.
12883 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
12884 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
12885 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
12886 (resolve_noninline_speculation): Update callee keys, too.
12888 2013-08-26 Jan Hubicka <jh@suse.cz>
12890 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
12891 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
12893 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
12895 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
12898 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
12900 * config/i386/i386.c (ix86_debug_options): Remove prototype.
12901 (x86_64_elf_select_section): Ditto.
12902 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
12904 (ix86_pass_by_reference): Ditto.
12905 (output_set_got): Ditto.
12906 (ix86_unary_operator_ok): Ditto.
12907 (ix86_expand_builtin): Ditto.
12909 2013-08-23 Jan Hubicka <jh@suse.cz>
12911 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
12913 2013-08-23 Jan Hubicka <jh@suse.cz>
12915 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
12916 (tree_decl_with_vis): Add FINAL field.
12918 2013-08-23 Jeff Law <law@redhat.com>
12920 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
12921 the debugging dump when the expression is fully redundant.
12923 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
12925 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
12926 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
12927 * pretty-print.c (pp_formatted_text_data): Likewise.
12928 (pp_write_text_to_stream): Likewise.
12929 (pp_write_text_as_dot_label_to_stream): Likewise.
12930 (pp_append_r): Likewise.
12931 (pp_format): Likewise.
12932 (pp_flush): Likewise.
12933 (pp_clear_output_area): Likewise.
12934 (pp_append_text): Likewise.
12935 (pp_formatted_text): Likewise.
12936 (pp_remaining_character_count_for_line): Likewise.
12937 (pp_newline): Likewise.
12938 (pp_character): Likewise.
12939 (output_buffer::~output_buffer): Define.
12940 (pretty_printer::~pretty_printer): Destruct output buffer.
12941 * pretty-print.h (output_buffer::~output_buffer): Declare.
12942 (pretty_printer::~pretty_printer): Declare virtual.
12944 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
12947 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
12948 HOST_WIDE_INT_M1U): New macros.
12949 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
12950 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
12951 unsigned -1 for lshift.
12952 * cse.c (cse_insn): Likewise.
12953 * double-int.c (rshift_double, lshift_double): Likewise.
12954 * builtins.c (fold_builtin_bitop): Likewise.
12955 * combine.c (force_to_mode): Likewise.
12956 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
12957 * simplify-rtx.c (simplify_const_unary_operation,
12958 simplify_const_binary_operation): Likewise.
12959 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
12960 check_va_list_escapes): Likewise.
12961 * rtlanal.c (nonzero_bits1): Likewise.
12962 * expmed.c (expand_smod_pow2): Likewise.
12963 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
12965 2013-08-23 Jan Hubicka <jh@suse.cz>
12967 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
12968 as having address taken.
12970 2013-08-23 Jan Hubicka <jh@suse.cz>
12972 * ipa-utils.h (method_class_type): Declare.
12973 * ipa-devirt.c (method_class_type): Export.
12975 * cgraphunit.c (analyze_functions): Do basic devirtualization;
12976 do not walk base classes of anonymous types.
12978 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
12980 PR rtl-optimization/58220
12981 PR regression/58221
12982 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
12983 handle SEQUENCE insns properly.
12985 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
12987 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
12989 (pp_newline_and_indent): Likewise.
12990 (pp_separate_with): Likewise.
12991 * pretty-print.c (pp_newline_and_flush): Define.
12992 (pp_newline_and_indent): Likewise.
12993 (pp_separate_with): Likewise.
12995 2013-08-23 Jakub Jelinek <jakub@redhat.com>
12998 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
12999 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
13001 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
13003 * config/i386/predicates.md (ext_sse_reg_operand): New.
13004 * config/i386/i386.md (*movti_internal): Use
13005 predicate to determine if EVEX is needed.
13006 (*movsi_internal): Ditto.
13007 (*movdf_internal): Ditto.
13008 (*movsf_internal): Ditto.
13009 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
13011 2013-08-23 Jakub Jelinek <jakub@redhat.com>
13013 PR tree-optimization/58209
13014 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
13015 (find_tail_calls): Give up for pointer result types if m is non-NULL.
13016 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
13017 emit POINTER_PLUS_EXPR.
13018 (create_tailcall_accumulator): For pointer result type accumulate in
13021 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
13023 * configure.ac: Add backslashes missing from the last change.
13024 * configure: Regenerate.
13026 2013-08-22 Jan Hubicka <jh@susue.cz>
13028 * ipa.c (function_and_variable_visibility): First remember function
13029 was global and then make it local.
13031 2013-08-22 Julian Brown <julian@codesourcery.com>
13033 * configure.ac: Add aarch64 to list of arches which use "nop" in
13035 * configure: Regenerate.
13037 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13039 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
13040 gnu_libc_has_function.
13041 * config/s390/tpf.h: Likewise.
13043 2013-08-22 Jan Hubicka <jh@susue.cz>
13045 * timevar.c (validate_phases): Add cast.
13047 2013-08-22 Jan Hubicka <jh@susue.cz>
13049 * timevar.c (validate_phases): Use size_t for memory.
13050 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
13052 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
13054 * pretty-print.h (output_buffer::output_buffer): Declare.
13055 (pretty_printer::pretty_printer): Likewise.
13056 (pp_construct): Remove.
13057 * pretty-print.c (output_buffer::output_buffer): Define.
13058 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
13059 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
13060 (print_gimple_expr): Likewise.
13061 (print_gimple_seq): Likewise.
13062 (gimple_dump_bb): Likewise.
13063 * sched-vis.c (dump_value_slim): Likewise.
13064 (dump_insn_slim): Likewise.
13065 (dump_rtl_slim): Likewise.
13066 (str_pattern_slim): Likewise.
13067 * tree-mudflap.c (mf_varname_tree): Likewise.
13068 * graph.c (print_graph_cfg): Likewise.
13069 (start_graph_dump): Likewise.
13070 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
13072 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
13073 pretty printer initialization.
13074 * coretypes.h (diagnostic_context): Remove superflous type alias
13076 (pretty_printer): Likewise. Declare directly as a class.
13077 (pretty_print_info): Remove declaration as class.
13078 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
13079 and pp_clear_output_area.
13080 (asan_add_global): Likewise.
13082 2013-08-22 Jan Hubicka <jh@suse.cz>
13084 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
13085 * ipa-utils.h (update_type_inheritance_graph): Declare.
13086 (possible_polymorphic_call_target_p): Declare.
13087 (possible_polymorphic_call_target_p): New.
13088 * ipa-devirt.c: Update toplevel comments.
13089 (cached_polymorphic_call_targets): Move up.
13090 (odr_type_d): Move ID down.
13091 (polymorphic_type_binfo_p): Update comment.
13092 (odr_hasher::remove): Likewise;
13093 (get_odr_type): Set anonymous_namespace.
13094 (dump_odr_type): Dump it.
13095 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
13096 (maybe_record_node): Record node in cached_polymorphic_call_targets.
13097 (record_binfo): Add comment.
13098 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
13100 (devirt_node_removal_hook): Do not iCE when cache is freed.
13101 (possible_polymorphic_call_target_p): New predicate.
13102 (update_type_inheritance_graph): New function.
13104 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
13105 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
13106 Sergey Lega <sergey.s.lega@intel.com>
13107 Anna Tikhonova <anna.tikhonova@intel.com>
13108 Ilya Tocar <ilya.tocar@intel.com>
13109 Andrey Turetskiy <andrey.turetskiy@intel.com>
13110 Ilya Verbin <ilya.verbin@intel.com>
13111 Kirill Yukhin <kirill.yukhin@intel.com>
13112 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
13114 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
13115 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
13116 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
13117 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
13118 (OPTION_MASK_ISA_AVX2_UNSET): Update.
13119 (OPTION_MASK_ISA_AVX512F_UNSET): New.
13120 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
13121 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
13122 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
13123 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
13124 OPT_mavx512pf, OPT_mavx512er cases.
13125 * config/i386/constraints.md (v): New constraint.
13126 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
13127 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
13128 (bit_AVX512CD): New.
13129 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13130 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
13131 * config/i386/i386-c.c (ix86_target_macros_internal):
13132 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
13134 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
13135 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
13136 * config/i386/i386.c (regclass_map, dbx_register_map)
13137 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
13138 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
13139 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
13140 -mavx512pf options.
13141 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
13142 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
13143 -mavx512pf options. Fix formatting.
13144 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
13145 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
13146 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
13147 -mavx512cd, -mavx512pf options.
13148 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
13149 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
13150 (ix86_preferred_output_reload_class): Replace SSE_REGS with
13152 (ix86_hard_regno_mode_ok): Support 512-bit registers.
13153 (ix86_set_reg_reg_cost): Ditto.
13154 (x86_order_regs_for_local_alloc): Ditto.
13155 (MAX_VECT_LEN): Extend to 64-byte.
13156 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
13157 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
13158 (TARGET_AVX512ER, TARGET_AVX512CD): New.
13159 (BIGGEST_ALIGNMENT): Extend to 512-bits.
13160 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
13161 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
13162 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
13163 (SSE_REG_MODE_P): Support new modes.
13164 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
13165 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
13166 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
13167 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
13168 (REG_CLASS_CONTENTS): Add new registers.
13169 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
13170 (EXT_REX_SSE_REGNO_P): New.
13171 (HI_REGISTER_NAMES): Add new registers.
13172 * config/i386/i386.md: Define constants for new registers.
13173 (mode): Add new 512-bit modes.
13174 (prefix): Support evex prefix.
13175 (isa): Support avx512f, noavx512f, fma_avx512f.
13176 (ssemodesuffix): Add new 512-bit modes.
13178 (*movxi_internal_avx512f): Ditto.
13179 (*movdi_internal): Replace constraint "x" with the new constraint "v".
13181 (*movsi_internal): Likewise.
13182 (*movdf_internal): Likewise.
13183 (*movsf_internal): Likewise.
13184 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
13185 (<code><mode>3): Likewise.
13186 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
13188 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
13189 with the new constraint "v".
13190 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
13192 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
13193 with the new constraint "v".
13194 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
13195 (<sse2>_storedqu<avxsizesuffix>): Likewise.
13196 (*<plusminus_insn><mode>3): Likewise.
13197 (<sse>_vm<plusminus_insn><mode>3): Likewise.
13198 (*mul<mode>3): Likewise.
13199 (<sse>_vmmul<mode>3): Likewise.
13200 (<sse>_div<mode>3): Likewise.
13201 (<sse>_vmdiv<mode>3): Likewise.
13202 (<sse>_sqrt<mode>2): Likewise.
13203 (<sse>_vmsqrt<mode>2): Likewise.
13204 (*<code><mode>3_finite): Likewise.
13205 (*<code><mode>3) <smaxmin>: Likewise.
13206 (<sse>_vm<code><mode>3): Likewise.
13207 (*<code><mode>3) <any_logic>: Likewise.
13208 (*fma_fmadd_<mode>): Likewise.
13209 (*fma_fmsub_<mode>): Likewise.
13210 (*fma_fnmadd_<mode>): Likewise.
13211 (*fma_fnmsub_<mode>): Likewise.
13212 (*fma_fmaddsub_<mode>): Likewise.
13213 (*fma_fmsubadd_<mode>): Likewise.
13214 (*fmai_fmadd_<mode>): Likewise.
13215 (*fmai_fmsub_<mode>): Likewise.
13216 (*fmai_fnmadd_<mode>): Likewise.
13217 (*fmai_fnmsub_<mode>): Likewise.
13218 (sse_cvtsi2ss): Likewise.
13219 (sse_cvtsi2ssq): Likewise.
13220 (sse_cvtss2si): Likewise.
13221 (sse_cvtss2si_2): Likewise.
13222 (sse_cvtss2siq): Likewise.
13223 (sse_cvtss2siq_2): Likewise.
13224 (sse_cvttss2si): Likewise.
13225 (sse_cvtss2siq_2): Likewise.
13226 (float<sseintvecmodelower><mode>2): Likewise.
13227 (sse2_cvtsd2si_2): Likewise.
13228 (sse2_cvtsd2siq_2): Likewise.
13229 (*<plusminus_insn><mode>3): Likewise.
13230 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
13231 (*<sse4_1_avx2>_mul<mode>3): Likewise.
13232 (ashr<mode>3): Likewise.
13233 (<shift_insn><mode>3): Likewise.
13234 (avx2_<code><mode>3): Likewise.
13235 (*avx2_<code><mode>3): Likewise.
13236 (*andnot<mode>3): Likewise.
13237 (*<code><mode>3) <any_logic>: Likewise.
13238 (abs<mode>2): Likewise.
13239 (avx2_permvar<mode>): Likewise.
13240 (avx2_perm<mode>_1): Likewise.
13241 (*avx_vpermilp<mode>): Likewise.
13242 (avx_vpermilvar<mode>3): Likewise.
13243 (avx2_ashrv<mode>): Likewise.
13244 (avx2_<shift_insn>v<mode>): Likewise.
13245 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
13247 * doc/rtl.texi: Document XImode.
13249 2013-08-21 Jeff Law <law@redhat.com>
13251 * tree-flow.h (register_jump_thread): Pass vector of edges
13252 instead of each important edge.
13253 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
13254 thread path into a vector and pass that to register_jump_thread.
13255 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
13256 passed in edge vector to the current 3-edge form.
13259 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
13261 * dce.c (fini_dce): Call df_analyze again just in case
13262 delete_unmarked_insns removed anything.
13264 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
13266 * reload.h (struct reg_equivs): Rename to ..
13267 (struct reg_equivs_s): .. this.
13269 2013-08-20 Martin Liska <marxin.liska@gmail.com>
13271 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
13273 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13275 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
13277 2013-08-21 Jeff Law <law@redhat.com>
13279 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
13280 simplify assignments too. If the RHS collapses to a singleton
13281 range, then return the value for the range.
13283 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
13285 * config/i386/sse.md (V16): Rename to...
13287 (mov<mode>): Update iterator name.
13288 (*mov<mode>_internal): Ditto.
13289 (push<mode>1): Ditto.
13290 (movmisalign<mode>): Ditto.
13292 2013-08-20 Jan Hubicka <jh@suse.cz>
13295 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
13296 entry for direct edges.
13297 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
13299 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13301 Revert my last two changes, r201865 and r201864:
13304 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13306 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
13307 instances can own GC refs.
13309 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
13310 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
13311 (gcc::context::gt_pch_nx): Likewise.
13312 (gcc::context::gt_pch_nx): Likewise.
13313 * ggc.h (gt_ggc_mx <T>): New.
13314 (gt_pch_nx_with_op <T>): New.
13315 (gt_pch_nx <T>): New.
13316 * passes.c (opt_pass::gt_ggc_mx): New.
13317 (opt_pass::gt_pch_nx): New.
13318 (opt_pass::gt_pch_nx_with_op): New.
13319 (pass_manager::gt_ggc_mx): New.
13320 (pass_manager::gt_pch_nx): New.
13321 (pass_manager::gt_pch_nx_with_op): New.
13322 (pass_manager::operator new): Use
13323 ggc_internal_cleared_alloc_stat rather than xcalloc.
13324 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
13325 (pass_manager::gt_ggc_mx): New.
13326 (pass_manager::gt_pch_nx): New.
13327 (pass_manager::gt_pch_nx_with_op): New.
13328 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
13329 (opt_pass::operator new): New.
13330 (opt_pass::gt_ggc_mx): New.
13331 (opt_pass::gt_pch_nx): New.
13332 (opt_pass::gt_pch_nx_with_op): New.
13335 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13337 * Makefile.in (GTFILES): Add context.h.
13338 * context.c (gcc::context::operator new): New.
13339 (gcc::context::gt_ggc_mx): New.
13340 (gcc::context::gt_pch_nx): New.
13341 (gcc::context::gt_pch_nx): New.
13342 * context.h (gcc::context): Add GTY((user)) marking.
13343 (gcc::context::operator new): New.
13344 (gcc::context::gt_ggc_mx): New.
13345 (gcc::context::gt_pch_nx): New.
13346 (gcc::context::gt_pch_nx): New.
13347 (g): Add GTY marking.
13348 (gt_ggc_mx (gcc::context *)): New.
13349 (gt_pch_nx (gcc::context *)): New.
13350 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
13351 void *cookie)): New.
13352 * gengtype.c (open_base_files) <ifiles>: Add context.h.
13354 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
13356 * dce.c (fini_dce): Call df_analyze again just in case
13357 delete_unmarked_insns removed anything.
13359 2013-08-20 Teresa Johnson <tejohnson@google.com>
13361 PR rtl-optimizations/57451
13362 * final.c (reemit_insn_block_notes): Prevent lexical blocks
13363 from crossing split section boundaries.
13365 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
13367 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
13368 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
13369 with MULTLIB_DEFAULTS.
13371 2013-08-20 Nick Clifton <nickc@redhat.com>
13373 * target.def (narrow_volatile_bitfield): Note that the default
13374 value is false, not !TARGET_STRICT_ALIGN.
13375 * doc/tm.texi: Regenerate.
13377 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
13379 Fix LIB_SPEC for systems without libpthread.
13381 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
13382 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
13384 * config/i386/linux-common.h: Likewise.
13385 * config/mips/linux-common.h: Likewise.
13387 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
13389 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
13392 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13394 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
13395 instances can own GC refs.
13397 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
13398 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
13399 (gcc::context::gt_pch_nx): Likewise.
13400 (gcc::context::gt_pch_nx): Likewise.
13401 * ggc.h (gt_ggc_mx <T>): New.
13402 (gt_pch_nx_with_op <T>): New.
13403 (gt_pch_nx <T>): New.
13404 * passes.c (opt_pass::gt_ggc_mx): New.
13405 (opt_pass::gt_pch_nx): New.
13406 (opt_pass::gt_pch_nx_with_op): New.
13407 (pass_manager::gt_ggc_mx): New.
13408 (pass_manager::gt_pch_nx): New.
13409 (pass_manager::gt_pch_nx_with_op): New.
13410 (pass_manager::operator new): Use
13411 ggc_internal_cleared_alloc_stat rather than xcalloc.
13412 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
13413 (pass_manager::gt_ggc_mx): New.
13414 (pass_manager::gt_pch_nx): New.
13415 (pass_manager::gt_pch_nx_with_op): New.
13416 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
13417 (opt_pass::operator new): New.
13418 (opt_pass::gt_ggc_mx): New.
13419 (opt_pass::gt_pch_nx): New.
13420 (opt_pass::gt_pch_nx_with_op): New.
13422 2013-08-20 David Malcolm <dmalcolm@redhat.com>
13424 * Makefile.in (GTFILES): Add context.h.
13425 * context.c (gcc::context::operator new): New.
13426 (gcc::context::gt_ggc_mx): New.
13427 (gcc::context::gt_pch_nx): New.
13428 (gcc::context::gt_pch_nx): New.
13429 * context.h (gcc::context): Add GTY((user)) marking.
13430 (gcc::context::operator new): New.
13431 (gcc::context::gt_ggc_mx): New.
13432 (gcc::context::gt_pch_nx): New.
13433 (gcc::context::gt_pch_nx): New.
13434 (g): Add GTY marking.
13435 (gt_ggc_mx (gcc::context *)): New.
13436 (gt_pch_nx (gcc::context *)): New.
13437 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
13438 void *cookie)): New.
13439 * gengtype.c (open_base_files) <ifiles>: Add context.h.
13441 2013-08-20 Alan Modra <amodra@gmail.com>
13444 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
13445 (rs6000_emit_epilogue): Likewise.
13447 2013-08-19 Dehao Chen <dehao@google.com>
13449 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
13451 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
13452 Jakub Jelinek <jakub@redhat.com>
13454 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
13455 (BUILT_IN_FABSD64): Likewise.
13456 (BUILT_IN_FABSD128): Likewise.
13457 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
13458 (fold_builtin_1): Likewise.
13459 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
13460 destination and source operands.
13461 (*abstd2_fpr): Likewise.
13462 (*nabstd2_fpr): Likewise.
13464 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
13466 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
13469 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
13471 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
13474 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
13476 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
13477 * builtins.c (default_libc_has_function): New.
13478 (gnu_libc_has_function): Ditto.
13479 (no_c99_libc_has_function): Ditto.
13480 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
13481 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
13482 (fold_builtin_sincos): Likewise.
13483 (fold_builtin_cexp): Likewise.
13484 * builtins.def (DEF_C94_BUILTIN): Likewise.
13485 (DEF_C99_BUILTIN): Likewise.
13486 (DEF_C99_C90RES_BUILTIN): Likewise.
13487 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
13488 definitions to using this define.
13489 * config/darwin-protos.h (darwin_libc_has_function): New.
13490 * config/darwin.c (darwin_libc_has_function): Ditto.
13491 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
13492 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
13493 * config/darwin.h: Ditto.
13494 * config/elfos.h: Ditto.
13495 * config/freebsd.h: Ditto.
13496 * config/i386/cygming.h: Ditto.
13497 * config/i386/djgpp.h: Ditto.
13498 * config/i386/i386-interix.h: Ditto.
13499 * config/microblaze/microblaze.h: Ditto.
13500 * config/mmix/mmix.h: Ditto.
13501 * config/gnu-user.h: Ditto.
13502 * config/ia64/hpux.h: Ditto.
13503 * config/pa/pa-hpux.h: Ditto.
13504 * config/pdp11/pdp11.h: Ditto.
13505 * config/picochip/picochip.h: Ditto.
13506 * config/linux.h: Ditto.
13507 * config/netbsd.h: Ditto.
13508 * config/openbsd.h: Ditto.
13509 * config/rs6000/aix43.h: Ditto.
13510 * config/rs6000/aix51.h: Ditto.
13511 * config/rs6000/aix52.h: Ditto.
13512 * config/rs6000/aix53.h: Ditto.
13513 * config/rs6000/aix61.h: Ditto.
13514 * config/rs6000/darwin.h: Ditto.
13515 * config/rs6000/linux.h: Ditto.
13516 * config/rs6000/linux64.h: Ditto.
13517 * config/s390/tpf.h: Ditto.
13518 * config/sol2-10.h: Ditto.
13519 * config/sol2.h: Ditto.
13520 * config/vms/vms.h: Ditto.
13521 * config/vxworks.h: Ditto.
13522 * config/linux-android.c (linux_android_libc_has_function):
13523 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
13524 * config/linux-protos.h (linux_android_libc_has_function):
13526 * config/i386/i386.c (ix86_libc_has_function): New.
13527 * config/i386/i386-protos.h
13528 (ix86_libc_has_function): New declaration.
13529 * config/i386/i386.md
13530 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
13531 ("isinf<mode>2): Likewise.
13532 * convert.c (convert_to_integer): Using new target hook
13533 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
13534 TARGET_C99_FUNCTIONS.
13535 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
13536 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
13537 * coretypes.h (function_class): New enum for different
13538 classes of functions.
13539 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
13540 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
13541 (TARGET_HAS_SINCOS): Likewise.
13542 (TARGET_LIBC_HAS_FUNCTION): New.
13543 * doc/tm.texi: Regenerated.
13544 * targhooks.h (default_libc_has_function): New declaration.
13545 (no_c99_libc_has_function): Ditto.
13546 (gnu_libc_has_function): Ditto.
13547 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
13548 and TARGET_HAS_SINCOS.
13550 2013-08-18 Jan Hubicka <jh@suse.cz>
13552 * Makeifle-in (ipa-devirt.o): New.
13553 (GTFILES): Add ipa-utils.h and ipa-devirt.c
13554 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
13555 (analyze_functions): Look into possible targets of polymorphic call.
13556 * dumpfile.c (dump_files): Add type-inheritance dump.
13557 * dumpfile.h (TDI_inheritance): New.
13558 * ipa-devirt.c: New file.
13559 * ipa-utils.h (odr_type_d): Forward declare.
13560 (odr_type): New type.
13561 (build_type_inheritance_graph): Declare.
13562 (possible_polymorphic_call_targets): Declare and introduce inline
13563 variant when only edge is pased.
13564 (dump_possible_polymorphic_call_targets): Likewise.
13565 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
13566 * tree.c (type_in_anonymous_namespace_p): Break out from ...
13567 (types_same_for_odr): ... here.
13568 * tree.h (type_in_anonymous_namespace_p): Declare.
13570 2013-08-18 Jakub Jelinek <jakub@redhat.com>
13572 PR tree-optimization/58006
13573 * tree-parloops.c (take_address_of): Don't ICE if get_name
13575 (eliminate_local_variables_stmt): Remove clobber stmts.
13577 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
13579 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
13580 error message is issued for an alias to undefined symbol.
13582 2013-08-18 Jan Hubicka <jh@suse.cz>
13584 * cgraph.c (cgraph_create_indirect_edge): Discover
13585 polymorphic calls and record basic info into indirect_info.
13586 * gimple-fold.c (gimple_fold_call): When doing BINFO based
13587 devirtualization, ignore objc function calls.
13588 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
13589 call with no parm index info.
13590 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
13591 * tree.c (virtual_method_call_p): New function.
13592 * tree.h (virtual_method_call_p): Declare.
13594 2013-08-16 Jan Hubicka <jh@suse.cz>
13596 PR middle-end/58179
13597 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
13599 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
13601 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
13604 2013-08-16 David Malcolm <dmalcolm@redhat.com>
13606 * gengtype.c (type_for_name): Add special-case support for
13607 locating types within the "gcc::" namespace.
13608 (open_base_files): Emit a "using namespace gcc" directive.
13610 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
13613 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
13614 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
13616 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
13617 array instead of each individual operand as a separate argument.
13618 (emit_fusion_gpr_load): Likewise.
13619 (expand_fusion_gpr_load): Add new function declaration.
13621 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
13622 signature to have the operands passed as an array, instead of as
13623 separate arguments. Allow ZERO_EXTEND to be in the memory
13624 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
13625 depend on the register live/dead flags when peepholes are run.
13626 (expand_fusion_gpr_load): New function to be called from the
13627 peephole2 pass, to change the register that addis sets to be the
13629 (emit_fusion_gpr_load): Change the calling signature to have the
13630 operands passed as an array, instead of as separate arguments.
13631 Allow ZERO_EXTEND to be in the memory address, and also
13632 SIGN_EXTEND if -mpower8-fusion-sign.
13634 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
13635 unspec enumeration.
13636 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
13637 adjust the register addis loads up in the peephole2 pass. Do not
13638 depend on the register live/dead state when the peephole pass is done.
13640 2013-08-16 David Malcolm <dmalcolm@redhat.com>
13642 * gengtype.c (create_user_defined_type): Ensure that the kind
13643 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
13644 declaration is seen before the GTY((user)) marking.
13646 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
13649 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
13651 2013-08-16 Jan Hubicka <jh@suse.cz>
13653 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
13654 arugment expected_type.
13655 (gimple_fold_call): Use it.
13656 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
13657 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
13658 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
13659 (try_make_edge_direct_virtual_call): Likewise.
13660 * tree.c (obj_type_ref_class): New.
13661 * tree.h (obj_type_ref_class): Use it.
13663 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
13665 * sched-vis.c (rtl_slim_pp_initialized): Remove.
13666 (rtl_slim_pp): Likewise.
13667 (init_rtl_slim_pretty_print): Likewise.
13668 (dump_value_slim): Don't call it. Use local pretty printer.
13669 (dump_insn_slim): Likewise.
13670 (dump_rtl_slim): Likewise.
13671 (str_pattern_slim): Likewise.
13672 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
13675 2013-08-16 Jakub Jelinek <jakub@redhat.com>
13677 PR tree-optimization/58164
13678 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
13679 walk gimple_goto_dest of GIMPLE_GOTO.
13681 PR tree-optimization/58165
13682 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
13683 bi_call must be the last stmt in a bb, don't split_block, instead
13684 use fallthru edge from it and give up if there is none.
13685 Release conds vector when returning early.
13687 2013-08-14 Xinliang David Li <davidxl@google.com>
13689 * config/i386/i386.c (ix86_option_override_internal):
13690 Remove unused variable and field.
13692 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13695 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
13696 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
13697 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
13698 and Linux, correct BLKmode alignment when 128-bit alignment is
13699 required and compatibility flag is not set.
13700 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
13701 for zero-size arguments when compatibility flag is not set.
13703 2013-08-14 Jakub Jelinek <jakub@redhat.com>
13705 PR tree-optimization/58145
13706 * tree-sra.c (build_ref_for_offset): If prev_base has
13707 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
13709 2013-08-14 Xinliang David Li <davidxl@google.com>
13711 * config/i386/i386.c (ix86_option_override_internal):
13712 Fix uninitialized variable error.
13714 2013-08-14 Xinliang David Li <davidxl@google.com>
13716 * config/i386/i386.opt: Define two new options.
13717 * config/i386/x86-tune.def: Add arch selector field in macros.
13718 * config/i386/i386.h: Adjust macro definition.
13719 * config/i386/i386.c (ix86_option_override_internal):
13721 (parse_mtune_ctrl_str): New function.
13722 (set_ix86_tune_features): New function.
13723 (ix86_function_specific_restore): Call the new helper function.
13725 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
13727 PR rtl-optimization/57662
13728 * sel-sched.c (code_motion_process_successors): When the current insn
13729 is removed after the recursive traversal, break from the loop.
13730 Add comments and debug printouts.
13732 2013-08-14 Jakub Jelinek <jakub@redhat.com>
13733 Alexandre Oliva <aoliva@redhat.com>
13736 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
13737 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
13738 there also UNSPEC_PLTOFF.
13740 2013-08-14 Marek Polacek <polacek@redhat.com>
13742 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
13743 NUM_CONDITIONS bit positions.
13745 2013-08-13 Cary Coutant <ccoutant@google.com>
13747 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
13748 (attr_checksum): Hash vector contents instead of pointer.
13749 (attr_checksum_ordered): Likewise.
13751 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
13753 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
13754 when Pmode != word_mode. Add length_address attribute.
13755 (sse3_monitor_<mode>): Merge from sse3_monitor and
13756 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
13757 Pmode != word_mode. Update insn length attribute.
13758 * config/i386/i386.c (ix86_option_override_internal): Update
13759 ix86_gen_monitor selection for merged sse3_monitor insn.
13761 2013-08-13 Julian Brown <julian@codesourcery.com>
13763 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
13764 perform invalid legitimization on greater-than-word-size modes for
13765 TARGET_E500_DOUBLE.
13767 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
13769 * ira.c (setup_class_translate_array): Use aclass instead of cl
13770 for classes not fully covered by allocno classes.
13772 2013-08-13 Jakub Jelinek <jakub@redhat.com>
13774 PR tree-optimization/57661
13775 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
13776 * tree-inline.c (tree_function_versioning): Initialize it.
13777 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
13778 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
13779 that is not being copied.
13782 * asan.c (instrument_strlen_call): Fix typo in comment.
13783 Use char * type even for the lhs of POINTER_PLUS_EXPR.
13785 2013-08-13 Steve Ellcey <sellcey@mips.com>
13787 * config/mips/mips.md (prefetch): Use lw instead of ld on
13788 loongson in 32bit mode.
13790 2013-08-13 Nick Clifton <nickc@redhat.com>
13792 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
13794 2013-08-13 Jan Hubicka <jh@suse.cz>
13796 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
13797 introduced edge; fix typo in sanity check.
13798 (cgraph_resolve_speculation): Export; improve diagnostic.
13799 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
13800 speculation at type mismatch.
13801 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
13802 (cgraph_resolve_speculation): Declare.
13803 (symtab_can_be_discarded): New function.
13804 * value-prof.c (gimple_ic_transform): Remove actual transform code.
13805 * ipa-inline-transform.c (speculation_removed): New global var.
13806 (clone_inlined_nodes): See if speculation can be removed.
13807 (inline_call): If speculations was removed, we growths may not match.
13808 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
13809 (speculation_useful_p): New function.
13810 (resolve_noninline_speculation): New function.
13811 (inline_small_functions): Resolve useless speculations.
13812 * ipa-inline.h (speculation_useful_p): Declare
13813 * ipa.c (can_replace_by_local_alias): Simplify.
13814 (ipa_profile): Produce speculative calls in non-lto, too;
13815 add simple cost model; produce local aliases.
13817 2013-08-13 David Malcolm <dmalcolm@redhat.com>
13819 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
13822 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
13824 * config/i386/i386.c (ix86_function_versions): Use error + inform.
13826 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
13828 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
13829 iterator instead of X87MODEF.
13831 2013-08-12 Perez Read <netfirewall@gmail.com>
13834 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
13835 operand 0 for intel asm alternative.
13836 (*movabs<mode>_2): Ditto for operand 1.
13838 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
13840 * config/aarch64/arm_none.h
13841 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
13843 2013-08-12 Nick Clifton <nickc@redhat.com>
13845 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
13847 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
13849 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
13850 expand for QI/HImode operand to produce more effictive code for
13851 unsigned char(short) --> float(double) conversion.
13853 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
13855 * doc/invoke.texi: Mention that -ftls-model does not force the final
13858 2013-08-12 Marek Polacek <polacek@redhat.com>
13859 Marc Glisse <marc.glisse@inria.fr>
13861 PR tree-optimization/57980
13862 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
13863 when creating -1 constant.
13865 2013-08-10 Jan Hubicka <jh@suse.cz>
13867 Workaround binutils PR14342.
13868 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
13869 (gimple_init_edge_profiler): Likewise.
13870 (gimple_gen_ic_func_profiler): Likewise.
13872 2013-08-09 Jan Hubicka <jh@suse.cz>
13874 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
13876 2013-08-09 Xinliang David Li <davidxl@google.com>
13878 * config/i386/stringop.def: New file.
13879 * config/i386/stringop.opt: New file.
13880 * config/i386/i386-opts.h: Include stringopt.def.
13881 * config/i386/i386.opt: Include stringopt.opt.
13882 * config/i386/i386.c (ix86_option_override_internal):
13883 Override default size based stringop inline strategies with options.
13884 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
13887 2013-08-09 Jan Hubicka <jh@suse.cz>
13889 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
13891 2013-08-09 Jan Hubicka <jh@suse.cz>
13893 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
13895 (dump_cgraph_node): Dump profile-id.
13896 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
13897 and common_target_probability.
13898 * lto-cgraph.c (lto_output_edge): Stream common targets.
13899 (lto_output_node): Stream profile ids.
13900 (input_node): Stream profile ids.
13901 (input_edge): Stream common targets.
13902 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
13903 * ipa.c: Include value-prof.h
13904 (ipa_profile_generate_summary): Turn indirect call statement histograms
13905 into common targets.
13906 (ipa_profile): Turn common targets into speculative edges.
13908 2013-08-09 Jan Hubicka <jh@suse.cz>
13910 * cgraph.h (cgraph_node): Add profile_id.
13911 * value-prof.c (cgraph_node_map): Turn into pointer_map.
13912 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
13913 (del_node_map): Update.
13914 (find_func_by_funcdef_no): Replace by ...
13915 (find_func_by_profile_id): ... this one.
13916 (gimple_ic_transform): Do not remove useful histograms when
13917 speculation is not done; dump info when indirect call removal
13919 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
13920 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
13921 (__gcov_indirect_call_profiler_v2): .. this one.
13922 * profile.h (init_node_map): Update.
13923 * coverage.c (coverage_compute_profile_id): New function.
13924 * coverage.h (coverage_compute_profile_id): Declare.
13925 * tree-profile.c (init_ic_make_global_vars): Make
13926 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
13927 (gimple_init_edge_profiler): Update prototype of
13928 __gcov_indirect_call_profiler.
13929 (gimple_gen_ic_func_profiler): Simplify.
13930 (tree_profiling): Use init_node_map
13932 2013-08-09 Jan Hubicka <jh@suse.cz>
13934 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
13935 non-speculative refs.
13936 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
13937 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
13938 (cgraph_set_call_stmt): Likewise.
13939 (cgraph_create_edge_1): Fix release checking compilatoin;
13940 clear lto_stmt_uid.
13941 (cgraph_free_edge): Free indirect info.
13942 (cgraph_turn_edge_to_speculative): New function.
13943 (cgraph_speculative_call_info): New function.
13944 (cgraph_make_edge_direct): Return direct edge; handle speculation.
13945 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
13946 (dump_cgraph_node): Dump speculation.
13947 (verify_edge_count_and_frequency): Accept speculative edges.
13948 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
13949 (verify_cgraph_node): Handle speculation.
13950 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
13951 (cgraph_set_call_stmt): Update prototype.
13952 (cgraph_make_edge_direct): Update prototype.
13953 (cgraph_speculative_call_info): Declare.
13954 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
13955 to change; update call of ipa_find_references.
13956 * ipa-ref.c (ipa_record_reference): Fix return value; clear
13957 lto_stmt_uid and speculative flags.
13958 (ipa_dump_references): Dump speculation.
13959 (ipa_clone_references): Clone speculative flag.
13960 (ipa_clone_referring): Likewise.
13961 (ipa_clone_ref): New function.
13962 (ipa_find_reference): Look into lto_stmt_uids
13963 (ipa_clear_stmts_in_references): Do not clear speculative calls.
13964 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
13965 (ipa_find_reference): Update declaration.
13966 (ipa_clone_ref): Declare.
13967 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
13968 stream speculative flag.
13969 (lto_output_ref): Stream statements uids and speculation.
13970 (input_ref): Likewise.
13971 (input_edge): Stream speuclation.
13972 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
13973 (cgraph_set_call_stmt_including_clones): Handle speculation.
13974 * ipa-inline.c (heap_edge_removal_hook): New function.
13975 (inline_small_functions): Register it.
13976 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
13977 also initialize refs.
13978 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
13980 (try_make_edge_direct_simple_call): Likewise.
13981 (try_make_edge_direct_simple_call): Likewise.
13982 (update_indirect_edges_after_inlining): Likewise.
13983 (remove_described_reference): Look proper lto_stmt_uid.
13984 (propagate_controlled_uses): Likewise.
13985 (propagate_controlled_uses): Liekwise.
13986 * tree-inline.c (copy_bb): Copy speculative edges.
13987 (redirect_all_calls): New function.
13988 (copy_cfg_body): Do redirection after loop info is updated.
13989 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
13991 2013-08-09 Jan Hubicka <jh@suse.cz>
13993 * lto-streamer-out.c (output_function): Renumber PHIs.
13994 * lto-streamer-in.c (input_function): Likewise.
13996 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
13998 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
13999 (get_lane_unsigned): Likewise.
14000 (dup_lane_scalar): Likewise.
14001 (get_lane): enable for VALL.
14002 * config/aarch64/aarch64-simd.md
14003 (aarch64_dup_lane_scalar<mode>): Remove.
14004 (aarch64_get_lane_signed<mode>): Likewise.
14005 (aarch64_get_lane_unsigned<mode>): Likewise.
14006 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
14007 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
14008 (aarch64_get_lane<mode>): Enable for all vector modes.
14009 (aarch64_get_lanedi): Remove misleading constraints.
14010 * config/aarch64/arm_neon.h
14011 (__aarch64_vget_lane_any): Define.
14012 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
14013 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
14014 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
14015 * config/aarch64/iterators.md (VDQQH): New.
14016 (VDQQHS): Likewise.
14017 (vwcore): Likewise.
14019 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
14021 * configure.ac: Add GAS check for LEON instructions on SPARC.
14022 * configure: Regenerate.
14023 * config.in: Likewise.
14024 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
14026 * config/sparc/sparc.opt (LEON, LEON3): New masks.
14027 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
14029 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
14030 (AS_LEON_FLAG): New macro.
14031 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
14032 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
14033 Deal with LEON and LEON3 for the memory model.
14034 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
14035 (atomic_compare_and_swap<mode>_1): Likewise.
14036 (*atomic_compare_and_swap<mode>_1): Likewise.
14038 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14040 * config/arm/neon.md (vcond): Fix floating-point vector
14041 comparisons against 0.
14043 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
14045 * lra-constraints.c (emit_spill_move): Remove assert.
14046 (process_alt_operands): Add more debugging
14047 output. Increase reject for spilling into memory. Decrease
14048 reject for reloading scratch.
14049 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
14051 2013-08-08 Steve Ellcey <sellcey@mips.com>
14053 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
14054 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
14055 micromips incompatible. Add nan2008.
14056 (MULTILIB_DIRNAMES): Add nan2008.
14057 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
14058 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
14059 and micromips incompatible. Add nan2008.
14060 (MULTILIB_DIRNAMES): Add nan2008.
14061 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
14063 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
14065 PR rtl-optimization/58079
14066 * combine.c (combine_simplify_rtx): Avoid using SUBST if
14067 simplify_comparison has widened a comparison with an integer.
14069 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14071 * config/arm/neon.md (movmisalign<mode>): Disable when we
14072 don't allow unaligned accesses.
14073 (*movmisalign<mode>_neon_store): Likewise.
14074 (*movmisalign<mode>_neon_load): Likewise.
14075 (*movmisalign<mode>_neon_store): Likewise.
14076 (*movmisalign<mode>_neon_load): Likewise.
14078 2013-08-08 Jan Hubicka <jh@suse.cz>
14080 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
14081 (make_pass_rebuild_cgraph_edges): Also clear references.
14082 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
14083 * ipa-inline-transform.c (inline_transform): Remove all references
14085 * cgraphunit.c (expand_function): Remove all references after
14087 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
14088 (ipa_find_reference): Rewrite to iterator.
14089 (remove_stmt_references): Likewise.
14090 (ipa_clear_stmts_in_references): New function.
14091 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
14092 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
14094 * ipa-split.c (split_function): Remove references in split function.
14096 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
14099 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
14100 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
14102 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
14105 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
14106 suggested mode for the assignment isn't compatible with the
14107 registers required.
14109 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
14112 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
14114 2013-08-07 Xinliang David Li <davidxl@google.com>
14116 * config/i386/i386.opt: New option -mtune-ctrl=.
14117 * config/i386/x86-tune.def: New file.
14118 * config/i386/i386.h: include x86-tune.def.
14119 * config/i386/i386.c (ix86_option_override_internal):
14120 Parsing -mtune-ctrl= option and set tune features.
14122 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
14125 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
14126 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
14129 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
14131 * rtl.h (update_alignments): Declare.
14132 * final.c (grow_label_align): New function extracted from...
14133 (shorten_branches): ...here. Call it.
14134 (update_alignments): New function.
14135 * reorg.c (sibling_labels): New variable.
14136 (get_label_before): Add SIBLING parameter. If it is non-zero, push
14137 the new label along with it onto the sibling_labels vector.
14138 (fill_simple_delay_slots): Adjust call to get_label_before.
14139 (fill_slots_from_thread): Likewise.
14140 (relax_delay_slots): Likewise.
14141 (make_return_insns): Likewise.
14142 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
14144 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
14146 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
14147 document its semantics.
14148 (diagnostic_report_diagnostic): Adjust accordingly.
14150 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14152 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
14153 (sparc_option_override): ...and port to new C++ pass API.
14154 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
14156 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
14158 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
14160 2013-08-06 Caroline Tice <cmtice@google.com>
14162 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
14163 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
14164 * tree-pass.h: Add pass_vtable_verify.
14165 * varasm.c (assemble_variable): Add code to properly set the comdat
14166 section and name for the .vtable_map_vars section.
14167 (assemble_vtyv_preinit_initializer): New function.
14168 (default_sectin_type_flags): Make sure .vtable_map_vars section has
14170 * output.h: Add function decl for assemble_vtv_preinit_initializer.
14171 * vtable-verify.c: New file.
14172 * vtable-verify.h: New file.
14173 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
14174 initialiation levels.
14175 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
14176 * passes.def: Insert pass_vtable_verify.
14177 * aclocal.m4: Reorder includes.
14178 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
14179 -fvtv-counts options.
14180 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
14181 as appropriate, if -fvtable-verify=... is used.
14182 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
14183 -fvtable-verify=... is used.
14184 * Makefile.in (OBJS): Add vtable-verify.o to list.
14185 (vtable-verify.o): Add new build rule.
14186 (GTFILES): Add vtable-verify.c to list.
14187 * common.opt (fvtable-verify=): New flag.
14188 (vtv_priority): Values for fvtable-verify= flag.
14189 (fvtv-counts): New flag.
14190 (fvtv-debug): New flag.
14191 * tree.h (save_vtable_map_decl): New extern function decl.
14193 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14195 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
14196 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
14197 (pass_data_rl78_devirt): ...new pass_data instance and...
14198 (make_pass_rl78_devirt): ...new function.
14199 (rl78_asm_file_start): Port pass registration to new C++ API.
14201 2013-08-07 David Malcolm <dmalcolm@redhat.com>
14203 * coretypes.h (rtl_opt_pass): Add.
14204 (gcc::context): Add.
14205 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
14206 (epiphany_init): Port to new C++ pass API.
14207 (epiphany_optimize_mode_switching): Likewise.
14208 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
14209 (pass_manager::get_pass_mode_switching): New.
14210 (pass_manager::get_pass_peephole2): New.
14211 * mode-switching.c (pass_mode_switching): Add clone method.
14212 * recog.c (pass_peephole2): Add clone method.
14213 (pass_split_all_insns): Add clone method.
14215 2013-08-06 David Malcolm <dmalcolm@redhat.com>
14217 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
14218 (mips_option_override): ...here, porting to new C++ API for passes.
14220 2013-08-06 Jan Hubicka <jh@suse.cz>
14222 * cgraph.c (cgraph_get_body): New function based on lto.c
14224 * cgraph.h (cgraph_get_body): Declare.
14225 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
14227 * cgraphunit.c (expand_function): Get body prior expanding.
14228 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
14229 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
14231 * passes.c (do_per_function_toporder): Get body.
14232 * tree-inline.c (expand_call_inline): Get body prior inlining it.
14233 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
14235 2013-08-06 Martin Jambor <mjambor@suse.cz>
14238 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
14239 is not re-finalized. Rename second parameter to no_collect.
14241 2013-08-06 Martin Jambor <mjambor@suse.cz>
14243 PR middle-end/58041
14244 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
14245 MEM_REF has proper alignment information.
14247 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
14250 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
14252 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
14253 rtx (*) (rtx, ...) with insn_gen_fn.
14254 * genoutput.c (output_insn_data): Cast gen_? function pointers to
14255 insn_gen_fn::stored_funcptr. Add initializer braces.
14257 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14259 Rewrite how instances of passes are cloned to remove assumptions
14260 about their sizes (thus allowing pass subclasses to have
14261 additional data fields, albeit non-GC-managed ones at this point).
14263 * passes.c (make_pass_instance): Now that passes have clone
14264 methods, rewrite this function to eliminate XNEW and memcpy
14265 calls that used hardcoded sizes. Since this function no longer
14266 creates pass instances, rename it to...
14267 (add_pass_instance): ...this. Document the old way that passes were
14268 numbered and flagged, and rework this function to continue using it.
14269 (next_pass_1): Add an initial_pass argument for use by
14271 (position_pass): When adding multiple instances of a pass, use
14272 the pass's clone method, rather than relying on the XNEW/memcpy
14273 within the former make_pass_instance (now add_pass_instance).
14274 (pass_manager::pass_manager): When invoking next_pass_1, also supply
14275 the initial instance of the current pass within the pass manager.
14277 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14279 This is the automated part of the conversion of passes from C
14280 structs to C++ classes.
14282 Patch autogenerated by refactor_passes.py from
14283 https://github.com/davidmalcolm/gcc-refactoring-scripts
14284 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
14286 * asan.c (pass_asan): Convert from a global struct to a subclass of
14287 gimple_opt_pass along with...
14288 (pass_data_asan): ...new pass_data instance and...
14289 (make_pass_asan): ...new function.
14290 (pass_asan_O0): Convert from a global struct to a subclass of
14291 gimple_opt_pass along with...
14292 (pass_data_asan_O0): ...new pass_data instance and...
14293 (make_pass_asan_O0): ...new function.
14294 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
14295 subclass of rtl_opt_pass along with...
14296 (pass_data_inc_dec): ...new pass_data instance and...
14297 (make_pass_inc_dec): ...new function.
14298 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
14299 a subclass of rtl_opt_pass along with...
14300 (pass_data_reorder_blocks): ...new pass_data instance and...
14301 (make_pass_reorder_blocks): ...new function.
14302 (pass_duplicate_computed_gotos): Convert from a global struct to a
14303 subclass of rtl_opt_pass along with...
14304 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
14305 (make_pass_duplicate_computed_gotos): ...new function.
14306 (pass_partition_blocks): Convert from a global struct to a subclass of
14307 rtl_opt_pass along with...
14308 (pass_data_partition_blocks): ...new pass_data instance and...
14309 (make_pass_partition_blocks): ...new function.
14310 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
14311 struct to a subclass of rtl_opt_pass along with...
14312 (pass_data_branch_target_load_optimize1): ...new pass_data instance
14314 (make_pass_branch_target_load_optimize1): ...new function.
14315 (pass_branch_target_load_optimize2): Convert from a global struct to a
14316 subclass of rtl_opt_pass along with...
14317 (pass_data_branch_target_load_optimize2): ...new pass_data instance
14319 (make_pass_branch_target_load_optimize2): ...new function.
14320 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
14321 of rtl_opt_pass along with...
14322 (pass_data_jump): ...new pass_data instance and...
14323 (make_pass_jump): ...new function.
14324 (pass_jump2): Convert from a global struct to a subclass of
14325 rtl_opt_pass along with...
14326 (pass_data_jump2): ...new pass_data instance and...
14327 (make_pass_jump2): ...new function.
14328 * cfgexpand.c (pass_expand): Convert from a global struct to a
14329 subclass of rtl_opt_pass along with...
14330 (pass_data_expand): ...new pass_data instance and...
14331 (make_pass_expand): ...new function.
14332 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
14333 of rtl_opt_pass along with...
14334 (pass_data_free_cfg): ...new pass_data instance and...
14335 (make_pass_free_cfg): ...new function.
14336 (pass_into_cfg_layout_mode): Convert from a global struct to a
14337 subclass of rtl_opt_pass along with...
14338 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
14339 (make_pass_into_cfg_layout_mode): ...new function.
14340 (pass_outof_cfg_layout_mode): Convert from a global struct to a
14341 subclass of rtl_opt_pass along with...
14342 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
14343 (make_pass_outof_cfg_layout_mode): ...new function.
14344 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
14345 struct to a subclass of gimple_opt_pass along with...
14346 (pass_data_build_cgraph_edges): ...new pass_data instance and...
14347 (make_pass_build_cgraph_edges): ...new function.
14348 (pass_rebuild_cgraph_edges): Convert from a global struct to a
14349 subclass of gimple_opt_pass along with...
14350 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
14351 (make_pass_rebuild_cgraph_edges): ...new function.
14352 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
14353 subclass of gimple_opt_pass along with...
14354 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
14356 (make_pass_remove_cgraph_callee_edges): ...new function.
14357 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
14358 struct to a subclass of rtl_opt_pass along with...
14359 (pass_data_stack_adjustments): ...new pass_data instance and...
14360 (make_pass_stack_adjustments): ...new function.
14361 * combine.c (pass_combine): Convert from a global struct to a subclass
14362 of rtl_opt_pass along with...
14363 (pass_data_combine): ...new pass_data instance and...
14364 (make_pass_combine): ...new function.
14365 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
14366 global struct to a subclass of rtl_opt_pass along with...
14367 (pass_data_compare_elim_after_reload): ...new pass_data instance
14369 (make_pass_compare_elim_after_reload): ...new function.
14370 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
14371 of rtl_opt_pass along with...
14372 (pass_data_rtl_cprop): ...new pass_data instance and...
14373 (make_pass_rtl_cprop): ...new function.
14374 * cse.c (pass_cse): Convert from a global struct to a subclass of
14375 rtl_opt_pass along with...
14376 (pass_data_cse): ...new pass_data instance and...
14377 (make_pass_cse): ...new function.
14378 (pass_cse2): Convert from a global struct to a subclass of
14379 rtl_opt_pass along with...
14380 (pass_data_cse2): ...new pass_data instance and...
14381 (make_pass_cse2): ...new function.
14382 (pass_cse_after_global_opts): Convert from a global struct to a
14383 subclass of rtl_opt_pass along with...
14384 (pass_data_cse_after_global_opts): ...new pass_data instance and...
14385 (make_pass_cse_after_global_opts): ...new function.
14386 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
14387 of rtl_opt_pass along with...
14388 (pass_data_ud_rtl_dce): ...new pass_data instance and...
14389 (make_pass_ud_rtl_dce): ...new function.
14390 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
14391 rtl_opt_pass along with...
14392 (pass_data_fast_rtl_dce): ...new pass_data instance and...
14393 (make_pass_fast_rtl_dce): ...new function.
14394 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
14395 a subclass of rtl_opt_pass along with...
14396 (pass_data_df_initialize_opt): ...new pass_data instance and...
14397 (make_pass_df_initialize_opt): ...new function.
14398 (pass_df_initialize_no_opt): Convert from a global struct to a
14399 subclass of rtl_opt_pass along with...
14400 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
14401 (make_pass_df_initialize_no_opt): ...new function.
14402 (pass_df_finish): Convert from a global struct to a subclass of
14403 rtl_opt_pass along with...
14404 (pass_data_df_finish): ...new pass_data instance and...
14405 (make_pass_df_finish): ...new function.
14406 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
14407 rtl_opt_pass along with...
14408 (pass_data_rtl_dse1): ...new pass_data instance and...
14409 (make_pass_rtl_dse1): ...new function.
14410 (pass_rtl_dse2): Convert from a global struct to a subclass of
14411 rtl_opt_pass along with...
14412 (pass_data_rtl_dse2): ...new pass_data instance and...
14413 (make_pass_rtl_dse2): ...new function.
14414 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
14415 subclass of rtl_opt_pass along with...
14416 (pass_data_dwarf2_frame): ...new pass_data instance and...
14417 (make_pass_dwarf2_frame): ...new function.
14418 * except.c (pass_set_nothrow_function_flags): Convert from a global
14419 struct to a subclass of rtl_opt_pass along with...
14420 (pass_data_set_nothrow_function_flags): ...new pass_data instance
14422 (make_pass_set_nothrow_function_flags): ...new function.
14423 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
14424 subclass of rtl_opt_pass along with...
14425 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
14427 (make_pass_convert_to_eh_region_ranges): ...new function.
14428 * final.c (pass_compute_alignments): Convert from a global struct to a
14429 subclass of rtl_opt_pass along with...
14430 (pass_data_compute_alignments): ...new pass_data instance and...
14431 (make_pass_compute_alignments): ...new function.
14432 (pass_final): Convert from a global struct to a subclass of
14433 rtl_opt_pass along with...
14434 (pass_data_final): ...new pass_data instance and...
14435 (make_pass_final): ...new function.
14436 (pass_shorten_branches): Convert from a global struct to a subclass of
14437 rtl_opt_pass along with...
14438 (pass_data_shorten_branches): ...new pass_data instance and...
14439 (make_pass_shorten_branches): ...new function.
14440 (pass_clean_state): Convert from a global struct to a subclass of
14441 rtl_opt_pass along with...
14442 (pass_data_clean_state): ...new pass_data instance and...
14443 (make_pass_clean_state): ...new function.
14444 * function.c (pass_instantiate_virtual_regs): Convert from a global
14445 struct to a subclass of rtl_opt_pass along with...
14446 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
14447 (make_pass_instantiate_virtual_regs): ...new function.
14448 (pass_leaf_regs): Convert from a global struct to a subclass of
14449 rtl_opt_pass along with...
14450 (pass_data_leaf_regs): ...new pass_data instance and...
14451 (make_pass_leaf_regs): ...new function.
14452 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
14453 subclass of rtl_opt_pass along with...
14454 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
14456 (make_pass_thread_prologue_and_epilogue): ...new function.
14457 (pass_match_asm_constraints): Convert from a global struct to a
14458 subclass of rtl_opt_pass along with...
14459 (pass_data_match_asm_constraints): ...new pass_data instance and...
14460 (make_pass_match_asm_constraints): ...new function.
14461 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
14462 subclass of rtl_opt_pass along with...
14463 (pass_data_rtl_fwprop): ...new pass_data instance and...
14464 (make_pass_rtl_fwprop): ...new function.
14465 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
14466 rtl_opt_pass along with...
14467 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
14468 (make_pass_rtl_fwprop_addr): ...new function.
14469 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
14470 rtl_opt_pass along with...
14471 (pass_data_rtl_pre): ...new pass_data instance and...
14472 (make_pass_rtl_pre): ...new function.
14473 (pass_rtl_hoist): Convert from a global struct to a subclass of
14474 rtl_opt_pass along with...
14475 (pass_data_rtl_hoist): ...new pass_data instance and...
14476 (make_pass_rtl_hoist): ...new function.
14477 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
14478 subclass of gimple_opt_pass along with...
14479 (pass_data_lower_cf): ...new pass_data instance and...
14480 (make_pass_lower_cf): ...new function.
14481 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
14482 from a global struct to a subclass of gimple_opt_pass along with...
14483 (pass_data_strength_reduction): ...new pass_data instance and...
14484 (make_pass_strength_reduction): ...new function.
14485 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
14486 of rtl_opt_pass along with...
14487 (pass_data_rtl_ifcvt): ...new pass_data instance and...
14488 (make_pass_rtl_ifcvt): ...new function.
14489 (pass_if_after_combine): Convert from a global struct to a subclass of
14490 rtl_opt_pass along with...
14491 (pass_data_if_after_combine): ...new pass_data instance and...
14492 (make_pass_if_after_combine): ...new function.
14493 (pass_if_after_reload): Convert from a global struct to a subclass of
14494 rtl_opt_pass along with...
14495 (pass_data_if_after_reload): ...new pass_data instance and...
14496 (make_pass_if_after_reload): ...new function.
14497 * init-regs.c (pass_initialize_regs): Convert from a global struct to
14498 a subclass of rtl_opt_pass along with...
14499 (pass_data_initialize_regs): ...new pass_data instance and...
14500 (make_pass_initialize_regs): ...new function.
14501 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
14502 of ipa_opt_pass_d along with...
14503 (pass_data_ipa_cp): ...new pass_data instance and...
14504 (make_pass_ipa_cp): ...new function.
14505 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
14506 global struct to a subclass of gimple_opt_pass along with...
14507 (pass_data_inline_parameters): ...new pass_data instance and...
14508 (make_pass_inline_parameters): ...new function.
14509 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
14510 subclass of gimple_opt_pass along with...
14511 (pass_data_early_inline): ...new pass_data instance and...
14512 (make_pass_early_inline): ...new function.
14513 (pass_ipa_inline): Convert from a global struct to a subclass of
14514 ipa_opt_pass_d along with...
14515 (pass_data_ipa_inline): ...new pass_data instance and...
14516 (make_pass_ipa_inline): ...new function.
14517 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
14518 struct to a subclass of gimple_opt_pass along with...
14519 (pass_data_local_pure_const): ...new pass_data instance and...
14520 (make_pass_local_pure_const): ...new function.
14521 (pass_ipa_pure_const): Convert from a global struct to a subclass of
14522 ipa_opt_pass_d along with...
14523 (pass_data_ipa_pure_const): ...new pass_data instance and...
14524 (make_pass_ipa_pure_const): ...new function.
14525 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
14526 to a subclass of ipa_opt_pass_d along with...
14527 (pass_data_ipa_reference): ...new pass_data instance and...
14528 (make_pass_ipa_reference): ...new function.
14529 * ipa-split.c (pass_split_functions): Convert from a global struct to
14530 a subclass of gimple_opt_pass along with...
14531 (pass_data_split_functions): ...new pass_data instance and...
14532 (make_pass_split_functions): ...new function.
14533 (pass_feedback_split_functions): Convert from a global struct to a
14534 subclass of gimple_opt_pass along with...
14535 (pass_data_feedback_split_functions): ...new pass_data instance and...
14536 (make_pass_feedback_split_functions): ...new function.
14537 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
14538 global struct to a subclass of simple_ipa_opt_pass along with...
14539 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
14541 (make_pass_ipa_function_and_variable_visibility): ...new function.
14542 (pass_ipa_free_inline_summary): Convert from a global struct to a
14543 subclass of simple_ipa_opt_pass along with...
14544 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
14545 (make_pass_ipa_free_inline_summary): ...new function.
14546 (pass_ipa_whole_program_visibility): Convert from a global struct to a
14547 subclass of ipa_opt_pass_d along with...
14548 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
14550 (make_pass_ipa_whole_program_visibility): ...new function.
14551 (pass_ipa_profile): Convert from a global struct to a subclass of
14552 ipa_opt_pass_d along with...
14553 (pass_data_ipa_profile): ...new pass_data instance and...
14554 (make_pass_ipa_profile): ...new function.
14555 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
14556 ipa_opt_pass_d along with...
14557 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
14558 (make_pass_ipa_cdtor_merge): ...new function.
14559 * ira.c (pass_ira): Convert from a global struct to a subclass of
14560 rtl_opt_pass along with...
14561 (pass_data_ira): ...new pass_data instance and...
14562 (make_pass_ira): ...new function.
14563 (pass_reload): Convert from a global struct to a subclass of
14564 rtl_opt_pass along with...
14565 (pass_data_reload): ...new pass_data instance and...
14566 (make_pass_reload): ...new function.
14567 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
14568 subclass of rtl_opt_pass along with...
14569 (pass_data_cleanup_barriers): ...new pass_data instance and...
14570 (make_pass_cleanup_barriers): ...new function.
14571 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
14572 of rtl_opt_pass along with...
14573 (pass_data_loop2): ...new pass_data instance and...
14574 (make_pass_loop2): ...new function.
14575 (pass_rtl_loop_init): Convert from a global struct to a subclass of
14576 rtl_opt_pass along with...
14577 (pass_data_rtl_loop_init): ...new pass_data instance and...
14578 (make_pass_rtl_loop_init): ...new function.
14579 (pass_rtl_loop_done): Convert from a global struct to a subclass of
14580 rtl_opt_pass along with...
14581 (pass_data_rtl_loop_done): ...new pass_data instance and...
14582 (make_pass_rtl_loop_done): ...new function.
14583 (pass_rtl_move_loop_invariants): Convert from a global struct to a
14584 subclass of rtl_opt_pass along with...
14585 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
14586 (make_pass_rtl_move_loop_invariants): ...new function.
14587 (pass_rtl_unswitch): Convert from a global struct to a subclass of
14588 rtl_opt_pass along with...
14589 (pass_data_rtl_unswitch): ...new pass_data instance and...
14590 (make_pass_rtl_unswitch): ...new function.
14591 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
14592 subclass of rtl_opt_pass along with...
14593 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
14595 (make_pass_rtl_unroll_and_peel_loops): ...new function.
14596 (pass_rtl_doloop): Convert from a global struct to a subclass of
14597 rtl_opt_pass along with...
14598 (pass_data_rtl_doloop): ...new pass_data instance and...
14599 (make_pass_rtl_doloop): ...new function.
14600 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
14601 a subclass of rtl_opt_pass along with...
14602 (pass_data_lower_subreg): ...new pass_data instance and...
14603 (make_pass_lower_subreg): ...new function.
14604 (pass_lower_subreg2): Convert from a global struct to a subclass of
14605 rtl_opt_pass along with...
14606 (pass_data_lower_subreg2): ...new pass_data instance and...
14607 (make_pass_lower_subreg2): ...new function.
14608 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
14609 struct to a subclass of ipa_opt_pass_d along with...
14610 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
14611 (make_pass_ipa_lto_gimple_out): ...new function.
14612 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
14613 of ipa_opt_pass_d along with...
14614 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
14615 (make_pass_ipa_lto_finish_out): ...new function.
14616 * mode-switching.c (pass_mode_switching): Convert from a global struct
14617 to a subclass of rtl_opt_pass along with...
14618 (pass_data_mode_switching): ...new pass_data instance and...
14619 (make_pass_mode_switching): ...new function.
14620 * modulo-sched.c (pass_sms): Convert from a global struct to a
14621 subclass of rtl_opt_pass along with...
14622 (pass_data_sms): ...new pass_data instance and...
14623 (make_pass_sms): ...new function.
14624 * omp-low.c (pass_expand_omp): Convert from a global struct to a
14625 subclass of gimple_opt_pass along with...
14626 (pass_data_expand_omp): ...new pass_data instance and...
14627 (make_pass_expand_omp): ...new function.
14628 (pass_lower_omp): Convert from a global struct to a subclass of
14629 gimple_opt_pass along with...
14630 (pass_data_lower_omp): ...new pass_data instance and...
14631 (make_pass_lower_omp): ...new function.
14632 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
14633 of gimple_opt_pass along with...
14634 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
14635 (make_pass_diagnose_omp_blocks): ...new function.
14636 * passes.c (pass_early_local_passes): Convert from a global struct to
14637 a subclass of simple_ipa_opt_pass along with...
14638 (pass_data_early_local_passes): ...new pass_data instance and...
14639 (make_pass_early_local_passes): ...new function.
14640 (pass_all_early_optimizations): Convert from a global struct to a
14641 subclass of gimple_opt_pass along with...
14642 (pass_data_all_early_optimizations): ...new pass_data instance and...
14643 (make_pass_all_early_optimizations): ...new function.
14644 (pass_all_optimizations): Convert from a global struct to a subclass
14645 of gimple_opt_pass along with...
14646 (pass_data_all_optimizations): ...new pass_data instance and...
14647 (make_pass_all_optimizations): ...new function.
14648 (pass_all_optimizations_g): Convert from a global struct to a subclass
14649 of gimple_opt_pass along with...
14650 (pass_data_all_optimizations_g): ...new pass_data instance and...
14651 (make_pass_all_optimizations_g): ...new function.
14652 (pass_rest_of_compilation): Convert from a global struct to a subclass
14653 of rtl_opt_pass along with...
14654 (pass_data_rest_of_compilation): ...new pass_data instance and...
14655 (make_pass_rest_of_compilation): ...new function.
14656 (pass_postreload): Convert from a global struct to a subclass of
14657 rtl_opt_pass along with...
14658 (pass_data_postreload): ...new pass_data instance and...
14659 (make_pass_postreload): ...new function.
14660 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
14661 subclass of rtl_opt_pass along with...
14662 (pass_data_gcse2): ...new pass_data instance and...
14663 (make_pass_gcse2): ...new function.
14664 * postreload.c (pass_postreload_cse): Convert from a global struct to
14665 a subclass of rtl_opt_pass along with...
14666 (pass_data_postreload_cse): ...new pass_data instance and...
14667 (make_pass_postreload_cse): ...new function.
14668 * predict.c (pass_profile): Convert from a global struct to a subclass
14669 of gimple_opt_pass along with...
14670 (pass_data_profile): ...new pass_data instance and...
14671 (make_pass_profile): ...new function.
14672 (pass_strip_predict_hints): Convert from a global struct to a subclass
14673 of gimple_opt_pass along with...
14674 (pass_data_strip_predict_hints): ...new pass_data instance and...
14675 (make_pass_strip_predict_hints): ...new function.
14676 * recog.c (pass_peephole2): Convert from a global struct to a subclass
14677 of rtl_opt_pass along with...
14678 (pass_data_peephole2): ...new pass_data instance and...
14679 (make_pass_peephole2): ...new function.
14680 (pass_split_all_insns): Convert from a global struct to a subclass of
14681 rtl_opt_pass along with...
14682 (pass_data_split_all_insns): ...new pass_data instance and...
14683 (make_pass_split_all_insns): ...new function.
14684 (pass_split_after_reload): Convert from a global struct to a subclass
14685 of rtl_opt_pass along with...
14686 (pass_data_split_after_reload): ...new pass_data instance and...
14687 (make_pass_split_after_reload): ...new function.
14688 (pass_split_before_regstack): Convert from a global struct to a
14689 subclass of rtl_opt_pass along with...
14690 (pass_data_split_before_regstack): ...new pass_data instance and...
14691 (make_pass_split_before_regstack): ...new function.
14692 (pass_split_before_sched2): Convert from a global struct to a subclass
14693 of rtl_opt_pass along with...
14694 (pass_data_split_before_sched2): ...new pass_data instance and...
14695 (make_pass_split_before_sched2): ...new function.
14696 (pass_split_for_shorten_branches): Convert from a global struct to a
14697 subclass of rtl_opt_pass along with...
14698 (pass_data_split_for_shorten_branches): ...new pass_data instance
14700 (make_pass_split_for_shorten_branches): ...new function.
14701 * ree.c (pass_ree): Convert from a global struct to a subclass of
14702 rtl_opt_pass along with...
14703 (pass_data_ree): ...new pass_data instance and...
14704 (make_pass_ree): ...new function.
14705 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
14706 subclass of rtl_opt_pass along with...
14707 (pass_data_stack_regs): ...new pass_data instance and...
14708 (make_pass_stack_regs): ...new function.
14709 (pass_stack_regs_run): Convert from a global struct to a subclass of
14710 rtl_opt_pass along with...
14711 (pass_data_stack_regs_run): ...new pass_data instance and...
14712 (make_pass_stack_regs_run): ...new function.
14713 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
14714 subclass of rtl_opt_pass along with...
14715 (pass_data_cprop_hardreg): ...new pass_data instance and...
14716 (make_pass_cprop_hardreg): ...new function.
14717 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
14718 subclass of rtl_opt_pass along with...
14719 (pass_data_reginfo_init): ...new pass_data instance and...
14720 (make_pass_reginfo_init): ...new function.
14721 * regmove.c (pass_regmove): Convert from a global struct to a subclass
14722 of rtl_opt_pass along with...
14723 (pass_data_regmove): ...new pass_data instance and...
14724 (make_pass_regmove): ...new function.
14725 * regrename.c (pass_regrename): Convert from a global struct to a
14726 subclass of rtl_opt_pass along with...
14727 (pass_data_regrename): ...new pass_data instance and...
14728 (make_pass_regrename): ...new function.
14729 * reorg.c (pass_delay_slots): Convert from a global struct to a
14730 subclass of rtl_opt_pass along with...
14731 (pass_data_delay_slots): ...new pass_data instance and...
14732 (make_pass_delay_slots): ...new function.
14733 (pass_machine_reorg): Convert from a global struct to a subclass of
14734 rtl_opt_pass along with...
14735 (pass_data_machine_reorg): ...new pass_data instance and...
14736 (make_pass_machine_reorg): ...new function.
14737 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
14738 of rtl_opt_pass along with...
14739 (pass_data_sched): ...new pass_data instance and...
14740 (make_pass_sched): ...new function.
14741 (pass_sched2): Convert from a global struct to a subclass of
14742 rtl_opt_pass along with...
14743 (pass_data_sched2): ...new pass_data instance and...
14744 (make_pass_sched2): ...new function.
14745 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
14746 to a subclass of rtl_opt_pass along with...
14747 (pass_data_stack_ptr_mod): ...new pass_data instance and...
14748 (make_pass_stack_ptr_mod): ...new function.
14749 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
14750 to a subclass of rtl_opt_pass along with...
14751 (pass_data_rtl_store_motion): ...new pass_data instance and...
14752 (make_pass_rtl_store_motion): ...new function.
14753 * tracer.c (pass_tracer): Convert from a global struct to a subclass
14754 of gimple_opt_pass along with...
14755 (pass_data_tracer): ...new pass_data instance and...
14756 (make_pass_tracer): ...new function.
14757 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
14758 to a subclass of gimple_opt_pass along with...
14759 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
14760 (make_pass_diagnose_tm_blocks): ...new function.
14761 (pass_lower_tm): Convert from a global struct to a subclass of
14762 gimple_opt_pass along with...
14763 (pass_data_lower_tm): ...new pass_data instance and...
14764 (make_pass_lower_tm): ...new function.
14765 (pass_tm_init): Convert from a global struct to a subclass of
14766 gimple_opt_pass along with...
14767 (pass_data_tm_init): ...new pass_data instance and...
14768 (make_pass_tm_init): ...new function.
14769 (pass_tm_mark): Convert from a global struct to a subclass of
14770 gimple_opt_pass along with...
14771 (pass_data_tm_mark): ...new pass_data instance and...
14772 (make_pass_tm_mark): ...new function.
14773 (pass_tm_edges): Convert from a global struct to a subclass of
14774 gimple_opt_pass along with...
14775 (pass_data_tm_edges): ...new pass_data instance and...
14776 (make_pass_tm_edges): ...new function.
14777 (pass_tm_memopt): Convert from a global struct to a subclass of
14778 gimple_opt_pass along with...
14779 (pass_data_tm_memopt): ...new pass_data instance and...
14780 (make_pass_tm_memopt): ...new function.
14781 (pass_ipa_tm): Convert from a global struct to a subclass of
14782 simple_ipa_opt_pass along with...
14783 (pass_data_ipa_tm): ...new pass_data instance and...
14784 (make_pass_ipa_tm): ...new function.
14785 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
14786 subclass of gimple_opt_pass along with...
14787 (pass_data_call_cdce): ...new pass_data instance and...
14788 (make_pass_call_cdce): ...new function.
14789 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
14790 subclass of gimple_opt_pass along with...
14791 (pass_data_build_cfg): ...new pass_data instance and...
14792 (make_pass_build_cfg): ...new function.
14793 (pass_split_crit_edges): Convert from a global struct to a subclass of
14794 gimple_opt_pass along with...
14795 (pass_data_split_crit_edges): ...new pass_data instance and...
14796 (make_pass_split_crit_edges): ...new function.
14797 (pass_warn_function_return): Convert from a global struct to a
14798 subclass of gimple_opt_pass along with...
14799 (pass_data_warn_function_return): ...new pass_data instance and...
14800 (make_pass_warn_function_return): ...new function.
14801 (pass_warn_function_noreturn): Convert from a global struct to a
14802 subclass of gimple_opt_pass along with...
14803 (pass_data_warn_function_noreturn): ...new pass_data instance and...
14804 (make_pass_warn_function_noreturn): ...new function.
14805 (pass_warn_unused_result): Convert from a global struct to a subclass
14806 of gimple_opt_pass along with...
14807 (pass_data_warn_unused_result): ...new pass_data instance and...
14808 (make_pass_warn_unused_result): ...new function.
14809 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
14810 a subclass of gimple_opt_pass along with...
14811 (pass_data_merge_phi): ...new pass_data instance and...
14812 (make_pass_merge_phi): ...new function.
14813 * tree-complex.c (pass_lower_complex): Convert from a global struct to
14814 a subclass of gimple_opt_pass along with...
14815 (pass_data_lower_complex): ...new pass_data instance and...
14816 (make_pass_lower_complex): ...new function.
14817 (pass_lower_complex_O0): Convert from a global struct to a subclass of
14818 gimple_opt_pass along with...
14819 (pass_data_lower_complex_O0): ...new pass_data instance and...
14820 (make_pass_lower_complex_O0): ...new function.
14821 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
14822 subclass of gimple_opt_pass along with...
14823 (pass_data_lower_eh): ...new pass_data instance and...
14824 (make_pass_lower_eh): ...new function.
14825 (pass_refactor_eh): Convert from a global struct to a subclass of
14826 gimple_opt_pass along with...
14827 (pass_data_refactor_eh): ...new pass_data instance and...
14828 (make_pass_refactor_eh): ...new function.
14829 (pass_lower_resx): Convert from a global struct to a subclass of
14830 gimple_opt_pass along with...
14831 (pass_data_lower_resx): ...new pass_data instance and...
14832 (make_pass_lower_resx): ...new function.
14833 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
14834 of gimple_opt_pass along with...
14835 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
14836 (make_pass_lower_eh_dispatch): ...new function.
14837 (pass_cleanup_eh): Convert from a global struct to a subclass of
14838 gimple_opt_pass along with...
14839 (pass_data_cleanup_eh): ...new pass_data instance and...
14840 (make_pass_cleanup_eh): ...new function.
14841 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
14842 to a subclass of simple_ipa_opt_pass along with...
14843 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
14844 (make_pass_ipa_lower_emutls): ...new function.
14845 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
14846 a subclass of gimple_opt_pass along with...
14847 (pass_data_if_conversion): ...new pass_data instance and...
14848 (make_pass_if_conversion): ...new function.
14849 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
14850 subclass of gimple_opt_pass along with...
14851 (pass_data_build_ssa): ...new pass_data instance and...
14852 (make_pass_build_ssa): ...new function.
14853 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
14854 global struct to a subclass of gimple_opt_pass along with...
14855 (pass_data_loop_distribution): ...new pass_data instance and...
14856 (make_pass_loop_distribution): ...new function.
14857 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
14858 subclass of gimple_opt_pass along with...
14859 (pass_data_mudflap_1): ...new pass_data instance and...
14860 (make_pass_mudflap_1): ...new function.
14861 (pass_mudflap_2): Convert from a global struct to a subclass of
14862 gimple_opt_pass along with...
14863 (pass_data_mudflap_2): ...new pass_data instance and...
14864 (make_pass_mudflap_2): ...new function.
14865 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
14866 subclass of gimple_opt_pass along with...
14867 (pass_data_mudflap_1): ...new pass_data instance and...
14868 (make_pass_mudflap_1): ...new function.
14869 (pass_mudflap_2): Convert from a global struct to a subclass of
14870 gimple_opt_pass along with...
14871 (pass_data_mudflap_2): ...new pass_data instance and...
14872 (make_pass_mudflap_2): ...new function.
14873 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
14874 gimple_opt_pass along with...
14875 (pass_data_nrv): ...new pass_data instance and...
14876 (make_pass_nrv): ...new function.
14877 (pass_return_slot): Convert from a global struct to a subclass of
14878 gimple_opt_pass along with...
14879 (pass_data_return_slot): ...new pass_data instance and...
14880 (make_pass_return_slot): ...new function.
14881 * tree-object-size.c (pass_object_sizes): Convert from a global struct
14882 to a subclass of gimple_opt_pass along with...
14883 (pass_data_object_sizes): ...new pass_data instance and...
14884 (make_pass_object_sizes): ...new function.
14885 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
14886 global struct to a subclass of gimple_opt_pass along with...
14887 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
14889 (make_pass_cleanup_cfg_post_optimizing): ...new function.
14890 (pass_fixup_cfg): Convert from a global struct to a subclass of
14891 gimple_opt_pass along with...
14892 (pass_data_fixup_cfg): ...new pass_data instance and...
14893 (make_pass_fixup_cfg): ...new function.
14894 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
14895 (make_pass_mudflap_1): ...new function.
14896 (pass_mudflap_2): Replace declaration with that of...
14897 (make_pass_mudflap_2): ...new function.
14898 (pass_asan): Replace declaration with that of...
14899 (make_pass_asan): ...new function.
14900 (pass_asan_O0): Replace declaration with that of...
14901 (make_pass_asan_O0): ...new function.
14902 (pass_tsan): Replace declaration with that of...
14903 (make_pass_tsan): ...new function.
14904 (pass_tsan_O0): Replace declaration with that of...
14905 (make_pass_tsan_O0): ...new function.
14906 (pass_lower_cf): Replace declaration with that of...
14907 (make_pass_lower_cf): ...new function.
14908 (pass_refactor_eh): Replace declaration with that of...
14909 (make_pass_refactor_eh): ...new function.
14910 (pass_lower_eh): Replace declaration with that of...
14911 (make_pass_lower_eh): ...new function.
14912 (pass_lower_eh_dispatch): Replace declaration with that of...
14913 (make_pass_lower_eh_dispatch): ...new function.
14914 (pass_lower_resx): Replace declaration with that of...
14915 (make_pass_lower_resx): ...new function.
14916 (pass_build_cfg): Replace declaration with that of...
14917 (make_pass_build_cfg): ...new function.
14918 (pass_early_tree_profile): Replace declaration with that of...
14919 (make_pass_early_tree_profile): ...new function.
14920 (pass_cleanup_eh): Replace declaration with that of...
14921 (make_pass_cleanup_eh): ...new function.
14922 (pass_sra): Replace declaration with that of...
14923 (make_pass_sra): ...new function.
14924 (pass_sra_early): Replace declaration with that of...
14925 (make_pass_sra_early): ...new function.
14926 (pass_early_ipa_sra): Replace declaration with that of...
14927 (make_pass_early_ipa_sra): ...new function.
14928 (pass_tail_recursion): Replace declaration with that of...
14929 (make_pass_tail_recursion): ...new function.
14930 (pass_tail_calls): Replace declaration with that of...
14931 (make_pass_tail_calls): ...new function.
14932 (pass_tree_loop): Replace declaration with that of...
14933 (make_pass_tree_loop): ...new function.
14934 (pass_tree_loop_init): Replace declaration with that of...
14935 (make_pass_tree_loop_init): ...new function.
14936 (pass_lim): Replace declaration with that of...
14937 (make_pass_lim): ...new function.
14938 (pass_tree_unswitch): Replace declaration with that of...
14939 (make_pass_tree_unswitch): ...new function.
14940 (pass_predcom): Replace declaration with that of...
14941 (make_pass_predcom): ...new function.
14942 (pass_iv_canon): Replace declaration with that of...
14943 (make_pass_iv_canon): ...new function.
14944 (pass_scev_cprop): Replace declaration with that of...
14945 (make_pass_scev_cprop): ...new function.
14946 (pass_empty_loop): Replace declaration with that of...
14947 (make_pass_empty_loop): ...new function.
14948 (pass_record_bounds): Replace declaration with that of...
14949 (make_pass_record_bounds): ...new function.
14950 (pass_graphite): Replace declaration with that of...
14951 (make_pass_graphite): ...new function.
14952 (pass_graphite_transforms): Replace declaration with that of...
14953 (make_pass_graphite_transforms): ...new function.
14954 (pass_if_conversion): Replace declaration with that of...
14955 (make_pass_if_conversion): ...new function.
14956 (pass_loop_distribution): Replace declaration with that of...
14957 (make_pass_loop_distribution): ...new function.
14958 (pass_vectorize): Replace declaration with that of...
14959 (make_pass_vectorize): ...new function.
14960 (pass_slp_vectorize): Replace declaration with that of...
14961 (make_pass_slp_vectorize): ...new function.
14962 (pass_complete_unroll): Replace declaration with that of...
14963 (make_pass_complete_unroll): ...new function.
14964 (pass_complete_unrolli): Replace declaration with that of...
14965 (make_pass_complete_unrolli): ...new function.
14966 (pass_parallelize_loops): Replace declaration with that of...
14967 (make_pass_parallelize_loops): ...new function.
14968 (pass_loop_prefetch): Replace declaration with that of...
14969 (make_pass_loop_prefetch): ...new function.
14970 (pass_iv_optimize): Replace declaration with that of...
14971 (make_pass_iv_optimize): ...new function.
14972 (pass_tree_loop_done): Replace declaration with that of...
14973 (make_pass_tree_loop_done): ...new function.
14974 (pass_ch): Replace declaration with that of...
14975 (make_pass_ch): ...new function.
14976 (pass_ccp): Replace declaration with that of...
14977 (make_pass_ccp): ...new function.
14978 (pass_phi_only_cprop): Replace declaration with that of...
14979 (make_pass_phi_only_cprop): ...new function.
14980 (pass_build_ssa): Replace declaration with that of...
14981 (make_pass_build_ssa): ...new function.
14982 (pass_build_alias): Replace declaration with that of...
14983 (make_pass_build_alias): ...new function.
14984 (pass_build_ealias): Replace declaration with that of...
14985 (make_pass_build_ealias): ...new function.
14986 (pass_dominator): Replace declaration with that of...
14987 (make_pass_dominator): ...new function.
14988 (pass_dce): Replace declaration with that of...
14989 (make_pass_dce): ...new function.
14990 (pass_dce_loop): Replace declaration with that of...
14991 (make_pass_dce_loop): ...new function.
14992 (pass_cd_dce): Replace declaration with that of...
14993 (make_pass_cd_dce): ...new function.
14994 (pass_call_cdce): Replace declaration with that of...
14995 (make_pass_call_cdce): ...new function.
14996 (pass_merge_phi): Replace declaration with that of...
14997 (make_pass_merge_phi): ...new function.
14998 (pass_split_crit_edges): Replace declaration with that of...
14999 (make_pass_split_crit_edges): ...new function.
15000 (pass_pre): Replace declaration with that of...
15001 (make_pass_pre): ...new function.
15002 (pass_profile): Replace declaration with that of...
15003 (make_pass_profile): ...new function.
15004 (pass_strip_predict_hints): Replace declaration with that of...
15005 (make_pass_strip_predict_hints): ...new function.
15006 (pass_lower_complex_O0): Replace declaration with that of...
15007 (make_pass_lower_complex_O0): ...new function.
15008 (pass_lower_complex): Replace declaration with that of...
15009 (make_pass_lower_complex): ...new function.
15010 (pass_lower_vector): Replace declaration with that of...
15011 (make_pass_lower_vector): ...new function.
15012 (pass_lower_vector_ssa): Replace declaration with that of...
15013 (make_pass_lower_vector_ssa): ...new function.
15014 (pass_lower_omp): Replace declaration with that of...
15015 (make_pass_lower_omp): ...new function.
15016 (pass_diagnose_omp_blocks): Replace declaration with that of...
15017 (make_pass_diagnose_omp_blocks): ...new function.
15018 (pass_expand_omp): Replace declaration with that of...
15019 (make_pass_expand_omp): ...new function.
15020 (pass_expand_omp_ssa): Replace declaration with that of...
15021 (make_pass_expand_omp_ssa): ...new function.
15022 (pass_object_sizes): Replace declaration with that of...
15023 (make_pass_object_sizes): ...new function.
15024 (pass_strlen): Replace declaration with that of...
15025 (make_pass_strlen): ...new function.
15026 (pass_fold_builtins): Replace declaration with that of...
15027 (make_pass_fold_builtins): ...new function.
15028 (pass_stdarg): Replace declaration with that of...
15029 (make_pass_stdarg): ...new function.
15030 (pass_early_warn_uninitialized): Replace declaration with that of...
15031 (make_pass_early_warn_uninitialized): ...new function.
15032 (pass_late_warn_uninitialized): Replace declaration with that of...
15033 (make_pass_late_warn_uninitialized): ...new function.
15034 (pass_cse_reciprocals): Replace declaration with that of...
15035 (make_pass_cse_reciprocals): ...new function.
15036 (pass_cse_sincos): Replace declaration with that of...
15037 (make_pass_cse_sincos): ...new function.
15038 (pass_optimize_bswap): Replace declaration with that of...
15039 (make_pass_optimize_bswap): ...new function.
15040 (pass_optimize_widening_mul): Replace declaration with that of...
15041 (make_pass_optimize_widening_mul): ...new function.
15042 (pass_warn_function_return): Replace declaration with that of...
15043 (make_pass_warn_function_return): ...new function.
15044 (pass_warn_function_noreturn): Replace declaration with that of...
15045 (make_pass_warn_function_noreturn): ...new function.
15046 (pass_cselim): Replace declaration with that of...
15047 (make_pass_cselim): ...new function.
15048 (pass_phiopt): Replace declaration with that of...
15049 (make_pass_phiopt): ...new function.
15050 (pass_forwprop): Replace declaration with that of...
15051 (make_pass_forwprop): ...new function.
15052 (pass_phiprop): Replace declaration with that of...
15053 (make_pass_phiprop): ...new function.
15054 (pass_tree_ifcombine): Replace declaration with that of...
15055 (make_pass_tree_ifcombine): ...new function.
15056 (pass_dse): Replace declaration with that of...
15057 (make_pass_dse): ...new function.
15058 (pass_nrv): Replace declaration with that of...
15059 (make_pass_nrv): ...new function.
15060 (pass_rename_ssa_copies): Replace declaration with that of...
15061 (make_pass_rename_ssa_copies): ...new function.
15062 (pass_sink_code): Replace declaration with that of...
15063 (make_pass_sink_code): ...new function.
15064 (pass_fre): Replace declaration with that of...
15065 (make_pass_fre): ...new function.
15066 (pass_check_data_deps): Replace declaration with that of...
15067 (make_pass_check_data_deps): ...new function.
15068 (pass_copy_prop): Replace declaration with that of...
15069 (make_pass_copy_prop): ...new function.
15070 (pass_vrp): Replace declaration with that of...
15071 (make_pass_vrp): ...new function.
15072 (pass_uncprop): Replace declaration with that of...
15073 (make_pass_uncprop): ...new function.
15074 (pass_return_slot): Replace declaration with that of...
15075 (make_pass_return_slot): ...new function.
15076 (pass_reassoc): Replace declaration with that of...
15077 (make_pass_reassoc): ...new function.
15078 (pass_rebuild_cgraph_edges): Replace declaration with that of...
15079 (make_pass_rebuild_cgraph_edges): ...new function.
15080 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
15081 (make_pass_remove_cgraph_callee_edges): ...new function.
15082 (pass_build_cgraph_edges): Replace declaration with that of...
15083 (make_pass_build_cgraph_edges): ...new function.
15084 (pass_local_pure_const): Replace declaration with that of...
15085 (make_pass_local_pure_const): ...new function.
15086 (pass_tracer): Replace declaration with that of...
15087 (make_pass_tracer): ...new function.
15088 (pass_warn_unused_result): Replace declaration with that of...
15089 (make_pass_warn_unused_result): ...new function.
15090 (pass_diagnose_tm_blocks): Replace declaration with that of...
15091 (make_pass_diagnose_tm_blocks): ...new function.
15092 (pass_lower_tm): Replace declaration with that of...
15093 (make_pass_lower_tm): ...new function.
15094 (pass_tm_init): Replace declaration with that of...
15095 (make_pass_tm_init): ...new function.
15096 (pass_tm_mark): Replace declaration with that of...
15097 (make_pass_tm_mark): ...new function.
15098 (pass_tm_memopt): Replace declaration with that of...
15099 (make_pass_tm_memopt): ...new function.
15100 (pass_tm_edges): Replace declaration with that of...
15101 (make_pass_tm_edges): ...new function.
15102 (pass_split_functions): Replace declaration with that of...
15103 (make_pass_split_functions): ...new function.
15104 (pass_feedback_split_functions): Replace declaration with that of...
15105 (make_pass_feedback_split_functions): ...new function.
15106 (pass_strength_reduction): Replace declaration with that of...
15107 (make_pass_strength_reduction): ...new function.
15108 (pass_ipa_lower_emutls): Replace declaration with that of...
15109 (make_pass_ipa_lower_emutls): ...new function.
15110 (pass_ipa_function_and_variable_visibility): Replace declaration with
15112 (make_pass_ipa_function_and_variable_visibility): ...new function.
15113 (pass_ipa_tree_profile): Replace declaration with that of...
15114 (make_pass_ipa_tree_profile): ...new function.
15115 (pass_early_local_passes): Replace declaration with that of...
15116 (make_pass_early_local_passes): ...new function.
15117 (pass_ipa_whole_program_visibility): Replace declaration with that
15119 (make_pass_ipa_whole_program_visibility): ...new function.
15120 (pass_ipa_lto_gimple_out): Replace declaration with that of...
15121 (make_pass_ipa_lto_gimple_out): ...new function.
15122 (pass_ipa_increase_alignment): Replace declaration with that of...
15123 (make_pass_ipa_increase_alignment): ...new function.
15124 (pass_ipa_inline): Replace declaration with that of...
15125 (make_pass_ipa_inline): ...new function.
15126 (pass_ipa_free_lang_data): Replace declaration with that of...
15127 (make_pass_ipa_free_lang_data): ...new function.
15128 (pass_ipa_free_inline_summary): Replace declaration with that of...
15129 (make_pass_ipa_free_inline_summary): ...new function.
15130 (pass_ipa_cp): Replace declaration with that of...
15131 (make_pass_ipa_cp): ...new function.
15132 (pass_ipa_reference): Replace declaration with that of...
15133 (make_pass_ipa_reference): ...new function.
15134 (pass_ipa_pure_const): Replace declaration with that of...
15135 (make_pass_ipa_pure_const): ...new function.
15136 (pass_ipa_pta): Replace declaration with that of...
15137 (make_pass_ipa_pta): ...new function.
15138 (pass_ipa_lto_finish_out): Replace declaration with that of...
15139 (make_pass_ipa_lto_finish_out): ...new function.
15140 (pass_ipa_tm): Replace declaration with that of...
15141 (make_pass_ipa_tm): ...new function.
15142 (pass_ipa_profile): Replace declaration with that of...
15143 (make_pass_ipa_profile): ...new function.
15144 (pass_ipa_cdtor_merge): Replace declaration with that of...
15145 (make_pass_ipa_cdtor_merge): ...new function.
15146 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
15148 (make_pass_cleanup_cfg_post_optimizing): ...new function.
15149 (pass_init_datastructures): Replace declaration with that of...
15150 (make_pass_init_datastructures): ...new function.
15151 (pass_fixup_cfg): Replace declaration with that of...
15152 (make_pass_fixup_cfg): ...new function.
15153 (pass_expand): Replace declaration with that of...
15154 (make_pass_expand): ...new function.
15155 (pass_instantiate_virtual_regs): Replace declaration with that of...
15156 (make_pass_instantiate_virtual_regs): ...new function.
15157 (pass_rtl_fwprop): Replace declaration with that of...
15158 (make_pass_rtl_fwprop): ...new function.
15159 (pass_rtl_fwprop_addr): Replace declaration with that of...
15160 (make_pass_rtl_fwprop_addr): ...new function.
15161 (pass_jump): Replace declaration with that of...
15162 (make_pass_jump): ...new function.
15163 (pass_jump2): Replace declaration with that of...
15164 (make_pass_jump2): ...new function.
15165 (pass_lower_subreg): Replace declaration with that of...
15166 (make_pass_lower_subreg): ...new function.
15167 (pass_cse): Replace declaration with that of...
15168 (make_pass_cse): ...new function.
15169 (pass_fast_rtl_dce): Replace declaration with that of...
15170 (make_pass_fast_rtl_dce): ...new function.
15171 (pass_ud_rtl_dce): Replace declaration with that of...
15172 (make_pass_ud_rtl_dce): ...new function.
15173 (pass_rtl_dce): Replace declaration with that of...
15174 (make_pass_rtl_dce): ...new function.
15175 (pass_rtl_dse1): Replace declaration with that of...
15176 (make_pass_rtl_dse1): ...new function.
15177 (pass_rtl_dse2): Replace declaration with that of...
15178 (make_pass_rtl_dse2): ...new function.
15179 (pass_rtl_dse3): Replace declaration with that of...
15180 (make_pass_rtl_dse3): ...new function.
15181 (pass_rtl_cprop): Replace declaration with that of...
15182 (make_pass_rtl_cprop): ...new function.
15183 (pass_rtl_pre): Replace declaration with that of...
15184 (make_pass_rtl_pre): ...new function.
15185 (pass_rtl_hoist): Replace declaration with that of...
15186 (make_pass_rtl_hoist): ...new function.
15187 (pass_rtl_store_motion): Replace declaration with that of...
15188 (make_pass_rtl_store_motion): ...new function.
15189 (pass_cse_after_global_opts): Replace declaration with that of...
15190 (make_pass_cse_after_global_opts): ...new function.
15191 (pass_rtl_ifcvt): Replace declaration with that of...
15192 (make_pass_rtl_ifcvt): ...new function.
15193 (pass_into_cfg_layout_mode): Replace declaration with that of...
15194 (make_pass_into_cfg_layout_mode): ...new function.
15195 (pass_outof_cfg_layout_mode): Replace declaration with that of...
15196 (make_pass_outof_cfg_layout_mode): ...new function.
15197 (pass_loop2): Replace declaration with that of...
15198 (make_pass_loop2): ...new function.
15199 (pass_rtl_loop_init): Replace declaration with that of...
15200 (make_pass_rtl_loop_init): ...new function.
15201 (pass_rtl_move_loop_invariants): Replace declaration with that of...
15202 (make_pass_rtl_move_loop_invariants): ...new function.
15203 (pass_rtl_unswitch): Replace declaration with that of...
15204 (make_pass_rtl_unswitch): ...new function.
15205 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
15206 (make_pass_rtl_unroll_and_peel_loops): ...new function.
15207 (pass_rtl_doloop): Replace declaration with that of...
15208 (make_pass_rtl_doloop): ...new function.
15209 (pass_rtl_loop_done): Replace declaration with that of...
15210 (make_pass_rtl_loop_done): ...new function.
15211 (pass_web): Replace declaration with that of...
15212 (make_pass_web): ...new function.
15213 (pass_cse2): Replace declaration with that of...
15214 (make_pass_cse2): ...new function.
15215 (pass_df_initialize_opt): Replace declaration with that of...
15216 (make_pass_df_initialize_opt): ...new function.
15217 (pass_df_initialize_no_opt): Replace declaration with that of...
15218 (make_pass_df_initialize_no_opt): ...new function.
15219 (pass_reginfo_init): Replace declaration with that of...
15220 (make_pass_reginfo_init): ...new function.
15221 (pass_inc_dec): Replace declaration with that of...
15222 (make_pass_inc_dec): ...new function.
15223 (pass_stack_ptr_mod): Replace declaration with that of...
15224 (make_pass_stack_ptr_mod): ...new function.
15225 (pass_initialize_regs): Replace declaration with that of...
15226 (make_pass_initialize_regs): ...new function.
15227 (pass_combine): Replace declaration with that of...
15228 (make_pass_combine): ...new function.
15229 (pass_if_after_combine): Replace declaration with that of...
15230 (make_pass_if_after_combine): ...new function.
15231 (pass_ree): Replace declaration with that of...
15232 (make_pass_ree): ...new function.
15233 (pass_partition_blocks): Replace declaration with that of...
15234 (make_pass_partition_blocks): ...new function.
15235 (pass_match_asm_constraints): Replace declaration with that of...
15236 (make_pass_match_asm_constraints): ...new function.
15237 (pass_regmove): Replace declaration with that of...
15238 (make_pass_regmove): ...new function.
15239 (pass_split_all_insns): Replace declaration with that of...
15240 (make_pass_split_all_insns): ...new function.
15241 (pass_fast_rtl_byte_dce): Replace declaration with that of...
15242 (make_pass_fast_rtl_byte_dce): ...new function.
15243 (pass_lower_subreg2): Replace declaration with that of...
15244 (make_pass_lower_subreg2): ...new function.
15245 (pass_mode_switching): Replace declaration with that of...
15246 (make_pass_mode_switching): ...new function.
15247 (pass_sms): Replace declaration with that of...
15248 (make_pass_sms): ...new function.
15249 (pass_sched): Replace declaration with that of...
15250 (make_pass_sched): ...new function.
15251 (pass_ira): Replace declaration with that of...
15252 (make_pass_ira): ...new function.
15253 (pass_reload): Replace declaration with that of...
15254 (make_pass_reload): ...new function.
15255 (pass_clean_state): Replace declaration with that of...
15256 (make_pass_clean_state): ...new function.
15257 (pass_branch_prob): Replace declaration with that of...
15258 (make_pass_branch_prob): ...new function.
15259 (pass_value_profile_transformations): Replace declaration with that
15261 (make_pass_value_profile_transformations): ...new function.
15262 (pass_postreload_cse): Replace declaration with that of...
15263 (make_pass_postreload_cse): ...new function.
15264 (pass_gcse2): Replace declaration with that of...
15265 (make_pass_gcse2): ...new function.
15266 (pass_split_after_reload): Replace declaration with that of...
15267 (make_pass_split_after_reload): ...new function.
15268 (pass_branch_target_load_optimize1): Replace declaration with that
15270 (make_pass_branch_target_load_optimize1): ...new function.
15271 (pass_thread_prologue_and_epilogue): Replace declaration with that
15273 (make_pass_thread_prologue_and_epilogue): ...new function.
15274 (pass_stack_adjustments): Replace declaration with that of...
15275 (make_pass_stack_adjustments): ...new function.
15276 (pass_peephole2): Replace declaration with that of...
15277 (make_pass_peephole2): ...new function.
15278 (pass_if_after_reload): Replace declaration with that of...
15279 (make_pass_if_after_reload): ...new function.
15280 (pass_regrename): Replace declaration with that of...
15281 (make_pass_regrename): ...new function.
15282 (pass_cprop_hardreg): Replace declaration with that of...
15283 (make_pass_cprop_hardreg): ...new function.
15284 (pass_reorder_blocks): Replace declaration with that of...
15285 (make_pass_reorder_blocks): ...new function.
15286 (pass_branch_target_load_optimize2): Replace declaration with that
15288 (make_pass_branch_target_load_optimize2): ...new function.
15289 (pass_leaf_regs): Replace declaration with that of...
15290 (make_pass_leaf_regs): ...new function.
15291 (pass_split_before_sched2): Replace declaration with that of...
15292 (make_pass_split_before_sched2): ...new function.
15293 (pass_compare_elim_after_reload): Replace declaration with that of...
15294 (make_pass_compare_elim_after_reload): ...new function.
15295 (pass_sched2): Replace declaration with that of...
15296 (make_pass_sched2): ...new function.
15297 (pass_stack_regs): Replace declaration with that of...
15298 (make_pass_stack_regs): ...new function.
15299 (pass_stack_regs_run): Replace declaration with that of...
15300 (make_pass_stack_regs_run): ...new function.
15301 (pass_df_finish): Replace declaration with that of...
15302 (make_pass_df_finish): ...new function.
15303 (pass_compute_alignments): Replace declaration with that of...
15304 (make_pass_compute_alignments): ...new function.
15305 (pass_duplicate_computed_gotos): Replace declaration with that of...
15306 (make_pass_duplicate_computed_gotos): ...new function.
15307 (pass_variable_tracking): Replace declaration with that of...
15308 (make_pass_variable_tracking): ...new function.
15309 (pass_free_cfg): Replace declaration with that of...
15310 (make_pass_free_cfg): ...new function.
15311 (pass_machine_reorg): Replace declaration with that of...
15312 (make_pass_machine_reorg): ...new function.
15313 (pass_cleanup_barriers): Replace declaration with that of...
15314 (make_pass_cleanup_barriers): ...new function.
15315 (pass_delay_slots): Replace declaration with that of...
15316 (make_pass_delay_slots): ...new function.
15317 (pass_split_for_shorten_branches): Replace declaration with that of...
15318 (make_pass_split_for_shorten_branches): ...new function.
15319 (pass_split_before_regstack): Replace declaration with that of...
15320 (make_pass_split_before_regstack): ...new function.
15321 (pass_convert_to_eh_region_ranges): Replace declaration with that
15323 (make_pass_convert_to_eh_region_ranges): ...new function.
15324 (pass_shorten_branches): Replace declaration with that of...
15325 (make_pass_shorten_branches): ...new function.
15326 (pass_set_nothrow_function_flags): Replace declaration with that of...
15327 (make_pass_set_nothrow_function_flags): ...new function.
15328 (pass_dwarf2_frame): Replace declaration with that of...
15329 (make_pass_dwarf2_frame): ...new function.
15330 (pass_final): Replace declaration with that of...
15331 (make_pass_final): ...new function.
15332 (pass_rtl_seqabstr): Replace declaration with that of...
15333 (make_pass_rtl_seqabstr): ...new function.
15334 (pass_release_ssa_names): Replace declaration with that of...
15335 (make_pass_release_ssa_names): ...new function.
15336 (pass_early_inline): Replace declaration with that of...
15337 (make_pass_early_inline): ...new function.
15338 (pass_inline_parameters): Replace declaration with that of...
15339 (make_pass_inline_parameters): ...new function.
15340 (pass_update_address_taken): Replace declaration with that of...
15341 (make_pass_update_address_taken): ...new function.
15342 (pass_convert_switch): Replace declaration with that of...
15343 (make_pass_convert_switch): ...new function.
15344 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
15345 to a subclass of simple_ipa_opt_pass along with...
15346 (pass_data_ipa_tree_profile): ...new pass_data instance and...
15347 (make_pass_ipa_tree_profile): ...new function.
15348 * tree-sra.c (pass_sra_early): Convert from a global struct to a
15349 subclass of gimple_opt_pass along with...
15350 (pass_data_sra_early): ...new pass_data instance and...
15351 (make_pass_sra_early): ...new function.
15352 (pass_sra): Convert from a global struct to a subclass of
15353 gimple_opt_pass along with...
15354 (pass_data_sra): ...new pass_data instance and...
15355 (make_pass_sra): ...new function.
15356 (pass_early_ipa_sra): Convert from a global struct to a subclass of
15357 gimple_opt_pass along with...
15358 (pass_data_early_ipa_sra): ...new pass_data instance and...
15359 (make_pass_early_ipa_sra): ...new function.
15360 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
15361 subclass of gimple_opt_pass along with...
15362 (pass_data_ccp): ...new pass_data instance and...
15363 (make_pass_ccp): ...new function.
15364 (pass_fold_builtins): Convert from a global struct to a subclass of
15365 gimple_opt_pass along with...
15366 (pass_data_fold_builtins): ...new pass_data instance and...
15367 (make_pass_fold_builtins): ...new function.
15368 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
15369 subclass of gimple_opt_pass along with...
15370 (pass_data_copy_prop): ...new pass_data instance and...
15371 (make_pass_copy_prop): ...new function.
15372 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
15373 global struct to a subclass of gimple_opt_pass along with...
15374 (pass_data_rename_ssa_copies): ...new pass_data instance and...
15375 (make_pass_rename_ssa_copies): ...new function.
15376 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
15377 subclass of gimple_opt_pass along with...
15378 (pass_data_dce): ...new pass_data instance and...
15379 (make_pass_dce): ...new function.
15380 (pass_dce_loop): Convert from a global struct to a subclass of
15381 gimple_opt_pass along with...
15382 (pass_data_dce_loop): ...new pass_data instance and...
15383 (make_pass_dce_loop): ...new function.
15384 (pass_cd_dce): Convert from a global struct to a subclass of
15385 gimple_opt_pass along with...
15386 (pass_data_cd_dce): ...new pass_data instance and...
15387 (make_pass_cd_dce): ...new function.
15388 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
15389 subclass of gimple_opt_pass along with...
15390 (pass_data_dominator): ...new pass_data instance and...
15391 (make_pass_dominator): ...new function.
15392 (pass_phi_only_cprop): Convert from a global struct to a subclass of
15393 gimple_opt_pass along with...
15394 (pass_data_phi_only_cprop): ...new pass_data instance and...
15395 (make_pass_phi_only_cprop): ...new function.
15396 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
15397 subclass of gimple_opt_pass along with...
15398 (pass_data_dse): ...new pass_data instance and...
15399 (make_pass_dse): ...new function.
15400 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
15401 a subclass of gimple_opt_pass along with...
15402 (pass_data_forwprop): ...new pass_data instance and...
15403 (make_pass_forwprop): ...new function.
15404 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
15405 struct to a subclass of gimple_opt_pass along with...
15406 (pass_data_tree_ifcombine): ...new pass_data instance and...
15407 (make_pass_tree_ifcombine): ...new function.
15408 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
15409 subclass of gimple_opt_pass along with...
15410 (pass_data_ch): ...new pass_data instance and...
15411 (make_pass_ch): ...new function.
15412 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
15413 subclass of gimple_opt_pass along with...
15414 (pass_data_tree_loop): ...new pass_data instance and...
15415 (make_pass_tree_loop): ...new function.
15416 (pass_tree_loop_init): Convert from a global struct to a subclass of
15417 gimple_opt_pass along with...
15418 (pass_data_tree_loop_init): ...new pass_data instance and...
15419 (make_pass_tree_loop_init): ...new function.
15420 (pass_lim): Convert from a global struct to a subclass of
15421 gimple_opt_pass along with...
15422 (pass_data_lim): ...new pass_data instance and...
15423 (make_pass_lim): ...new function.
15424 (pass_tree_unswitch): Convert from a global struct to a subclass of
15425 gimple_opt_pass along with...
15426 (pass_data_tree_unswitch): ...new pass_data instance and...
15427 (make_pass_tree_unswitch): ...new function.
15428 (pass_predcom): Convert from a global struct to a subclass of
15429 gimple_opt_pass along with...
15430 (pass_data_predcom): ...new pass_data instance and...
15431 (make_pass_predcom): ...new function.
15432 (pass_vectorize): Convert from a global struct to a subclass of
15433 gimple_opt_pass along with...
15434 (pass_data_vectorize): ...new pass_data instance and...
15435 (make_pass_vectorize): ...new function.
15436 (pass_graphite): Convert from a global struct to a subclass of
15437 gimple_opt_pass along with...
15438 (pass_data_graphite): ...new pass_data instance and...
15439 (make_pass_graphite): ...new function.
15440 (pass_graphite_transforms): Convert from a global struct to a subclass
15441 of gimple_opt_pass along with...
15442 (pass_data_graphite_transforms): ...new pass_data instance and...
15443 (make_pass_graphite_transforms): ...new function.
15444 (pass_check_data_deps): Convert from a global struct to a subclass of
15445 gimple_opt_pass along with...
15446 (pass_data_check_data_deps): ...new pass_data instance and...
15447 (make_pass_check_data_deps): ...new function.
15448 (pass_iv_canon): Convert from a global struct to a subclass of
15449 gimple_opt_pass along with...
15450 (pass_data_iv_canon): ...new pass_data instance and...
15451 (make_pass_iv_canon): ...new function.
15452 (pass_scev_cprop): Convert from a global struct to a subclass of
15453 gimple_opt_pass along with...
15454 (pass_data_scev_cprop): ...new pass_data instance and...
15455 (make_pass_scev_cprop): ...new function.
15456 (pass_record_bounds): Convert from a global struct to a subclass of
15457 gimple_opt_pass along with...
15458 (pass_data_record_bounds): ...new pass_data instance and...
15459 (make_pass_record_bounds): ...new function.
15460 (pass_complete_unroll): Convert from a global struct to a subclass of
15461 gimple_opt_pass along with...
15462 (pass_data_complete_unroll): ...new pass_data instance and...
15463 (make_pass_complete_unroll): ...new function.
15464 (pass_complete_unrolli): Convert from a global struct to a subclass of
15465 gimple_opt_pass along with...
15466 (pass_data_complete_unrolli): ...new pass_data instance and...
15467 (make_pass_complete_unrolli): ...new function.
15468 (pass_parallelize_loops): Convert from a global struct to a subclass
15469 of gimple_opt_pass along with...
15470 (pass_data_parallelize_loops): ...new pass_data instance and...
15471 (make_pass_parallelize_loops): ...new function.
15472 (pass_loop_prefetch): Convert from a global struct to a subclass of
15473 gimple_opt_pass along with...
15474 (pass_data_loop_prefetch): ...new pass_data instance and...
15475 (make_pass_loop_prefetch): ...new function.
15476 (pass_iv_optimize): Convert from a global struct to a subclass of
15477 gimple_opt_pass along with...
15478 (pass_data_iv_optimize): ...new pass_data instance and...
15479 (make_pass_iv_optimize): ...new function.
15480 (pass_tree_loop_done): Convert from a global struct to a subclass of
15481 gimple_opt_pass along with...
15482 (pass_data_tree_loop_done): ...new pass_data instance and...
15483 (make_pass_tree_loop_done): ...new function.
15484 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
15485 struct to a subclass of gimple_opt_pass along with...
15486 (pass_data_cse_reciprocals): ...new pass_data instance and...
15487 (make_pass_cse_reciprocals): ...new function.
15488 (pass_cse_sincos): Convert from a global struct to a subclass of
15489 gimple_opt_pass along with...
15490 (pass_data_cse_sincos): ...new pass_data instance and...
15491 (make_pass_cse_sincos): ...new function.
15492 (pass_optimize_bswap): Convert from a global struct to a subclass of
15493 gimple_opt_pass along with...
15494 (pass_data_optimize_bswap): ...new pass_data instance and...
15495 (make_pass_optimize_bswap): ...new function.
15496 (pass_optimize_widening_mul): Convert from a global struct to a
15497 subclass of gimple_opt_pass along with...
15498 (pass_data_optimize_widening_mul): ...new pass_data instance and...
15499 (make_pass_optimize_widening_mul): ...new function.
15500 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
15501 subclass of gimple_opt_pass along with...
15502 (pass_data_phiopt): ...new pass_data instance and...
15503 (make_pass_phiopt): ...new function.
15504 (pass_cselim): Convert from a global struct to a subclass of
15505 gimple_opt_pass along with...
15506 (pass_data_cselim): ...new pass_data instance and...
15507 (make_pass_cselim): ...new function.
15508 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
15509 subclass of gimple_opt_pass along with...
15510 (pass_data_phiprop): ...new pass_data instance and...
15511 (make_pass_phiprop): ...new function.
15512 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
15513 subclass of gimple_opt_pass along with...
15514 (pass_data_pre): ...new pass_data instance and...
15515 (make_pass_pre): ...new function.
15516 (pass_fre): Convert from a global struct to a subclass of
15517 gimple_opt_pass along with...
15518 (pass_data_fre): ...new pass_data instance and...
15519 (make_pass_fre): ...new function.
15520 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
15521 subclass of gimple_opt_pass along with...
15522 (pass_data_reassoc): ...new pass_data instance and...
15523 (make_pass_reassoc): ...new function.
15524 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
15525 subclass of gimple_opt_pass along with...
15526 (pass_data_sink_code): ...new pass_data instance and...
15527 (make_pass_sink_code): ...new function.
15528 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
15529 subclass of gimple_opt_pass along with...
15530 (pass_data_strlen): ...new pass_data instance and...
15531 (make_pass_strlen): ...new function.
15532 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
15533 struct to a subclass of gimple_opt_pass along with...
15534 (pass_data_build_alias): ...new pass_data instance and...
15535 (make_pass_build_alias): ...new function.
15536 (pass_build_ealias): Convert from a global struct to a subclass of
15537 gimple_opt_pass along with...
15538 (pass_data_build_ealias): ...new pass_data instance and...
15539 (make_pass_build_ealias): ...new function.
15540 (pass_ipa_pta): Convert from a global struct to a subclass of
15541 simple_ipa_opt_pass along with...
15542 (pass_data_ipa_pta): ...new pass_data instance and...
15543 (make_pass_ipa_pta): ...new function.
15544 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
15545 subclass of gimple_opt_pass along with...
15546 (pass_data_uncprop): ...new pass_data instance and...
15547 (make_pass_uncprop): ...new function.
15548 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
15549 global struct to a subclass of gimple_opt_pass along with...
15550 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
15551 (make_pass_late_warn_uninitialized): ...new function.
15552 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
15553 to a subclass of gimple_opt_pass along with...
15554 (pass_data_init_datastructures): ...new pass_data instance and...
15555 (make_pass_init_datastructures): ...new function.
15556 (pass_early_warn_uninitialized): Convert from a global struct to a
15557 subclass of gimple_opt_pass along with...
15558 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
15559 (make_pass_early_warn_uninitialized): ...new function.
15560 (pass_update_address_taken): Convert from a global struct to a
15561 subclass of gimple_opt_pass along with...
15562 (pass_data_update_address_taken): ...new pass_data instance and...
15563 (make_pass_update_address_taken): ...new function.
15564 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
15565 struct to a subclass of gimple_opt_pass along with...
15566 (pass_data_release_ssa_names): ...new pass_data instance and...
15567 (make_pass_release_ssa_names): ...new function.
15568 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
15569 subclass of gimple_opt_pass along with...
15570 (pass_data_stdarg): ...new pass_data instance and...
15571 (make_pass_stdarg): ...new function.
15572 * tree-switch-conversion.c (pass_convert_switch): Convert from a
15573 global struct to a subclass of gimple_opt_pass along with...
15574 (pass_data_convert_switch): ...new pass_data instance and...
15575 (make_pass_convert_switch): ...new function.
15576 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
15577 to a subclass of gimple_opt_pass along with...
15578 (pass_data_tail_recursion): ...new pass_data instance and...
15579 (make_pass_tail_recursion): ...new function.
15580 (pass_tail_calls): Convert from a global struct to a subclass of
15581 gimple_opt_pass along with...
15582 (pass_data_tail_calls): ...new pass_data instance and...
15583 (make_pass_tail_calls): ...new function.
15584 * tree-vect-generic.c (pass_lower_vector): Convert from a global
15585 struct to a subclass of gimple_opt_pass along with...
15586 (pass_data_lower_vector): ...new pass_data instance and...
15587 (make_pass_lower_vector): ...new function.
15588 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
15589 gimple_opt_pass along with...
15590 (pass_data_lower_vector_ssa): ...new pass_data instance and...
15591 (make_pass_lower_vector_ssa): ...new function.
15592 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
15593 to a subclass of gimple_opt_pass along with...
15594 (pass_data_slp_vectorize): ...new pass_data instance and...
15595 (make_pass_slp_vectorize): ...new function.
15596 (pass_ipa_increase_alignment): Convert from a global struct to a
15597 subclass of simple_ipa_opt_pass along with...
15598 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
15599 (make_pass_ipa_increase_alignment): ...new function.
15600 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
15601 gimple_opt_pass along with...
15602 (pass_data_vrp): ...new pass_data instance and...
15603 (make_pass_vrp): ...new function.
15604 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
15605 subclass of simple_ipa_opt_pass along with...
15606 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
15607 (make_pass_ipa_free_lang_data): ...new function.
15608 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
15609 gimple_opt_pass along with...
15610 (pass_data_tsan): ...new pass_data instance and...
15611 (make_pass_tsan): ...new function.
15612 (pass_tsan_O0): Convert from a global struct to a subclass of
15613 gimple_opt_pass along with...
15614 (pass_data_tsan_O0): ...new pass_data instance and...
15615 (make_pass_tsan_O0): ...new function.
15616 * var-tracking.c (pass_variable_tracking): Convert from a global
15617 struct to a subclass of rtl_opt_pass along with...
15618 (pass_data_variable_tracking): ...new pass_data instance and...
15619 (make_pass_variable_tracking): ...new function.
15620 * web.c (pass_web): Convert from a global struct to a subclass of
15621 rtl_opt_pass along with...
15622 (pass_data_web): ...new pass_data instance and...
15623 (make_pass_web): ...new function.
15624 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
15625 declaration with that of...
15626 (make_pass_mode_switch_use): ...new function.
15627 (pass_resolve_sw_modes): Replace declaration with that of...
15628 (make_pass_resolve_sw_modes): ...new function.
15629 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
15630 from a global struct to a subclass of rtl_opt_pass along with...
15631 (pass_data_mode_switch_use): ...new pass_data instance and...
15632 (make_pass_mode_switch_use): ...new function.
15633 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
15634 from a global struct to a subclass of rtl_opt_pass along with...
15635 (pass_data_resolve_sw_modes): ...new pass_data instance and...
15636 (make_pass_resolve_sw_modes): ...new function.
15637 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
15638 struct to a subclass of rtl_opt_pass along with...
15639 (pass_data_insert_vzeroupper): ...new pass_data instance and...
15640 (make_pass_insert_vzeroupper): ...new function.
15641 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
15642 global struct to a subclass of rtl_opt_pass along with...
15643 (pass_data_work_around_errata): ...new pass_data instance and...
15644 (make_pass_work_around_errata): ...new function.
15645 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
15646 struct to a subclass of rtl_opt_pass along with...
15647 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
15648 (make_pass_mips_machine_reorg2): ...new function.
15650 2013-08-05 David Malcolm <dmalcolm@redhat.com>
15652 * passes.c (pass_manager::operator new): New.
15654 2013-08-05 David Malcolm <dmalcolm@redhat.com>
15656 Handwritten part of conversion of passes to C++ classes.
15658 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
15659 (toplev.o): Add dep on PASS_MANAGER_H.
15660 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
15661 of early local pases to reflect this moving from a global to a
15662 member of gcc::pass_manager.
15663 (cgraph_add_new_function): Likewise.
15664 * lto-cgraph.c (lto_output_node): Update for conversion of
15665 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
15666 * passes.c (opt_pass::clone): New.
15667 (opt_pass::gate): New.
15668 (opt_pass::execute): New.
15669 (opt_pass::opt_pass): New.
15670 (pass_manager::execute_early_local_passes): New.
15671 (pass_manager::execute_pass_mode_switching): new.
15672 (finish_optimization_passes): Convert to...
15673 (pass_manager::finish_optimization_passes): ...this.
15674 (finish_optimization_passes): Update for conversion of passes to
15676 (register_dump_files_1): Use has_gate since we cannot portably
15677 check a vtable entry against NULL.
15678 (dump_one_pass): Likewise.
15679 (ipa_write_summaries_2): Likewise.
15680 (ipa_write_optimization_summaries_1): Likewise.
15681 (ipa_read_summaries_1): Likewise.
15682 (ipa_read_optimization_summaries_1): Likewise.
15683 (execute_ipa_stmt_fixups): Likewise.
15684 (pass_manager::pass_manager): Rewrite pass-creation, invoking
15685 pass-creation functions rather than wiring up globals, and
15686 storing the results in fields of pass_manager generated using
15687 pass-instances.def.
15688 (pass_manager::dump_profile_report): Update for conversion of
15689 passes to C++ classes.
15690 (pass_manager::execute_ipa_summary_passes): Likewise.
15691 (execute_one_ipa_transform_pass): Likewise.
15692 (execute_one_pass): Use has_gate and has_execute since we cannot
15693 portably check a vtable entry against NULL.
15694 * pass_manager.h (pass_manager::finish_optimization_passes): New.
15695 (pass_manager): Use pass-instances.def to add fields for the
15696 various pass instances.
15697 * toplev.c (finalize): Update for move of
15698 finish_optimization_passes to a method of gcc::pass_manager.
15699 * toplev.h (finish_optimization_passes): Move to method of class
15701 * tree-pass.h (struct pass_data): New.
15702 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
15703 (opt_pass::gate): Convert to virtual function.
15704 (opt_pass::~opt_pass): New.
15705 (opt_pass::clone): New.
15706 (opt_pass::execute): Convert to virtual function.
15707 (opt_pass::opt_pass): New.
15708 (opt_pass::ctxt_): new.
15709 (gimple_opt_pass): Convert to subclass of opt_pass.
15710 (gimple_opt_pass::gimple_opt_pass): New.
15711 (rtl_opt_pass): Convert to subclass of opt_pass.
15712 (rtl_opt_pass::rtl_opt_pass): New.
15713 (ipa_opt_pass_d): Convert to subclass of opt_pass.
15714 (ipa_opt_pass_d::ipa_opt_pass_d): New.
15715 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
15716 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
15717 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
15718 invocation of pass_mode_switching to reflect this moving from a
15719 global to a member of gcc::pass_manager.
15720 (ix86_option_override): Rework how pass_insert_vzeroupper is
15721 added to the pass_manager to reflect autogenerated changes.
15722 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
15724 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
15726 PR rtl-optimization/57708
15727 * recog.c (peep2_find_free_register): Validate all regs in a
15730 2013-08-05 Jan Hubicka <jh@suse.cz>
15733 * cgraph.c (verify_cgraph_node): Accept local flags from other
15735 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
15736 (function_and_variable_visibility): Likewise.
15737 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
15739 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
15741 * graph.c (init_graph_slim_pretty_print): Remove.
15742 (print_graph_cfg): Do not call it. Use local pretty printer.
15743 (start_graph_dump): Likewise.
15745 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
15747 * gimple-pretty-print.c (buffer): Remove.
15748 (initialized): Likewise.
15749 (maybe_init_pretty_print): Likewise.
15750 (print_gimple_stmt): Do not call it. Use non-static local
15751 pretty_printer variable.
15752 (print_gimple_expr): Likewise.
15753 (print_gimple_seq): Likewise.
15754 (gimple_dump_bb): Likewise.
15756 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
15758 * asan.c (asan_pp): Remove.
15759 (asan_pp_initialized): Likewise.
15760 (asan_pp_initialize): Likewise.
15761 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
15762 (asan_emit_stack_protection): Tidy. Use local pretty printer.
15763 (asan_add_global): Likewise.
15765 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
15767 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
15768 * diagnostic.h (diagnostic_flush_buffer): Adjust.
15769 * pretty-print.c (pp_formatted_text_data): Likewise.
15770 (pp_indent): Rename from pp_base_indent.
15771 (pp_format): Rename from pp_base_format.
15772 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
15773 (pp_format_verbatim): Rename from pp_base_format_verbatim.
15774 (pp_flush): Rename from pp_base_flush.
15775 (pp_set_line_maximum_length): Rename from
15776 pp_base_set_line_maximum_length.
15777 (pp_clear_output_area): Rename from pp_base_clear_output_area.
15778 (pp_set_prefix): Rename from pp_base_set_prefix.
15779 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
15780 (pp_emit_prefix): Rename from pp_base_emit_prefix.
15781 (pp_append_text): Rename from pp_base_append_text.
15782 (pp_formatted_text): Rename from pp_base_formatted_text.
15783 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
15784 (pp_remaining_character_count_for_line): Rename from
15785 pp_base_remaining_character_count_for_line.
15786 (pp_newline): Rename from pp_base_newline.
15787 (pp_character): Rename from pp_base_character.
15788 (pp_string): Rename from pp_base_string.
15789 (pp_maybe_space): Rename from pp_base_maybe_space.
15790 * asan.c (asan_pp_string): Adjust.
15791 (asan_emit_stack_protection): Likewise.
15792 (asan_add_global): Likewise.
15793 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
15794 * tree-mudflap.c (mf_varname_tree): Likewise.
15795 * tree-pretty-print.c (pp_tree_identifier): Rename from
15796 pp_base_tree_identifier.
15797 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
15798 Declare as function.
15800 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
15802 * pretty-print.h (pp_bar_bar): New.
15803 (pp_ampersand_ampersand): Likewise.
15804 (pp_less_equal): Likewise.
15805 (pp_greater_equal): Likewise.
15806 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
15807 printer functions instead of pp_string or operators and punctuators.
15808 (dump_gimple_call): Likewise.
15809 (dump_gimple_omp_for): Likewise.
15810 (dump_gimple_transaction): Likewise.
15811 (dump_gimple_phi): Likewise.
15812 (pp_gimple_stmt_1): Likewise.
15813 * sched-vis.c (print_insn): Likewise.
15814 * tree-mudflap.c (mf_varname_tree): Likewise.
15815 * tree-pretty-print.c (dump_block_node): Likewise.
15816 (dump_generic_node): Likewise.
15818 2013-08-02 Jan Hubicka <jh@suse.cz>
15820 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
15822 * lto-streamer-out.c (tree_is_indexable): Results decls and
15823 parm decls are not indexable.
15824 (DFS_write_tree_body): Do not follow args and results.
15825 (hash_tree): Likewise.
15826 (output_functions): Rearrange so struct function is needed
15827 only when real body is output; be able to also ouptut abstract
15828 functions; output DECL_ARGUMENTS and DECL_RESULT.
15829 (lto_output): When not in WPA, ale store abstract functions.
15830 (write_symbol): Do not care about RESULT_DECL.
15831 (output_symbol_p): Handle correctly sbtract decls.
15832 * lto-streamer-in.c (input_function): Rearrange so struct
15833 function can be NULL at entry; allow streaming of
15834 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
15835 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
15836 sanity check during LTO.
15837 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
15838 RESULT_DECl and DECL_ARGUMENTS.
15839 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
15842 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
15844 * pretty-print.h (pp_underscore): New.
15846 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
15847 printer functions instead of pp_character.
15848 (dump_binary_rhs): Likewise.
15849 (dump_ternary_rhs): Likewise.
15850 (dump_gimple_call_args): Likewise.
15851 (pp_points_to_solution): Likewise.
15852 (dump_gimple_call): Likewise.
15853 (dump_gimple_switch): Likewise.
15854 (dump_gimple_cond): Likewise.
15855 (dump_gimple_bind): Likewise.
15856 (dump_gimple_try): Likewise.
15857 (dump_gimple_omp_for): Likewise.
15858 (dump_gimple_omp_continue): Likewise.
15859 (dump_gimple_omp_single): Likewise.
15860 (dump_gimple_omp_sections): Likewise.
15861 (dump_gimple_omp_block): Likewise.
15862 (dump_gimple_omp_critical): Likewise.
15863 (dump_gimple_transaction): Likewise.
15864 (dump_gimple_asm): Likewise.
15865 (dump_gimple_phi): Likewise.
15866 (dump_gimple_omp_parallel): Likewise.
15867 (dump_gimple_omp_task): Likewise.
15868 (dump_gimple_omp_atomic_load): Likewise.
15869 (dump_gimple_omp_atomic_store): Likewise.
15870 (dump_gimple_mem_ops): Likewise.
15871 (pp_gimple_stmt_1): Likewise.
15872 (pp_cfg_jump): Likewise.
15873 (dump_implicit_edges): Likewise.
15874 (gimple_dump_bb_for_graph): Likewise.
15875 * graph.c (draw_cfg_node): Likewise.
15876 * langhooks.c (lhd_print_error_function): Likewise.
15877 * sched-vis.c (print_exp): Likewise.
15878 (print_value): Likewise.
15879 (print_pattern): Likewise.
15880 (print_insn): Likewise.
15881 (rtl_dump_bb_for_graph): Likewise.
15882 * tree-pretty-print.c (dump_function_declaration): Likewise.
15883 (dump_array_domain): Likewise.
15884 (dump_omp_clause): Likewise.
15885 (dump_location): Likewise.
15886 (dump_generic_node): Likewise.
15887 (print_struct_decl): Likewise.
15888 * diagnostic.c (diagnostic_show_locus): Use pp_space.
15890 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15892 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
15893 candidate table when replacing a candidate statement.
15894 (replace_rhs_if_not_dup): Likewise.
15895 (replace_one_candidate): Likewise.
15897 2013-08-02 Jan Hubicka <jh@suse.cz>
15898 Martin Liska <marxin.liska@gmail.com>
15900 * cgraphunit.c (add_new_function): Fix logic when adding from
15902 (assemble_thunk): Rename to ...
15903 (expand_thunk); .. this one; export; get it working with
15904 general functions; make produced gimple valid.
15905 * cgraph.h (expand_thunk): Declare.
15907 2013-08-02 Jan Hubicka <jh@suse.cz>
15909 * ipa-cp.c (gather_context_independent_values): Use
15910 ipa_get_param_move_cost.
15911 (get_replacement_map): Remove PARAM; move parameter folding
15913 (create_specialized_node): Update.
15914 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
15915 assert that we have gimple body; update move_cost.
15916 (count_formal_params): Assert that we have gimple body.
15917 (ipa_dump_param): New function.
15918 (ipa_alloc_node_params): Break out from ...
15919 (ipa_initialize_node_params): ... here.
15920 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
15921 (ipa_write_node_info): Stream move costs.
15922 (ipa_read_node_info): Read move costs.
15923 (ipa_update_after_lto_read): Do not recompute node params.
15924 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
15925 (ipa_get_param): Check we are not in WPA.
15926 (ipa_get_param_move_cost): New.
15927 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
15928 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
15929 parm numbers to be present.
15931 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
15933 PR rtl-optimization/58048
15934 * lra-constraints.c (process_alt_operands): Don't check asm
15935 operand on register.
15937 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
15939 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
15940 the implied StoreLoad barrier for atomic operations if before.
15942 2013-08-02 Jan Hubicka <jh@suse.cz>
15943 Martin Liska <marxin.liska@gmail.com>
15945 * cgraph.c (cgraph_function_body_availability): Do not check
15947 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
15948 symtab_node_availability): Declare.
15949 * ipa.c (can_replace_by_local_alias): New.
15950 (function_and_variable_visibility): Use it.
15951 * symtab.c (symtab_for_node_and_aliases,
15952 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
15954 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
15956 PR rtl-optimization/57963
15957 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
15958 (lra_constraints): Use them.
15960 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
15962 * config/arm/types.md (define_attr "type"): Add "load_acq"
15964 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
15966 (cortex_a53_store1): Likewise.
15968 2013-08-01 Jan Hubicka <jh@suse.cz>
15970 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
15971 partitions are not needed.
15973 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
15975 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
15976 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
15977 MAYBE_NON_Q_CLASS_P where appropriate.
15979 2013-08-01 Jan Hubicka <jh@suse.cz>
15981 * cgraph.h (release_function_body): Declare.
15982 * tree.c (free_lang_data_in_decl): Free, parameters and return values
15983 of unused delcarations.
15985 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15987 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
15988 RTL form when subtracting a constant.
15990 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15992 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
15993 Generate canonical plus rtx with negated immediate instead of minus
15995 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
15997 2013-08-01 Jan Hubicka <jh@suse.cz>
15999 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
16000 (cgraph_release_function_body): Likewise.
16001 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
16002 * cgraph.h (cgrpah_node): Rename abstract_and_needed
16003 to used_as_abstract_origin.
16004 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
16005 symbols used as abstract origins.
16006 * cgraphunit.c (analyze_functions): Update.
16007 * ipa.c (symtab_remove_unreachable_nodes): Recompute
16008 used_as_abstract_origin.
16009 * tree-inline.c (tree_function_versioning): Update
16010 used_as_abstract_origin; be ready for DECL_RESULT and
16011 DECL_ARGUMENTS to be NULL.
16013 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
16014 for abstract functions.
16015 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
16018 2013-08-01 Jan Hubicka <jh@suse.cz>
16020 * profile.c (compute_value_histograms): Fix thinko.
16022 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
16024 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
16025 aarch-common-protos.h to extra_headers.
16026 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
16027 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
16028 * config/aarch64/t-aarch64 (aarch-common.o): Define.
16030 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
16032 * config/aarch64/aarch64.md (define_attr "type"): Delete.
16033 Include "../arm/types.md". Define "type" attribute for all patterns.
16034 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
16037 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
16039 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
16040 to support power8 load fusion.
16041 (fusion_gpr_mem_load): Likewise.
16043 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
16045 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
16046 declarations for power8 load fusion.
16047 (emit_fusion_gpr_load): Likewise.
16049 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
16050 tuning for power8, turn on fusion mode by default. Turn on sign
16051 extending fusion mode if normal fusion mode is on, and we are at
16053 (fusion_gpr_load_p): New function, return true if we can fuse an
16054 addis instruction with a dependent load to a GPR.
16055 (emit_fusion_gpr_load): Emit the instructions for power8 load
16058 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
16059 (VSX load fusion peepholes): New peepholes to fuse together an
16060 addi instruction with a VSX load instruction.
16062 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
16063 peepholes to fuse an addis instruction with a load to a GPR base
16064 register. If we are supporting sign extending fusions, convert
16065 sign extending loads to zero extending loads and add an explicit
16068 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
16070 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
16071 aarch-common-protos.h to extra_headers.
16072 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
16073 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
16074 (arm_early_store_addr_dep): Likewise.
16075 (arm_no_early_alu_shift_dep): Likewise.
16076 (arm_no_early_alu_shift_value_dep): Likewise.
16077 (arm_no_early_mul_dep): Likewise.
16078 (arm_no_early_store_addr_dep): Likewise.
16079 (arm_mac_accumulator_is_mul_result): Likewise.
16080 (arm_mac_accumulator_is_result): Likewise.
16081 * config/arm/aarch-common.c: ... here. New file.
16082 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
16084 (arm_early_store_addr_dep): Likewise.
16085 (arm_no_early_alu_shift_dep): Likewise.
16086 (arm_no_early_alu_shift_value_dep): Likewise.
16087 (arm_no_early_mul_dep): Likewise.
16088 (arm_no_early_store_addr_dep): Likewise.
16089 (arm_mac_accumulator_is_mul_result): Likewise.
16090 (arm_mac_accumulator_is_result): Likewise.
16091 * config/arm/aarch-common-protos.h: ... here. New file.
16092 * config/arm/t-arm (aarch-common.o): Define.
16094 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
16096 * config/arm/arm.md: Include new file "types.md".
16097 (define_attr "type"): Move from here to ...
16098 (define_attr "mul32"): Likewise.
16099 (define_attr "mul64"): Likewise.
16100 * config/arm/types.md: ... here. New file.
16102 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
16104 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
16105 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
16107 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16109 * gen-pass-instances.awk: Fix offset of substr().
16111 2013-07-31 David Malcolm <dmalcolm@redhat.com>
16113 * Makefile.in (pass-instances.def): New.
16114 (passes.o): Replace dependency on passes.def with one on
16117 * gen-pass-instances.awk: New.
16119 * passes.c (pass_manager::pass_manager): Use pass-instances.def
16120 rather than passes.def, updating local definition of NEXT_PASS
16121 macro to add an extra NUM parameter (currently unused).
16123 2013-07-30 David Malcolm <dmalcolm@redhat.com>
16125 * Makefile.in (PASS_MANAGER_H): New.
16126 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
16127 (passes.o): Likewise.
16128 (statistics.o): Likewise.
16129 (cgraphunit.o): Likewise.
16130 (context.o): Depend on PASS_MANAGER_H.
16132 * pass_manager.h: New.
16134 * cgraphunit.c (cgraph_add_new_function): Update for moves
16135 of globals to fields of pass_manager.
16136 (analyze_function): Likewise.
16137 (expand_function): Likewise.
16138 (ipa_passes): Likewise.
16139 (compile): Likewise.
16141 * context.c (context::context): New.
16142 * context.h (context::context): New.
16143 (context::get_passes): New.
16144 (context::passes_): New.
16146 * lto-cgraph.c (input_node): Update for moves of globals to
16147 fields of pass_manager.
16149 * passes.c (all_passes): Remove, in favor of a field of the
16150 same name within the new class pass_manager.
16151 (all_small_ipa_passes): Likewise.
16152 (all_lowering_passes): Likewise.
16153 (all_regular_ipa_passes): Likewise.
16154 (all_late_ipa_passes): Likewise.
16155 (all_lto_gen_passes): Likewise.
16156 (passes_by_id): Likewise.
16157 (passes_by_id_size): Likewise.
16158 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
16159 the new class pass_manager.
16160 (set_pass_for_id): Convert to...
16161 (pass_manager::set_pass_for_id): ...method.
16162 (get_pass_for_id): Convert to...
16163 (pass_manager::get_pass_for_id): ...method.
16164 (register_one_dump_file): Move body of implementation into...
16165 (pass_manager::register_one_dump_file): ...here.
16166 (register_dump_files_1): Convert to...
16167 (pass_manager::register_dump_files_1): ...method.
16168 (register_dump_files): Convert to...
16169 (pass_manager::register_dump_files): ...method.
16170 (create_pass_tab): Update for moves of globals to fields of
16172 (dump_passes): Move body of implementation into...
16173 (pass_manager::dump_passes): ...here.
16174 (register_pass): Move body of implementation into...
16175 (pass_manager::register_pass): ...here.
16176 (init_optimization_passes): Convert into...
16177 (pass_manager::pass_manager): ...constructor for new
16178 pass_manager class, and initialize the pass_lists array.
16179 (check_profile_consistency): Update for moves of globals to
16180 fields of pass_manager.
16181 (dump_profile_report): Move body of implementation into...
16182 (pass_manager::dump_profile_report): ...here.
16183 (ipa_write_summaries_1): Update for moves of pass lists from
16184 being globals to fields of pass_manager.
16185 (ipa_write_optimization_summaries): Likewise.
16186 (ipa_read_summaries): Likewise.
16187 (ipa_read_optimization_summaries): Likewise.
16188 (execute_all_ipa_stmt_fixups): Likewise.
16190 * statistics.c (statistics_fini): Update for moves of globals to
16191 fields of pass_manager.
16193 * toplev.c (general_init): Replace call to
16194 init_optimization_passes with construction of the pass_manager
16197 * tree-pass.h (all_passes): Remove, in favor of a field of the
16198 same name within the new class pass_manager.
16199 (all_small_ipa_passes): Likewise.
16200 (all_lowering_passes): Likewise.
16201 (all_regular_ipa_passes): Likewise.
16202 (all_lto_gen_passes): Likewise.
16203 (all_late_ipa_passes): Likewise.
16204 (passes_by_id): Likewise.
16205 (passes_by_id_size): Likewise.
16206 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
16207 the new class pass_manager.
16208 (get_pass_for_id): Remove.
16210 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
16212 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
16215 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
16217 * arm.md (mulhi3): New expand pattern.
16219 2013-07-30 Jan Hubicka <jh@suse.cz>
16220 Martin Liska <marxin.liska@gmail.com>
16222 * profile.c (compute_value_histograms): Do not ICE when
16223 there is mismatch only on some counters.
16225 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16227 PR rtl-optimization/57637
16228 * function.c (move_insn_for_shrink_wrap): Also check the
16229 GEN set of the LIVE problem for the liveness analysis
16230 if it exists, otherwise give up.
16232 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16234 PR tree-optimization/57993
16235 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
16236 replaced statement in the candidate table.
16237 (phi_add_costs): Return infinite cost when the hidden basis does
16238 not dominate all phis on which the candidate is dependent.
16239 (replace_one_candidate): Record replaced statement in the
16242 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
16244 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
16245 (ashlv2si3): New expander.
16246 (*ashlv2si3_i): New define_insn_and_split.
16247 * predicates.md (float_operation): Allow patterns with three
16248 basic sub-patterns.
16250 PR rtl-optimization/58021
16251 * mode-switching.c (create_pre_exit): Always split off preceding
16252 insns if we are not at the basic block head.
16254 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
16256 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
16257 (UCLIBC_DYNAMIC_LINKER): New macro.
16258 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
16260 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
16261 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
16263 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
16264 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
16265 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
16266 for SF and DF modes. Use ieee_quad_format for TF mode.
16267 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
16268 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
16269 (mips_option_override): Handle `-mnan=legacy'.
16270 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
16271 `-mabs=2008' and `-mnan=2008'.
16272 (OPTION_DEFAULT_SPECS): Add "nan" default.
16273 (ASM_SPEC): Handle `-mnan='.
16274 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
16275 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
16276 comment accordingly.
16277 (neg<mode>2): Likewise.
16278 * config/mips/mips.opt (mabs, mnan): New options.
16279 * doc/install.texi (Configuration): Document `--with-nan=' option.
16280 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
16282 (MIPS Options): Document them.
16283 * config.gcc <mips*-*-*>: Handle `--with-nan='.
16284 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
16285 * configure: Regenerate.
16286 * config.in: Regenerate.
16288 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
16290 * config/i386/i386.md (float post-reload splitters): Do not check
16291 for subregs of SSE registers.
16293 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
16294 H.J. Lu <hongjiu.lu@intel.com>
16298 * config/i386/i386.md (post-reload splitter
16299 to avoid partial SSE reg dependency stalls): New pattern.
16301 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
16303 * config/s390/s390.md ("movcc"): Swap load and store instructions.
16305 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
16307 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16308 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
16310 2013-07-26 Cary Coutant <ccoutant@google.com>
16312 * dwarf2out.c (die_checksum_ordered): Don't include template
16313 instantiations in signature.
16314 (is_template_parameter): New function.
16315 (is_template_instantiation): New function.
16316 (generate_skeleton_bottom_up): Don't include template instantiations
16318 (generate_skeleton): Likewise.
16319 (break_out_comdat_types): Move recursive call to break out nested
16321 (prune_unused_types_mark_generic_parms_dies): Call
16322 is_template_parameter.
16324 2013-07-26 Ian Bolton <ian.bolton@arm.com>
16326 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
16327 uses vector registers.
16328 * config/aarch64/iterators.md: Add attributes rtn and vas.
16330 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16331 Richard Earnshaw <richard.earnshaw@arm.com>
16333 * combine.c (simplify_comparison): Re-canonicalize operands
16335 * config/arm/arm.md (movcond_addsi): New splitter.
16337 2013-07-25 Sterling Augustine <saugustine@google.com>
16339 * dwarf2out.c (size_of_pubnames): Move code to...
16340 (include_pubname_in_output): ...here. New.
16341 (want_pubnames): Rearrange.
16342 (output_pubnames): Call include_pubname_in_output. Move assertion.
16344 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
16346 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
16348 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
16351 * doc/extend.texi: Remove obsolete builtins. Fix
16352 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
16354 2013-07-25 Jan Hubicka <jh@suse.cz>
16356 * cgraph.c (release_function_body): Break out from ...
16357 (cgraph_release_function_body): ... this one; also release DECL_RESULT
16358 and DECL_ARGUMENTS.
16359 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
16360 old_tree in the map.
16361 (create_specialized_node): Update.
16362 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
16364 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
16365 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
16366 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
16367 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
16370 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16372 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
16373 addsi3_carryin_alt2_<optab>): Correct output template.
16375 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16377 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
16378 Adjust for arm_restrict_it.
16379 Remove trailing whitespace.
16381 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
16383 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
16384 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
16386 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
16388 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
16390 PR rtl-optimization/57960
16391 * lra-constraints.c (process_alt_operands): Use the right mode
16392 when checking strict_low.
16394 2013-07-25 Jan Hubicka <jh@suse.cz>
16396 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
16397 * cgraph.c (cgraph_remove_node): Do not release function body
16398 when in cgraph streaming.
16399 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
16400 in other partitions are not considered reachable; fix handling of
16403 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16405 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
16407 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16412 * config/arm/arm.md ("*sibcall_insn): Replace use of
16413 Ss with US. Adjust output for v5 and v4t.
16414 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
16416 * config/arm/constraints.md ("Ss"): Rename to US.
16418 2013-07-25 Terry Guo <terry.guo@arm.com>
16420 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
16421 shift_add/shift_sub0/shift_sub1 RTXs.
16423 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
16424 Anton Blanchard <anton@au1.ibm.com>
16426 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
16427 (altivec_vpks<VI_char>ss): Likewise.
16428 (altivec_vpks<VI_char>us): Likewise.
16429 (altivec_vpku<VI_char>us): Likewise.
16430 (altivec_vpku<VI_char>um): Likewise.
16432 2013-07-24 David Malcolm <dmalcolm@redhat.com>
16434 Introduce context class.
16436 * Makefile.in (CONTEXT_H): New.
16437 (OBJS): Add context.o.
16438 (toplev.o): Add CONTEXT_H to dependencies.
16441 * toplev.c (general_init): Create the singleton gcc::context instance.
16447 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
16449 PR rtl-optimization/57968
16450 * mode-switching.c (create_pre_exit): Allow instructions that
16451 don't set a return register to need a non-exit mode.
16453 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16454 Anton Blanchard <anton@au1.ibm.com>
16456 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
16457 operands to vperm for little endian.
16458 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
16459 of lvsl to create the control mask for a vperm for little endian.
16461 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16462 Anton Blanchard <anton@au1.ibm.com>
16464 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
16465 two operands for little-endian.
16467 2013-07-23 Steve Ellcey <sellcey@mips.com>
16469 * config/mips/mips.c (mips_case_values_threshold): New.
16470 (TARGET_CASE_VALUES_THRESHOLD): Define.
16472 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16473 Anton Blanchard <anton@au1.ibm.com>
16475 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
16476 selection of field for vector splat in little endian mode.
16478 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
16480 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
16481 expanders to rs6000.md.
16482 (ior<mode>3): Likewise.
16483 (and<mode>3): Likewise.
16484 (one_cmpl<mode>2): Likewise.
16485 (nor<mode>3): Likewise.
16486 (andc<mode>3): Likewise.
16487 (eqv<mode>3): Likewise.
16488 (nand<mode>3): Likewise.
16489 (orc<mode>3): Likewise.
16491 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
16494 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
16495 to split multi-word logical operations.
16496 (rs6000_split_logical_di): Likewise.
16497 (rs6000_split_logical): Likewise.
16499 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
16500 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
16501 and allow TImode operations in 32-bit.
16502 (vsx_and<mode>3_64bit): Likewise.
16503 (vsx_ior<mode>3_32bit): Likewise.
16504 (vsx_ior<mode>3_64bit): Likewise.
16505 (vsx_xor<mode>3_32bit): Likewise.
16506 (vsx_xor<mode>3_64bit): Likewise.
16507 (vsx_one_cmpl<mode>2_32bit): Likewise.
16508 (vsx_one_cmpl<mode>2_64bit): Likewise.
16509 (vsx_nor<mode>3_32bit): Likewise.
16510 (vsx_nor<mode>3_64bit): Likewise.
16511 (vsx_andc<mode>3_32bit): Likewise.
16512 (vsx_andc<mode>3_64bit): Likewise.
16513 (vsx_eqv<mode>3_32bit): Likewise.
16514 (vsx_eqv<mode>3_64bit): Likewise.
16515 (vsx_nand<mode>3_32bit): Likewise.
16516 (vsx_nand<mode>3_64bit): Likewise.
16517 (vsx_orc<mode>3_32bit): Likewise.
16518 (vsx_orc<mode>3_64bit): Likewise.
16520 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
16521 logical types in GPRs.
16523 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
16524 logical insns to rs6000.md, and allow TImode operations in
16526 (altivec_ior<mode>3): Likewise.
16527 (altivec_xor<mode>3): Likewise.
16528 (altivec_one_cmpl<mode>2): Likewise.
16529 (altivec_nor<mode>3): Likewise.
16530 (altivec_andc<mode>3): Likewise.
16532 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
16533 attributes for moving the 128-bit logical operations into
16535 (BOOL_REGS_OUTPUT): Likewise.
16536 (BOOL_REGS_OP1): Likewise.
16537 (BOOL_REGS_OP2): Likewise.
16538 (BOOL_REGS_UNARY): Likewise.
16539 (BOOL_REGS_AND_CR0): Likewise.
16540 (one_cmpl<mode>2): Add support for DI logical operations on
16541 32-bit, splitting the operations to 32-bit.
16542 (anddi3): Likewise.
16543 (iordi3): Likewise.
16544 (xordi3): Likewise.
16545 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
16546 changes to combine the 32/64-bit code, allow logical operations on
16547 TI mode in 32-bit, and to use similar match_operator patterns like
16548 scalar mode uses. Combine the Altivec and VSX code for logical
16549 operations, and move it here.
16550 (ior<mode>3, 128-bit types): Likewise.
16551 (xor<mode>3, 128-bit types): Likewise.
16552 (one_cmpl<mode>3, 128-bit types): Likewise.
16553 (nor<mode>3, 128-bit types): Likewise.
16554 (andc<mode>3, 128-bit types): Likewise.
16555 (eqv<mode>3, 128-bit types): Likewise.
16556 (nand<mode>3, 128-bit types): Likewise.
16557 (orc<mode>3, 128-bit types): Likewise.
16558 (and<mode>3_internal): Likewise.
16559 (bool<mode>3_internal): Likewise.
16560 (boolc<mode>3_internal1): Likewise.
16561 (boolc<mode>3_internal2): Likewise.
16562 (boolcc<mode>3_internal1): Likewise.
16563 (boolcc<mode>3_internal2): Likewise.
16564 (eqv<mode>3_internal1): Likewise.
16565 (eqv<mode>3_internal2): Likewise.
16566 (one_cmpl1<mode>3_internal): Likewise.
16568 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
16570 * config/microblaze/microblaze.c (microblaze_expand_prologue):
16571 Rename flag_stack_usage to flag_stack_usage_info.
16573 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
16575 * config/microblaze/sync.md: New file.
16576 * config/microblaze/microblaze.md: Include sync.md
16577 * config/microblaze/microblaze.c: Add print_operand 'y'.
16578 * config/microblaze/constraints.md: Add memory_contraint
16579 'Q' which is a single register.
16581 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
16583 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
16585 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
16587 * reload.c (find_reloads): Exit loop once we find this operand
16588 cannot be reloaded somehow for this alternative.
16590 * reload.c (find_reloads): Exit loop once we find a hard register.
16592 * rtlanal.c (computed_jump_p): Exit loop once we find label
16595 * i386.c (ix86_pad_returns): Exit loop after setting replace.
16597 * cfgloopmanip.c (remove_path): Exit loop after setting
16600 * gensupport.c (subst_dup): Avoid loop if code is not
16601 MATCH_DUP nor MATCH_OP_DUP.
16603 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
16605 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
16607 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
16609 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
16610 true for SP_REGNUM if mode == ptr_mode.
16611 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
16612 with value R0_REGNUM + 31.
16614 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
16616 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
16617 pad pointer-typed argument downward.
16619 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
16621 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
16622 and __ILP32__ when the ILP32 model is in use.
16624 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
16626 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
16627 (aarch64_load_symref_appropriately): In the case of
16628 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
16629 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
16630 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
16631 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
16632 if 'mode' doesn't equal to 'ptr_mode'.
16633 (aarch64_output_mi_thunk): Add an assertion on the alignment of
16634 'vcall_offset'; change to call aarch64_emit_move differently depending
16635 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
16636 to calculate the upper bound of 'vcall_offset'.
16637 (aarch64_cannot_force_const_mem): Change to also return true if
16639 (aarch64_legitimize_reload_address): In the case of large
16640 displacements, add new local variable 'xmode' and an assertion
16641 based on it; change to use 'xmode' to generate the new rtx and
16643 (aarch64_asm_trampoline_template): Change to generate the template
16644 differently depending on TARGET_ILP32 or not; change to use
16645 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
16646 (aarch64_trampoline_size): Removed.
16647 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
16648 and replace immediate literals with it. Change to use 'ptr_mode'
16649 instead of 'DImode' and call convert_memory_address if the mode
16650 of 'fnaddr' doesn't equal to 'ptr_mode'.
16651 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
16653 (aarch64_elf_asm_destructor): Likewise.
16654 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
16655 on TARGET_ILP32 instead of aarch64_trampoline_size.
16656 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
16657 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
16658 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
16659 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
16660 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
16661 (storewb_pair<GPI:mode>_<P:mode>): ... this.
16662 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
16663 depending on the value of 'mode'.
16664 (add_losym_<mode>): New.
16665 (ldr_got_small_<mode>): New, based on ldr_got_small.
16666 (ldr_got_small): Remove.
16667 (ldr_got_small_sidi): New.
16668 * config/aarch64/iterators.md (P): New.
16669 (PTR): Change to 'ptr_mode' in the condition.
16671 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
16673 * config.gcc (aarch64*-*-*): Support --with-abi.
16674 (aarch64*-*-elf): Support --with-multilib-list.
16675 (aarch64*-*-linux*): Likewise.
16676 (supported_defaults): Add abi to aarch64*-*-*.
16677 * configure.ac: Mention AArch64 for --with-multilib-list.
16678 * configure: Re-generated.
16679 * config/aarch64/biarchilp32.h: New file.
16680 * config/aarch64/biarchlp64.h: New file.
16681 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
16683 (MULTILIB_DEFAULTS): Ditto.
16684 (DRIVER_SELF_SPECS): Ditto.
16685 (ASM_SPEC): Update to also substitute -mabi.
16686 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
16687 file whose name depends on -mabi= and -mbig-endian.
16688 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
16690 (POINTER_SIZE): New define.
16691 (POINTERS_EXTEND_UNSIGNED): Ditto.
16692 (enum aarch64_abi_type): New enumeration tag.
16693 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
16694 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
16695 (TARGET_ILP32): New define.
16696 * config/aarch64/aarch64.opt (mabi): New.
16697 (aarch64_abi): New.
16698 (ilp32, lp64): New values for -mabi.
16699 * config/aarch64/t-aarch64 (comma): New define.
16700 (MULTILIB_OPTIONS): Ditto.
16701 (MULTILIB_DIRNAMES): Ditto.
16702 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
16703 * doc/invoke.texi: Document -mabi for AArch64.
16705 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
16707 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
16709 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
16710 Anton Blanchard <anton@au1.ibm.com>
16712 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
16713 endianness when selecting field to splat.
16715 2013-07-22 Eric Christopher <echristo@gmail.com>
16717 * dwarf2out.c (die_odr_checksum): New function to use
16718 CHECKSUM_ macros and ULEB128 for DIE tag.
16719 (generate_type_signature): Use.
16721 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
16723 * config.gcc (sparc*-*-*): Accept leon3 processor.
16724 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
16725 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
16726 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
16727 * config/sparc/sparc.opt (enum processor_type): Add leon3.
16728 (mfix-ut699): Adjust comment.
16729 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
16730 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
16731 (CPP_CPU_SPEC): Likewise.
16732 (ASM_CPU_SPEC): Likewise.
16733 * config/sparc/sparc.c (leon3_cost): New constant.
16734 (sparc_option_override): Add leon3 support.
16735 (mem_ref): New function.
16736 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
16737 (sparc_do_work_around_errata): Look into the instruction in the delay
16738 slot and adjust accordingly. Add fix for the data cache nullify issues
16739 of the UT699. Change insertion position for the NOP.
16740 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
16741 (leon3_load): New reservation.
16742 (leon_store): Bump latency to 2.
16743 (grfpu): New automaton.
16744 (grfpu_alu): New unit.
16745 (grfpu_ds): Likewise.
16746 (leon_fp_alu): Adjust.
16747 (leon_fp_mult): Delete.
16748 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
16749 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
16750 * config/sparc/sparc.md (cpu): Add leon3.
16751 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
16752 (swapsi): Likewise.
16753 (atomic_test_and_set): Likewise.
16754 (ldstub): Likewise.
16756 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
16758 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
16759 default for R5900 targets.
16760 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
16761 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
16762 * config/mips/mips.c (mips_option_override): Report an error for
16763 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
16764 for -march=r5900 -mhard-float.
16766 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
16768 * df-problems.c (can_move_insns_across): Exit loop once we
16769 find a non-fixed, non-global register.
16771 * ipa-pure-const.c (propagate_nothrow): Exit loop after
16774 * omega.c (omega_eliminate_red): Break after setting red_found.
16775 (omega_problem_has_red_equations): Similarly after setting found.
16776 (omega_query_variable): Similarly after setting coupled.
16778 2013-07-22 Marek Polacek <polacek@redhat.com>
16780 * gimplify.c: Don't include gimple.h twice.
16782 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16784 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
16785 instead of TARGET_THUMB1.
16786 (Pz): New constraint.
16787 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
16789 (compare_negsi_si): Likewise.
16790 (compare_addsi2_op0): Likewise.
16791 (compare_addsi2_op1): Likewise.
16792 (addsi3_carryin_<optab>): Likewise.
16793 (addsi3_carryin_alt2_<optab>): Likewise.
16794 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
16795 for arm_restrict_it.
16796 (subsi3_carryin): Likewise.
16797 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
16798 (minmax_arithsi): Disable for arm_restrict_it.
16799 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
16800 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
16801 (satsi_<SAT:code>_shift): Likewise.
16802 (arm_shiftsi3): Add alternative for 16-bit encoding.
16803 (arm32_movhf): Disable for arm_restrict_it.
16804 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
16805 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
16807 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
16809 * config/arm/arm.md (attribute "insn"): Delete.
16810 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
16811 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
16812 (not_shiftsi): Update for attribute change.
16813 (not_shiftsi_compare0): Likewise.
16814 (not_shiftsi_compare0_scratch): Likewise.
16815 (arm_one_cmplsi2): Likewise.
16816 (thumb1_one_cmplsi2): Likewise.
16817 (notsi_compare0): Likewise.
16818 (notsi_compare0_scratch): Likewise.
16819 (thumb1_movdi_insn): Likewise.
16820 (arm_movsi_insn): Likewise.
16821 (movhi_insn_arch4): Likewise.
16822 (movhi_bytes): Likewise.
16823 (arm_movqi_insn): Likewise.
16824 (thumb1_movqi_insn): Likewise.
16825 (arm32_movhf): Likewise.
16826 (thumb1_movhf): Likewise.
16827 (arm_movsf_soft_insn): Likewise.
16828 (thumb1_movsf_insn): Likewise.
16829 (thumb_movdf_insn): Likewise.
16830 (movsicc_insn): Likewise.
16831 (movsfcc_soft_insn): Likewise.
16832 (and_scc): Likewise.
16833 (cond_move): Likewise.
16834 (if_move_not): Likewise.
16835 (if_not_move): Likewise.
16836 (if_shift_move): Likewise.
16837 (if_move_shift): Likewise.
16838 (if_shift_shift): Likewise.
16839 (if_not_arith): Likewise.
16840 (if_arith_not): Likewise.
16841 (cond_move_not): Likewise.
16842 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
16843 (neon_mov<mode>): Likewise.
16844 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
16845 (thumb2_movsi_vfp): Likewise.
16846 (movsf_vfp): Likewise.
16847 (thumb2_movsf_vfp): Likewise.
16848 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
16850 (cortexa7_older_only): Likewise.
16851 (cortexa7_younger): Likewise.
16852 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
16853 (1020alu_shift_op): Likewise.
16854 (1020alu_shift_reg_op): Likewise.
16855 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
16856 (alu_shift_op): Likewise.
16857 (alu_shift_reg_op): Likewise.
16858 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
16859 (11_alu_shift_op): Likewise.
16860 (11_alu_shift_reg_op): Likewise.
16861 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
16862 (9_alu_shift_reg_op): Likewise.
16863 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
16865 (cortex_a15_alu_shift): Likewise.
16866 (cortex_a15_alu_shift_reg): Likewise.
16867 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
16868 (cortex_a5_alu_shift): Likewise.
16869 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
16871 (cortex_a53_alu_shift): Likewise.
16872 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
16874 (cortex_a7_alu_reg): Likewise.
16875 (cortex_a7_alu_shift): Likewise.
16876 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
16877 (cortex_a8_alu_shift): Likewise.
16878 (cortex_a8_alu_shift_reg): Likewise.
16879 (cortex_a8_mov): Likewise.
16880 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
16881 (cortex_a9_dp_shift): Likewise.
16882 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
16883 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
16884 (cortex_r4_mov): Likewise.
16885 (cortex_r4_alu_shift): Likewise.
16886 (cortex_r4_alu_shift_reg): Likewise.
16887 * config/arm/fa526.md (526_alu_op): Update for attribute change.
16888 (526_alu_shift_op): Likewise.
16889 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
16890 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
16891 (626te_alu_shift_op): Likewise.
16892 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
16893 (726te_alu_op): Likewise.
16894 (726te_alu_shift_op): Likewise.
16895 (726te_alu_shift_reg_op): Likewise.
16896 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
16897 (mp626_alu_shift_op): Likewise.
16898 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
16899 (pj4_alu_e1_conds): Likewise.
16900 (pj4_alu): Likewise.
16901 (pj4_alu_conds): Likewise.
16902 (pj4_shift): Likewise.
16903 (pj4_shift_conds): Likewise.
16904 (pj4_alu_shift): Likewise.
16905 (pj4_alu_shift_conds): Likewise.
16907 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16909 * config/arm/predicates.md (shiftable_operator_strict_it):
16911 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
16912 Disable cond_exec version for arm_restrict_it.
16913 (thumb2_smaxsi3): Convert to generate cond_exec.
16914 (thumb2_sminsi3): Likewise.
16915 (thumb32_umaxsi3): Likewise.
16916 (thumb2_uminsi3): Likewise.
16917 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
16918 (thumb2_neg_abssi2): Likewise.
16919 (thumb2_mov_scc): Add alternative for 16-bit encoding.
16920 (thumb2_movsicc_insn): Adjust alternatives.
16921 (thumb2_mov_negscc): Disable for arm_restrict_it.
16922 (thumb2_mov_negscc_strict_it): New pattern.
16923 (thumb2_mov_notscc_strict_it): New pattern.
16924 (thumb2_mov_notscc): Disable for arm_restrict_it.
16925 (thumb2_ior_scc): Likewise.
16926 (thumb2_ior_scc_strict_it): New pattern.
16927 (thumb2_cond_move): Adjust for arm_restrict_it.
16928 (thumb2_cond_arith): Disable for arm_restrict_it.
16929 (thumb2_cond_arith_strict_it): New pattern.
16930 (thumb2_cond_sub): Adjust for arm_restrict_it.
16931 (thumb2_movcond): Likewise.
16932 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
16933 (thumb2_zero_extendhisi2_v6): Likewise.
16934 (thumb2_zero_extendqisi2_v6): Likewise.
16935 (orsi_notsi_si): Likewise.
16936 (orsi_not_shiftsi_si): Likewise.
16938 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
16940 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
16941 instruction sequence is 1 byte shorter.
16943 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
16945 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
16946 it is not needed after split.
16948 2013-07-20 Iain Sandoe <iain@codesourcery.com>
16951 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
16952 second label for nonlocal goto receivers. Don't output pic base labels
16953 unless we're producing PIC; mark that action unreachable().
16954 (ix86_save_reg): If the function contains a nonlocal label, save the
16956 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
16957 * config/darwin.c (emitted_pic_label_num): New GTY.
16958 (update_pic_label_number_if_needed): New.
16959 (machopic_output_function_base_name): Adjust for nonlocal receiver
16961 (machopic_should_output_picbase_label): New.
16962 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
16963 (nonlocal_goto_receiver): New insn and split.
16965 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
16967 * config/aarch64/aarch64-builtins.c
16968 (aarch64_fold_builtin): Fold abs in all modes.
16969 * config/aarch64/aarch64-simd-builtins.def
16970 (abs): Enable for all modes.
16971 * config/aarch64/arm_neon.h
16972 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
16973 (vabs_f64): Add missing intrinsic.
16975 2013-07-19 Ian Bolton <ian.bolton@arm.com>
16977 * config/aarch64/arm_neon.h (vabs_s64): New function
16979 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
16982 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
16983 * config/avr/avr.md (adjust_len): Add `round'.
16984 * config/avr/avr-protos.h (avr_out_round): New prototype.
16985 (avr_out_plus): Add `out_label' argument.
16986 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
16987 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
16988 Handle the case where `insn' is just a pattern.
16989 (avr_out_bitop): Handle the case where `insn' is just a pattern.
16990 (avr_out_round): New function.
16991 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
16993 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
16995 * config/microblaze/microblaze.c (microblaze_expand_prologue):
16996 Add check for flag_stack_usage to handle -fstack-usage support
16998 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
17000 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
17001 interaction for new Power8 flags and VSX.
17003 2013-07-18 Sriraman Tallam <tmsriram@google.com>
17005 PR middle-end/57698
17006 * tree-inline.c (expand_call_inline): Emit errors during
17007 early_inlining only if optimization is not turned on.
17009 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17013 * passes.c (init_optimization_passes): Move the construction of
17014 the pass hierarchy into a new passes.def file.
17016 * Makefile.in (passes.o): Add dependency on passes.def.
17018 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17020 * passes.c (init_optimization_passes): Introduce macros for
17021 constructing the tree of passes (INSERT_PASSES_AFTER,
17022 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
17023 TERMINATE_PASS_LIST).
17025 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
17026 Wei Mi <wmi@google.com>
17028 PR rtl-optimization/57878
17029 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
17031 (reload_pseudo_compare_func): Check nregs first for reload
17034 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17036 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
17038 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
17040 * read-rtl.c (validate_const_int): Once an invalid character is
17041 seen, quit the loop.
17043 * gengtype.c (write_roots): Similarly once we find the "deletable"
17044 or "if_marked" option.
17046 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
17048 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
17049 "xtab" and "sat". Move value "clz" from here to ...
17050 (attriubte "type"): ... here.
17051 (satsi_<SAT:code>): Delete "insn" attribute.
17052 (satsi_<SAT:code>_shift): Likewise.
17053 (arm_zero_extendqisi2addsi): Likewise.
17054 (arm_extendqisi2addsi): Likewise.
17055 (clzsi2): Update for attribute changes.
17056 (rbitsi2): Likewise.
17057 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
17059 (arm_usatsihi): Likewise.
17060 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
17062 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
17064 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
17065 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
17066 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
17067 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
17068 in alphabetical order.
17069 (attribute "core_cycles"): Update for attribute changes.
17070 (arm_addsi3): Likewise.
17071 (addsi3_compare0): Likewise.
17072 (addsi3_compare0_scratch): Likewise.
17073 (addsi3_compare_op1): Likewise.
17074 (addsi3_compare_op2): Likewise.
17075 (compare_addsi2_op0): Likewise.
17076 (compare_addsi2_op1): Likewise.
17077 (addsi3_carryin_shift_<optab>): Likewise.
17078 (subsi3_carryin_shift): Likewise.
17079 (rsbsi3_carryin_shift): Likewise.
17080 (arm_subsi3_insn): Likewise.
17081 (subsi3_compare0): Likewise.
17082 (subsi3_compare): Likewise.
17083 (arm_andsi3_insn): Likewise.
17084 (thumb1_andsi3_insn): Likewise.
17085 (andsi3_compare0): Likewise.
17086 (andsi3_compare0_scratch): Likewise.
17087 (zeroextractsi_compare0_scratch
17088 (andsi_not_shiftsi_si): Likewise.
17089 (iorsi3_insn): Likewise.
17090 (iorsi3_compare0): Likewise.
17091 (iorsi3_compare0_scratch): Likewise.
17092 (arm_xorsi3): Likewise.
17093 (thumb1_xorsi3_insn): Likewise.
17094 (xorsi3_compare0): Likewise.
17095 (xorsi3_compare0_scratch): Likewise.
17096 (satsi_<SAT:code>_shift): Likewise.
17098 (arm_shiftsi3): Likewise.
17099 (shiftsi3_compare0): Likewise.
17100 (not_shiftsi): Likewise.
17101 (not_shiftsi_compare0): Likewise.
17102 (not_shiftsi_compare0_scratch): Likewise.
17103 (arm_one_cmplsi2): Likewise.
17104 (thumb_one_complsi2): Likewise.
17105 (notsi_compare0): Likewise.
17106 (notsi_compare0_scratch): Likewise.
17107 (thumb1_zero_extendhisi2): Likewise.
17108 (arm_zero_extendhisi2): Likewise.
17109 (arm_zero_extendhisi2_v6): Likewise.
17110 (arm_zero_extendhisi2addsi): Likewise.
17111 (thumb1_zero_extendqisi2): Likewise.
17112 (thumb1_zero_extendqisi2_v6): Likewise.
17113 (arm_zero_extendqisi2): Likewise.
17114 (arm_zero_extendqisi2_v6): Likewise.
17115 (arm_zero_extendqisi2addsi): Likewise.
17116 (thumb1_extendhisi2): Likewise.
17117 (arm_extendhisi2): Likewise.
17118 (arm_extendhisi2_v6): Likewise.
17119 (arm_extendqisi): Likewise.
17120 (arm_extendqisi_v6): Likewise.
17121 (arm_extendqisi2addsi): Likewise.
17122 (thumb1_extendqisi2): Likewise.
17123 (thumb1_movdi_insn): Likewise.
17124 (arm_movsi_insn): Likewise.
17125 (movsi_compare0): Likewise.
17126 (movhi_insn_arch4): Likewise.
17127 (movhi_bytes): Likewise.
17128 (arm_movqi_insn): Likewise.
17129 (thumb1_movqi_insn): Likewise.
17130 (arm32_movhf): Likewise.
17131 (thumb1_movhf): Likewise.
17132 (arm_movsf_soft_insn): Likewise.
17133 (thumb1_movsf_insn): Likewise.
17134 (movdf_soft_insn): Likewise.
17135 (thumb_movdf_insn): Likewise.
17136 (arm_cmpsi_insn): Likewise.
17137 (cmpsi_shiftsi): Likewise.
17138 (cmpsi_shiftsi_swp): Likewise.
17139 (arm_cmpsi_negshiftsi_si): Likewise.
17140 (movsicc_insn): Likewise.
17141 (movsfcc_soft_insn): Likewise.
17142 (arith_shiftsi): Likewise.
17143 (arith_shiftsi_compare0
17144 (arith_shiftsi_compare0_scratch
17145 (sub_shiftsi): Likewise.
17146 (sub_shiftsi_compare0
17147 (sub_shiftsi_compare0_scratch
17148 (and_scc): Likewise.
17149 (cond_move): Likewise.
17150 (if_plus_move): Likewise.
17151 (if_move_plus): Likewise.
17152 (if_move_not): Likewise.
17153 (if_not_move): Likewise.
17154 (if_shift_move): Likewise.
17155 (if_move_shift): Likewise.
17156 (if_shift_shift): Likewise.
17157 (if_not_arith): Likewise.
17158 (if_arith_not): Likewise.
17159 (cond_move_not): Likewise.
17160 (thumb1_ashlsi3): Set type attribute.
17161 (thumb1_ashrsi3): Likewise.
17162 (thumb1_lshrsi3): Likewise.
17163 (thumb1_rotrsi3): Likewise.
17164 (shiftsi3_compare0_scratch): Likewise.
17165 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
17166 (neon_mov<mode>): Likewise.
17167 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
17169 (thumb2_movsi_insn): Likewise.
17170 (thumb2_cmpsi_neg_shiftsi): Likewise.
17171 (thumb2_extendqisi_v6): Likewise.
17172 (thumb2_zero_extendhisi2_v6): Likewise.
17173 (thumb2_zero_extendqisi2_v6): Likewise.
17174 (thumb2_shiftsi3_short): Likewise.
17175 (thumb2_addsi3_compare0_scratch): Likewise.
17176 (orsi_not_shiftsi_si): Likewise.
17177 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
17178 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
17180 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
17181 (1020alu_shift_op): Likewise.
17182 (1020alu_shift_reg_op): Likewise.
17183 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
17184 (alu_shift_op): Likewise.
17185 (alu_shift_reg_op): Likewise.
17186 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
17187 (11_alu_shift_op): Likewise.
17188 (11_alu_shift_reg_op): Likewise.
17189 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
17190 (9_alu_shift_reg_op): Likewise.
17191 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
17193 (cortex_a15_alu_shift): Likewise.
17194 (cortex_a15_alu_shift_reg): Likewise.
17195 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
17197 (cortex_a5_alu_shift): Likewise.
17198 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
17200 (cortex_a53_alu_shift): Likewise.
17201 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
17203 (cortex_a7_alu_reg): Likewise.
17204 (cortex_a7_alu_shift): Likewise.
17205 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
17207 (cortex_a8_alu_shift): Likewise.
17208 (cortex_a8_alu_shift_reg): Likewise.
17209 (cortex_a8_mov): Likewise.
17210 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
17211 (cortex_a9_dp_shift): Likewise.
17212 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
17214 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
17216 (cortex_r4_mov): Likewise.
17217 (cortex_r4_alu_shift): Likewise.
17218 (cortex_r4_alu_shift_reg): Likewise.
17219 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
17220 (526_alu_shift_op): Likewise.
17221 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
17222 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
17223 (626te_alu_shift_op): Likewise.
17224 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
17225 (726te_alu_op): Likewise.
17226 (726te_alu_shift_op): Likewise.
17227 (726te_alu_shift_reg_op): Likewise.
17228 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
17229 (mp626_alu_shift_op): Likewise.
17230 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
17231 (pj4_alu_e1_conds): Likewise.
17232 (pj4_alu): Likewise.
17233 (pj4_alu_conds): Likewise.
17234 (pj4_shift): Likewise.
17235 (pj4_shift_conds): Likewise.
17236 (pj4_alu_shift): Likewise.
17237 (pj4_alu_shift_conds): Likewise.
17238 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
17240 (cortexa7_older_only): Likewise.
17241 (cortexa7_younger): Likewise.
17243 2013-07-18 David Malcolm <dmalcolm@redhat.com>
17245 * ipa-pure-const.c (generate_summary): Rename to...
17246 (pure_const_generate_summary): ... this.
17248 2013-07-17 Iain Sandoe <iain@codesourcery.com>
17250 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
17252 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
17255 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
17258 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17260 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
17261 enabled without -march=zEC12.
17262 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
17265 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17267 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
17268 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
17269 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
17270 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
17271 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
17272 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
17273 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
17275 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
17276 and ISA_HAS_NMADD3_NMSUB3.
17277 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
17278 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
17279 (nmsub4<mode>, nmsub3<mode>): Likewise.
17280 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
17282 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17284 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
17285 TARGET_MIPS5400 checking.
17287 2013-07-16 Jakub Jelinek <jakub@redhat.com>
17288 Peter Bergner <bergner@vnet.ibm.com>
17290 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
17291 registers in the comment.
17292 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
17293 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
17294 rather than FIRST_PSEUDO_REGISTERS.
17296 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
17298 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
17299 enable extra ISA flags with TARGET_HTM.
17301 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
17303 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
17306 2013-07-15 Cong Hou <congh@google.com>
17308 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
17309 in compare function for sorting.
17311 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
17313 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
17314 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
17315 * config/rs6000/rs6000.opt: Add -mhtm option.
17316 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
17317 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
17318 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
17319 __HTM__ if the HTM instructions are available.
17320 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
17321 htm_spr_reg_operand): New define_predicates.
17322 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
17323 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
17325 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
17326 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
17327 HTM builtin functions.
17328 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
17329 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
17330 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
17331 (rs6000_builtin_mask_calculate): Likewise.
17332 (rs6000_option_override_internal): Likewise.
17333 (bdesc_htm): Add new HTM builtin support.
17334 (htm_spr_num): New function.
17335 (htm_spr_regno): Likewise.
17336 (rs6000_htm_spr_icode): Likewise.
17337 (htm_expand_builtin): Likewise.
17338 (htm_init_builtins): Likewise.
17339 (rs6000_expand_builtin): Add support for HTM builtin functions.
17340 (rs6000_init_builtins): Likewise.
17341 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
17343 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
17344 (TARGET_HTM, MASK_HTM): Define macros.
17345 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
17346 (FIXED_REGISTERS): Likewise.
17347 (CALL_USED_REGISTERS): Likewise.
17348 (CALL_REALLY_USED_REGISTERS): Likewise.
17349 (REG_ALLOC_ORDER): Likewise.
17350 (enum reg_class): Likewise.
17351 (REG_CLASS_NAMES): Likewise.
17352 (REG_CLASS_CONTENTS): Likewise.
17353 (REGISTER_NAMES): Likewise.
17354 (ADDITIONAL_REGISTER_NAMES): Likewise.
17355 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
17356 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
17357 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
17358 * config/rs6000/htm.md: New file.
17359 * config/rs6000/htmintrin.h: New file.
17360 * config/rs6000/htmxlintrin.h: New file.
17362 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
17364 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
17365 Define SYMBOL_TINY_GOT, update comment.
17366 * config/aarch64/aarch64.c
17367 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
17368 (aarch64_expand_mov_immediate): Likewise.
17369 (aarch64_print_operand): Likewise.
17370 (aarch64_classify_symbol): Likewise.
17371 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
17372 (ldr_got_tiny): Define.
17374 2013-07-13 Tobias Grosser <tobias@grosser.es>
17376 PR tree-optimization/54094
17377 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
17378 scheduling dimension for the parallelism check from the polyhedral
17379 information in the AST.
17380 * graphite-dependences.c (carries_deps): Do not assume the schedule is
17383 2013-07-13 Jason Merrill <jason@redhat.com>
17385 * print-tree.c (debug_vec_tree): Use debug_raw.
17386 (debug_raw (vec<tree, va_gc> &)): New.
17387 (debug_raw (vec<tree, va_gc> *)): New.
17388 * tree.h: Declare them.
17390 2013-07-13 Bin Cheng <bin.cheng@arm.com>
17392 * ifcvt.c (ifcvt_after_combine): New static variable.
17393 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
17395 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
17396 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
17397 rest_of_handle_if_after_reload): Pass new argument for if_convert.
17399 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
17401 * config/mips/mips.c (mips_expand_call): Remove empty statement.
17403 2013-07-12 Michael Matz <matz@suse.de>
17405 PR middle-end/55771
17406 * convert.c (convert_to_real): Reject non-float inner types.
17408 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
17410 * config/aarch64/aarch64-protos.h
17411 (aarch64_simd_immediate_valid_for_move): Remove.
17412 * config/aarch64/aarch64.c (simd_immediate_info): New member.
17413 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
17415 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
17417 2013-07-11 Steve Ellcey <sellcey@mips.com>
17419 * config/mips/mips.c (mips_conditional_register_usage): Do not
17420 use t[0-7] registers in MIPS16 mode when optimizing for size.
17422 2013-07-11 Sriraman Tallam <tmsriram@google.com>
17424 * config/i386/i386.c (dispatch_function_versions): Fix array
17425 indexing of function_version_info to match actual_versions.
17427 2013-07-11 Teresa Johnson <tejohnson@google.com>
17429 * vec.h (struct va_gc): Move release out-of-line.
17430 (va_gc::release): Call ggc_free on released vec.
17432 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17434 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
17435 Require GOT register as additional operand in UNSPEC.
17436 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
17437 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
17438 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
17439 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
17440 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
17441 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
17442 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
17444 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
17447 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
17448 name seen by assembler/linker rather if available.
17450 2013-07-11 Andreas Schwab <schwab@suse.de>
17452 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
17454 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
17456 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
17458 2013-07-10 Joseph Myers <joseph@codesourcery.com>
17460 * doc/tm.texi.in: Move hook documentation to ....
17461 * target.def: ... here.
17463 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
17464 text on @hook line.
17465 * doc/tm.texi: Regenerate.
17467 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
17470 * doc/invoke.texi: Document -Wconditionally-supported.
17472 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
17475 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
17478 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
17481 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
17482 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
17483 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
17484 Remove duplicate devices.
17485 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
17486 * config/avr/t-multilib: Regenerate.
17487 * config/avr/avr-tables.opt: Regenerate.
17488 * doc/avr-mmcu.texi: Regenerate.
17490 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
17493 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
17495 2013-07-10 Graham Stott <graham.stott@btinternet.com>
17497 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
17498 the cost of MULT when optimizing for size.
17500 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17502 * config/cr16/cr16-protos.h: Don't include target.h.
17504 2013-07-09 Joseph Myers <joseph@codesourcery.com>
17506 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
17507 adjust register size for TDmode and TFmode for VSX registers.
17509 2013-07-08 Kai Tietz <ktietz@redhat.com>
17512 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
17513 hook_bool_const_tree_true.
17515 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17517 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
17518 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
17519 * config/s390/s390.md: Define FPR*_REGNUM constants.
17520 Fix FPR2_REGNUM constant (18 -> 17).
17521 ("*trunc<BFP:mode><DFP_ALL:mode>2")
17522 ("*trunc<DFP_ALL:mode><BFP:mode>2")
17523 ("trunc<BFP:mode><DFP_ALL:mode>2")
17524 ("trunc<DFP_ALL:mode><BFP:mode>2")
17525 ("*extend<BFP:mode><DFP_ALL:mode>2")
17526 ("*extend<DFP_ALL:mode><BFP:mode>2")
17527 ("extend<BFP:mode><DFP_ALL:mode>2")
17528 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
17531 2013-07-08 Graham Stott <graham.stott@btinternet.com>
17533 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
17535 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17537 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
17538 and cfun_fpr_bit_p to cfun_fpr_save_p.
17539 (s390_frame_area, s390_register_info, s390_frame_info)
17540 (s390_emit_prologue, s390_emit_epilogue)
17541 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
17543 * config/s390/s390.h: Define *_REGNUM macros for floating point
17546 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
17548 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
17550 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
17552 PR rtl-optimization/57786
17553 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
17554 and break out of the loop when it is set to false.
17556 2013-07-08 Jakub Jelinek <jakub@redhat.com>
17559 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
17560 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
17561 (const_int 63)) 0)).
17562 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
17563 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
17564 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
17566 PR rtl-optimization/57829
17567 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
17568 mask bits outside of mode are just sign-extension from mode to HWI.
17570 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
17572 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
17573 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
17574 adjust_address instead of change_address to keep info about alignment.
17575 (emit_strmov): Remove.
17576 (emit_memmov): New function.
17577 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
17578 (expand_movmem_epilogue): Likewise and return updated rtx for
17580 (expand_constant_movmem_prologue): Likewise and return updated rtx for
17581 destination and source.
17582 (decide_alignment): Refactor, handle vector_loop.
17583 (ix86_expand_movmem): Likewise.
17584 (ix86_expand_setmem): Likewise.
17585 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
17587 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
17589 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
17590 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
17592 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
17594 * config/i386/sse.md (sse_movlhps): Change alternative 3
17595 of operand 2 to "m".
17597 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
17600 * config/i386/sse.md (iptr): New mode attribute.
17601 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
17602 (<sse>_vm<plusminus_insn><mode>3): Ditto.
17603 (<sse>_vmmul<mode>3): Ditto.
17604 (<sse>_vmdiv<mode>3): Ditto.
17605 (sse_vmrcpv4sf2): Ditto.
17606 (<sse>_vmsqrt<mode>2): Ditto.
17607 (sse_vmrsqrtv4sf2): Ditto.
17608 (<sse>_vm<code><mode>3): Ditto.
17609 (avx_vmcmp<mode>3): Ditto.
17610 (<sse>_vmmaskcmp<mode>3): Ditto.
17611 (<sse>_comi): Ditto.
17612 (<sse>_ucomi): Ditto.
17613 (*xop_vmfrcz_<mode>): Ditto.
17614 (*fmai_fmadd_<mode>): Ditto.
17615 (*fmai_fmsub_<mode>): Ditto.
17616 (*fmai_fnmadd_<mode>): Ditto.
17617 (*fmai_fnmsub_<mode>): Ditto.
17618 (*fma4i_vmfmadd_<mode>): Ditto.
17619 (*fma4i_vmfmsub_<mode>): Ditto.
17620 (*fma4i_vmfnmadd_<mode>): Ditto.
17621 (*fma4i_vmfnmsub_<mode>): Ditto.
17622 (*xop_vmfrcz_<mode>): Ditto.
17623 (sse_cvtps2pi): Ditto.
17624 (sse_cvttps2pi): Ditto.
17625 (sse_cvtss2si): Ditto.
17626 (sse_cvtss2si_2): Ditto.
17627 (sse_cvtss2siq_2): Ditto.
17628 (sse_cvttss2si): Ditto.
17629 (sse_cvttss2siq): Ditto.
17630 (sse_cvtsd2si): Ditto.
17631 (sse_cvtsd2si_2): Ditto.
17632 (sse_cvtsd2siq_2): Ditto.
17633 (sse_cvttsd2si): Ditto.
17634 (sse_cvttsd2siq): Ditto.
17635 (sse_cvtsd2ss): Ditto.
17636 (sse_cvtss2sd): Ditto.
17637 (avx2_pbroadcast<mode>): Ditto.
17638 (avx2_pbroadcast<mode>_1): Ditto.
17639 (*avx_vperm_broadcast_v4sf): Ditto.
17641 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
17642 (sse_movlhps): Ditto.
17643 (sse_storehps): Ditto.
17644 (sse_loadhps): Ditto.
17645 (sse_storelps): Ditto.
17646 (sse_loadlps): Ditto.
17647 (*vec_concatv4sf): Ditto.
17648 (*vec_interleave_highv2df): Ditto.
17649 (*vec_interleave_lowv2df): Ditto.
17650 (*vec_extractv2df_1_sse): Ditto.
17651 (*vec_extractv2df_0_sse): Ditto.
17652 (sse2_storelpd): Ditto.
17653 (sse2_loadlpd): Ditto.
17654 (sse2_movsd): Ditto.
17655 (*vec_concatv4si): Ditto.
17656 (vec_concatv2di): Ditto.
17658 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
17659 for Intel asm dialect.
17660 (mmx_punpcklwd): Ditto.
17661 (mmx_punpckldq): Ditto.
17663 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
17664 for intel assembler dialect.
17666 2013-07-06 Jakub Jelinek <jakub@redhat.com>
17669 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
17670 for BUILT_IN_C{LZ,LRSB}*.
17671 * tree.h (CASE_INT_FN): Add FN##IMAX case.
17672 * tree-vrp.c (extract_range_basic): Handle
17673 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
17674 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
17675 fall thru to code calling set_value*.
17676 * builtins.c (expand_builtin): Remove *IMAX cases.
17677 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
17678 if width is bigger than 2*HWI.
17680 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
17682 PR rtl-optimization/55342
17683 * lra-int.h (lra_subreg_reload_pseudos): New.
17684 * lra.c: Add undoing optional reloads to the block diagram.
17685 (lra_subreg_reload_pseudos): New.
17686 (lra_optional_reload_pseudos): Change comments.
17687 (lra): Init and clear lra_subreg_reload_pseudos. Clear
17688 lra_optional_reload_pseudos after undo transformations.
17689 * lra-assigns.c (pseudo_prefix_title): New.
17690 (lra_setup_reg_renumber): Use it.
17691 (spill_for): Ditto. Check subreg reload pseudos too.
17692 (assign_by_spills): Consider subreg reload pseudos too.
17693 * lra-constraints.c (simplify_operand_subreg): Use
17694 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
17695 (curr_insn_transform): Recognize and do optional reloads.
17696 (undo_optional_reloads): New.
17697 (lra_undo_inheritance): Call undo_optional_reloads.
17699 2013-07-05 Thomas Quinot <quinot@adacore.com>
17701 * tree-complex.c (expand_complex_operations_1): Fix typo.
17703 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
17705 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
17706 (tune_params): New member 'const vec_costs'.
17707 * config/aarch64/aarch64.c (generic_vector_cost): New.
17708 (generic_tunings): New member 'generic_vector_cost'.
17709 (aarch64_builtin_vectorization_cost): New.
17710 (aarch64_add_stmt_cost): New.
17711 (TARGET_VECTORIZE_ADD_STMT_COST): New.
17712 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
17714 2013-07-03 Jakub Jelinek <jakub@redhat.com>
17717 * config/i386/predicates.md (vsib_address_operand): Disallow
17718 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
17720 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
17722 PR middle-end/55030
17723 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
17724 expand_builtin_setjmp_receiver.
17725 (expand_label): Adjust, call expand_builtin_setjmp_receiver
17726 with NULL for the label parameter.
17727 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
17728 the frame-pointer. Adjust comments.
17729 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
17730 only if LABEL is non-NULL.
17732 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
17734 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
17735 (ARM_ABI_AAPCS64): Ditto.
17737 (ARM_DEFAULT_ABI): Ditto.
17739 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
17741 * config/aarch64/aarch64-builtins.c
17742 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
17743 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
17745 * config/aarch64/aarch64-simd.md
17746 (aarch64_ld1<VALL:mode>): New.
17747 (aarch64_st1<VALL:mode>): Likewise.
17748 * config/aarch64/arm_neon.h
17749 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
17751 2013-07-02 Sriraman Tallam <tmsriram@google.com>
17753 * config/i386/i386.c (gate_insert_vzeroupper): Check if
17755 (ix86_option_override_internal):Turn on all -mavx target flags by
17756 default as they are dependent on AVX anyway.
17758 2013-07-02 Cary Coutant <ccoutant@google.com>
17760 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
17761 deterministic hash.
17762 (loc_checksum_ordered): Likewise.
17763 (hash_loc_operands): Remove inline keyword.
17765 2013-07-02 Jakub Jelinek <jakub@redhat.com>
17767 PR tree-optimization/57741
17768 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
17769 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
17770 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
17771 Allow REAL_CST step_exprs if flag_associative_math.
17772 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
17774 2013-07-02 Ian Bolton <ian.bolton@arm.com>
17776 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
17778 2013-07-02 Ian Bolton <ian.bolton@arm.com>
17780 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
17782 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17784 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
17786 (iorsi3_insn): Likewise.
17787 (arm_xorsi3): Likewise.
17789 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
17791 * arm.md (attribute "wtype"): Delete. Move attribute values from here
17793 (attribute "type"): ... here, and prefix with "wmmx_".
17794 (attribute "core_cycles"): Update for attribute changes.
17795 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
17796 (tbcstv4hi): Likewise.
17797 (tbcstv2si): Likewise.
17798 (iwmmxt_iordi3): Likewise.
17799 (iwmmxt_xordi3): Likewise.
17800 (iwmmxt_anddi3): Likewise.
17801 (iwmmxt_nanddi3): Likewise.
17802 (iwmmxt_arm_movdi): Likewise.
17803 (iwmmxt_movsi_insn): Likewise.
17804 (mov<mode>_internal): Likewise.
17805 (and<mode>3_iwmmxt): Likewise.
17806 (ior<mode>3_iwmmxt): Likewise.
17807 (xor<mode>3_iwmmxt): Likewise.
17808 (add<mode>3_iwmmxt): Likewise.
17809 (ssaddv8qi3): Likewise.
17810 (ssaddv4hi3): Likewise.
17811 (ssaddv2si3): Likewise.
17812 (usaddv8qi3): Likewise.
17813 (usaddv4hi3): Likewise.
17814 (usaddv2si3): Likewise.
17815 (sub<mode>3_iwmmxt): Likewise.
17816 (sssubv8qi3): Likewise.
17817 (sssubv4hi3): Likewise.
17818 (sssubv2si3): Likewise.
17819 (ussubv8qi3): Likewise.
17820 (ussubv4hi3): Likewise.
17821 (ussubv2si3): Likewise.
17822 (mulv4hi3_iwmmxt): Likewise.
17823 (smulv4hi3_highpart): Likewise.
17824 (umulv4hi3_highpart): Likewise.
17825 (iwmmxt_wmacs): Likewise.
17826 (iwmmxt_wmacsz): Likewise.
17827 (iwmmxt_wmacu): Likewise.
17828 (iwmmxt_wmacuz): Likewise.
17829 (iwmmxt_clrdi): Likewise.
17830 (iwmmxt_clrv8qi): Likewise.
17831 (iwmmxt_clr4hi): Likewise.
17832 (iwmmxt_clr2si): Likewise.
17833 (iwmmxt_uavgrndv8qi3): Likewise.
17834 (iwmmxt_uavgrndv4hi3): Likewise.
17835 (iwmmxt_uavgv8qi3): Likewise.
17836 (iwmmxt_uavgv4hi3): Likewise.
17837 (iwmmxt_tinsrb): Likewise.
17838 (iwmmxt_tinsrh): Likewise.
17839 (iwmmxt_tinsrw): Likewise.
17840 (iwmmxt_textrmub): Likewise.
17841 (iwmmxt_textrmsb): Likewise.
17842 (iwmmxt_textrmuh): Likewise.
17843 (iwmmxt_textrmsh): Likewise.
17844 (iwmmxt_textrmw): Likewise.
17845 (iwmxxt_wshufh): Likewise.
17846 (eqv8qi3): Likewise.
17847 (eqv4hi3): Likewise.
17848 (eqv2si3): Likewise.
17849 (gtuv8qi3): Likewise.
17850 (gtuv4hi3): Likewise.
17851 (gtuv2si3): Likewise.
17852 (gtv8qi3): Likewise.
17853 (gtv4hi3): Likewise.
17854 (gtv2si3): Likewise.
17855 (smax<mode>3_iwmmxt): Likewise.
17856 (umax<mode>3_iwmmxt): Likewise.
17857 (smin<mode>3_iwmmxt): Likewise.
17858 (umin<mode>3_iwmmxt): Likewise.
17859 (iwmmxt_wpackhss): Likewise.
17860 (iwmmxt_wpackwss): Likewise.
17861 (iwmmxt_wpackdss): Likewise.
17862 (iwmmxt_wpackhus): Likewise.
17863 (iwmmxt_wpackwus): Likewise.
17864 (iwmmxt_wpackdus): Likewise.
17865 (iwmmxt_wunpckihb): Likewise.
17866 (iwmmxt_wunpckihh): Likewise.
17867 (iwmmxt_wunpckihw): Likewise.
17868 (iwmmxt_wunpckilb): Likewise.
17869 (iwmmxt_wunpckilh): Likewise.
17870 (iwmmxt_wunpckilw): Likewise.
17871 (iwmmxt_wunpckehub): Likewise.
17872 (iwmmxt_wunpckehuh): Likewise.
17873 (iwmmxt_wunpckehuw): Likewise.
17874 (iwmmxt_wunpckehsb): Likewise.
17875 (iwmmxt_wunpckehsh): Likewise.
17876 (iwmmxt_wunpckehsw): Likewise.
17877 (iwmmxt_wunpckelub): Likewise.
17878 (iwmmxt_wunpckeluh): Likewise.
17879 (iwmmxt_wunpckeluw): Likewise.
17880 (iwmmxt_wunpckelsb): Likewise.
17881 (iwmmxt_wunpckelsh): Likewise.
17882 (iwmmxt_wunpckelsw): Likewise.
17883 (ror<mode>3): Likewise.
17884 (ashr<mode>3_iwmmxt): Likewise.
17885 (lshr<mode>3_iwmmxt): Likewise.
17886 (ashl<mode>3_iwmmxt): Likewise.
17887 (ror<mode>3_di): Likewise.
17888 (ashr<mode>3_di): Likewise.
17889 (lshr<mode>3_di): Likewise.
17890 (ashl<mode>3_di): Likewise.
17891 (iwmmxt_wmadds): Likewise.
17892 (iwmmxt_wmaddu): Likewise.
17893 (iwmmxt_tmia): Likewise.
17894 (iwmmxt_tmiaph): Likewise.
17895 (iwmmxt_tmiabb): Likewise.
17896 (iwmmxt_tmiatb): Likewise.
17897 (iwmmxt_tmiabt): Likewise.
17898 (iwmmxt_tmiatt): Likewise.
17899 (iwmmxt_tmovmskb): Likewise.
17900 (iwmmxt_tmovmskh): Likewise.
17901 (iwmmxt_tmovmskw): Likewise.
17902 (iwmmxt_waccb): Likewise.
17903 (iwmmxt_wacch): Likewise.
17904 (iwmmxt_waccw): Likewise.
17905 (iwmmxt_waligni): Likewise.
17906 (iwmmxt_walignr): Likewise.
17907 (iwmmxt_walignr0): Likewise.
17908 (iwmmxt_walignr1): Likewise.
17909 (iwmmxt_walignr2): Likewise.
17910 (iwmmxt_walignr3): Likewise.
17911 (iwmmxt_wsadb): Likewise.
17912 (iwmmxt_wsadh): Likewise.
17913 (iwmmxt_wsadbz): Likewise.
17914 (iwmmxt_wsadhz): Likewise.
17915 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
17916 (iwmmxt_wabsdiffb): Likewise.
17917 (iwmmxt_wabsdiffh): Likewise.
17918 (iwmmxt_wabsdiffw): Likewise.
17919 (iwmmxt_waddsubhx): Likewise
17920 (iwmmxt_wsubaddhx): Likewise.
17921 (addc<mode>3): Likewise.
17922 (iwmmxt_avg4): Likewise.
17923 (iwmmxt_avg4r): Likewise.
17924 (iwmmxt_wmaddsx): Likewise.
17925 (iwmmxt_wmaddux): Likewise.
17926 (iwmmxt_wmaddsn): Likewise.
17927 (iwmmxt_wmaddun): Likewise.
17928 (iwmmxt_wmulwsm): Likewise.
17929 (iwmmxt_wmulwum): Likewise.
17930 (iwmmxt_wmulsmr): Likewise.
17931 (iwmmxt_wmulumr): Likewise.
17932 (iwmmxt_wmulwsmr): Likewise.
17933 (iwmmxt_wmulwumr): Likewise.
17934 (iwmmxt_wmulwl): Likewise.
17935 (iwmmxt_wqmulm): Likewise.
17936 (iwmmxt_wqmulwm): Likewise.
17937 (iwmmxt_wqmulmr): Likewise.
17938 (iwmmxt_wqmulwmr): Likewise.
17939 (iwmmxt_waddbhusm): Likewise.
17940 (iwmmxt_waddbhusl): Likewise.
17941 (iwmmxt_wqmiabb): Likewise.
17942 (iwmmxt_wqmiabt): Likewise.
17943 (iwmmxt_wqmiatb): Likewise.
17944 (iwmmxt_wqmiatt): Likewise.
17945 (iwmmxt_wqmiabbn): Likewise.
17946 (iwmmxt_wqmiabtn): Likewise.
17947 (iwmmxt_wqmiatbn): Likewise.
17948 (iwmmxt_wqmiattn): Likewise.
17949 (iwmmxt_wmiabb): Likewise.
17950 (iwmmxt_wmiabt): Likewise.
17951 (iwmmxt_wmiatb): Likewise.
17952 (iwmmxt_wmiatt): Likewise.
17953 (iwmmxt_wmiabbn): Likewise.
17954 (iwmmxt_wmiabtn): Likewise.
17955 (iwmmxt_wmiatbn): Likewise.
17956 (iwmmxt_wmiattn): Likewise.
17957 (iwmmxt_wmiawbb): Likewise.
17958 (iwmmxt_wmiawbt): Likewise.
17959 (iwmmxt_wmiawtb): Likewise.
17960 (iwmmxt_wmiawtt): Likewise.
17961 (iwmmxt_wmiawbbn): Likewise.
17962 (iwmmxt_wmiawbtn): Likewise.
17963 (iwmmxt_wmiawtbn): Likewise.
17964 (iwmmxt_wmiawttn): Likewise.
17965 (iwmmxt_wmerge): Likewise.
17966 (iwmmxt_tandc<mode>3): Likewise.
17967 (iwmmxt_torc<mode>3): Likewise.
17968 (iwmmxt_torvsc<mode>3): Likewise.
17969 (iwmmxt_textrc<mode>3): Likewise.
17970 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
17971 (wmmxt_pack): Likewise.
17972 (wmmxt_mult_c1): Likewise.
17973 (wmmxt_mult_c2): Likewise.
17974 (wmmxt_alu_c1): Likewise.
17975 (wmmxt_alu_c2): Likewise.
17976 (wmmxt_alu_c3): Likewise.
17977 (wmmxt_transfer_c1): Likewise.
17978 (wmmxt_transfer_c2): Likewise.
17979 (wmmxt_transfer_c3): Likewise.
17980 (marvell_f_iwmmxt_wstr): Likewise.
17981 (marvell_f_iwmmxt_wldr): Likewise.
17983 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
17985 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
17987 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
17990 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
17991 * lra-constraints.c (need_for_split_p): Check call used hard regs
17992 living through calls.
17994 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
17995 call used regs for call insn.
17997 2013-06-28 Jakub Jelinek <jakub@redhat.com>
18000 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
18001 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
18003 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18005 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
18007 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
18009 * lra-constraints.c (need_for_split_p): Check call used hard regs
18010 living through calls.
18012 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
18015 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
18016 to tie with any other modes. Eliminate Altivec vector mode tests,
18017 since these are a subset of ALTIVEC or VSX vector modes. Simplify
18018 code, to return 0 if testing MODE2 for a condition, if we've
18019 already tested MODE1 for the same condition.
18021 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18023 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
18026 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18028 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
18029 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
18031 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
18033 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
18035 (aarch64_symbolic_constant_p): Remove.
18036 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
18037 static. Fix line length and white space.
18038 (aarch64_symbolic_constant_p): Remove.
18039 * config/aarch64/predicates.md (aarch64_valid_symref):
18040 Use aarch64_classify_symbol_expression.
18042 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18044 * config/arm/constraints.md (Ts): New constraint.
18045 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
18047 (compare_scc): Use "Ts" constraint for operand 0.
18048 (ior_scc_scc): Likewise.
18049 (and_scc_scc): Likewise.
18050 (and_scc_scc_nodom): Likewise.
18051 (ior_scc_scc_cmp): Likewise for operand 7.
18052 (and_scc_scc_cmp): Likewise.
18053 * config/arm/thumb2.md (thumb2_movsi_insn):
18054 Add alternatives for 16-bit encodings.
18055 (thumb2_movhi_insn): Likewise.
18056 (thumb2_movsicc_insn): Likewise.
18057 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
18058 (thumb2_negscc): Use "Ts" constraint.
18059 Move mvn instruction outside cond_exec block.
18060 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
18061 for 16-bit encodings.
18063 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18065 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
18067 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
18068 (mulsi3subsi): Likewise.
18069 (mulsidi3adddi): Likewise.
18070 (mulsidi3_v6): Likewise.
18071 (umulsidi3_v6): Likewise.
18072 (umulsidi3adddi_v6): Likewise.
18073 (smulsi3_highpart_v6): Likewise.
18074 (umulsi3_highpart_v6): Likewise.
18075 (mulhisi3tb): Likewise.
18076 (mulhisi3bt): Likewise.
18077 (mulhisi3tt): Likewise.
18078 (maddhisi4): Likewise.
18079 (maddhisi4tb): Likewise.
18080 (maddhisi4tt): Likewise.
18081 (maddhidi4): Likewise.
18082 (maddhidi4tb): Likewise.
18083 (maddhidi4tt): Likewise.
18084 (zeroextractsi_compare0_scratch): Likewise.
18085 (insv_zero): Likewise.
18086 (insv_t2): Likewise.
18087 (anddi_notzesidi_di): Likewise.
18088 (anddi_notsesidi_di): Likewise.
18089 (andsi_notsi_si): Likewise.
18090 (iordi_zesidi_di): Likewise.
18091 (xordi_zesidi_di): Likewise.
18092 (andsi_iorsi3_notsi): Likewise.
18093 (smax_0): Likewise.
18094 (smax_m1): Likewise.
18095 (smin_0): Likewise.
18096 (not_shiftsi): Likewise.
18097 (unaligned_loadsi): Likewise.
18098 (unaligned_loadhis): Likewise.
18099 (unaligned_loadhiu): Likewise.
18100 (unaligned_storesi): Likewise.
18101 (unaligned_storehi): Likewise.
18102 (extv_reg): Likewise.
18103 (extzv_t2): Likewise.
18104 (divsi3): Likewise.
18105 (udivsi3): Likewise.
18106 (arm_zero_extendhisi2addsi): Likewise.
18107 (arm_zero_extendqisi2addsi): Likewise.
18108 (compareqi_eq0): Likewise.
18109 (arm_extendhisi2_v6): Likewise.
18110 (arm_extendqisi2addsi): Likewise.
18111 (arm_movt): Likewise.
18112 (thumb2_ldrd): Likewise.
18113 (thumb2_ldrd_base): Likewise.
18114 (thumb2_ldrd_base_neg): Likewise.
18115 (thumb2_strd): Likewise.
18116 (thumb2_strd_base): Likewise.
18117 (thumb2_strd_base_neg): Likewise.
18118 (arm_negsi2): Add alternative for 16-bit encoding.
18119 (arm_one_cmplsi2): Likewise.
18121 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18123 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
18124 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
18125 (movdfcc): Likewise.
18126 * config/arm/vfp.md (*thumb2_movsf_vfp):
18127 Disable predication for arm_restrict_it.
18128 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
18129 (*thumb2_movdfcc_vfp): Likewise.
18130 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
18131 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
18132 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
18133 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
18134 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
18135 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
18136 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
18137 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
18138 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
18139 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
18140 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
18141 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
18142 Disable predication for arm_restrict_it.
18144 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
18146 * config/i386/bmiintrin.h (_bextr_u32): New.
18147 (_bextr_u64): Ditto.
18149 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
18151 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
18152 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
18153 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
18154 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
18155 * config/mips/n32-elf.h: ...this new file.
18157 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
18160 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
18161 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
18163 2013-06-27 Catherine Moore <clm@codesourcery.com>
18165 * config/mips/mips-tables.opt: Regenerate.
18166 * config/mips/mips-cpus.def: Add m14ke and m14kec.
18167 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
18168 * doc/invoke.texi: Add -m14kc.
18170 2013-06-27 Jakub Jelinek <jakub@redhat.com>
18173 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
18174 constraints of operand 1 and 2.
18177 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
18178 to match RTL canonicalization. Swap predicates and
18179 constraints of operand 1 and 2.
18181 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
18183 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
18184 Process OP_INOUT regs for splitting too.
18186 2013-06-27 Jakub Jelinek <jakub@redhat.com>
18188 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
18189 decl before the loop, initialize to NULL.
18190 (vectorizable_load): Initialize ptr_incr to NULL.
18192 2013-06-27 Martin Jambor <mjambor@suse.cz>
18195 * ipa-ref.h (ipa_maybe_record_reference): Declare.
18196 * ipa-ref.c (ipa_maybe_record_reference): New function.
18197 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
18198 * ipa-cp.c (create_specialized_node): Record potential references from
18200 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
18202 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
18204 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
18205 parameter 'mode' of type 'enum machine_mode mode'; change to pass
18206 'mode' to force_reg.
18207 (aarch64_add_offset): Update calls to aarch64_force_temporary.
18208 (aarch64_expand_mov_immediate): Likewise.
18210 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
18212 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
18213 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
18215 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18217 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
18218 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
18219 (struct machine_function): Add tbegin_p.
18220 (s390_canonicalize_comparison): Fold CC mode compares to
18221 conditional jump if possible.
18222 (s390_emit_jump): Return the emitted jump.
18223 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
18224 Handle CCRAWmode compares.
18225 (s390_option_override): Default to -mhtm if available.
18226 (s390_reg_clobbered_rtx): Handle floating point regs as well.
18227 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
18228 FPRs instead of df_regs_ever_live_p.
18229 (s390_optimize_nonescaping_tx): New function.
18230 (s390_init_frame_layout): Extend clobbered_regs array to cover
18232 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
18233 (s390_expand_tbegin): New function.
18234 (enum s390_builtin): New enum definition.
18235 (code_for_builtin): New array definition.
18236 (s390_init_builtins): New function.
18237 (s390_expand_builtin): New function.
18238 (TARGET_INIT_BUILTINS): Define.
18239 (TARGET_EXPAND_BUILTIN): Define.
18240 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
18241 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
18242 (s390_alc_comparison): Likewise.
18243 * config/s390/s390-modes.def: Add CCRAWmode.
18244 * config/s390/s390.h (processor_flags): Add PF_TX.
18245 (TARGET_CPU_HTM): Define macro.
18246 (TARGET_HTM): Define macro.
18247 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
18248 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
18249 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
18250 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
18251 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
18252 (TBEGIN_MASK, TBEGINC_MASK): New constants.
18253 ("*cc_to_int"): Move up.
18254 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
18255 constants other than 0.
18256 ("*ccraw_to_int"): New insn and splitter definition.
18257 ("tbegin", "tbegin_nofloat", "tbegin_retry")
18258 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
18259 ("tx_assist"): New expander.
18260 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
18261 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
18262 * config/s390/s390.opt: Add -mhtm option.
18263 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
18264 * config/s390/htmxlintrin.h: New file.
18265 * config/s390/htmintrin.h: New file.
18266 * config/s390/s390intrin.h: New file.
18267 * doc/extend.texi: Document htm builtins.
18268 * config.gcc: Add the new header files to extra_headers.
18270 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
18272 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
18273 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
18275 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
18276 Pat Haugen <pthaugen@us.ibm.com>
18277 Peter Bergner <bergner@vnet.ibm.com>
18279 * config/rs6000/power8.md: New.
18280 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
18281 setting for power8 entry.
18282 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
18283 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
18284 test for Power4/Power5 only.
18285 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
18287 (force_new_group): Adjust comment.
18288 * config/rs6000/rs6000.md: Include power8.md.
18290 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
18292 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
18293 * config/arm/arm-protos.h (arm_max_conditional_execute): New
18295 (tune_params): Update comment.
18296 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
18297 (arm_max_conditional_execute): New function.
18298 (thumb2_final_prescan_insn): Use max_insn_skipped and
18299 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
18301 2013-06-25 Jakub Jelinek <jakub@redhat.com>
18303 PR tree-optimization/57705
18304 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
18305 SSA_NAME step, provided that it is not defined inside the loop.
18306 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
18307 (get_initial_def_for_induction): Handle SSA_NAME IV step.
18309 2013-06-25 Martin Jambor <mjambor@suse.cz>
18311 PR middle-end/57670
18312 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
18313 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
18315 (ipa_note_param_call): Initialize member_ptr flag.
18316 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
18317 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
18318 (ipa_write_indirect_edge_info): Stream member_ptr flag.
18319 (ipa_read_indirect_edge_info): Likewise.
18321 2013-06-25 Richard Biener <rguenther@suse.de>
18323 PR middle-end/56977
18324 * passes.c (init_optimization_passes): Move pass_fold_builtins
18325 and pass_dce earlier with -Og.
18327 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
18329 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
18330 <BIT_FIELD_REF>: Remove trailing TAB.
18331 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
18334 2013-06-24 Martin Jambor <mjambor@suse.cz>
18336 PR tree-optimization/57358
18337 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
18338 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
18339 (ipa_analyze_params_uses): Generate pessimistic info when true.
18341 2013-06-24 Martin Jambor <mjambor@suse.cz>
18343 PR tree-optimization/57539
18344 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
18345 global.inlined_to of the new node to it. All callers changed.
18346 * ipa-inline-transform.c (clone_inlined_nodes): New variable
18347 inlining_into, pass it to cgraph_clone_node.
18348 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
18349 ipa_free_edge_args_substructures.
18350 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
18351 rdesc linked list. Do not assert rdesc edges have inlined caller.
18352 Assert we have found an rdesc in the rdesc list.
18354 2013-06-24 Richard Biener <rguenther@suse.de>
18356 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
18357 (pointer_set_lookup): Declare.
18358 (class pointer_map): New template class implementing a
18359 generic pointer to T map.
18360 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
18361 pointer_map<T>::contains, pointer_map<T>::insert,
18362 pointer_map<T>::traverse): New functions.
18363 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
18364 (pointer_set_lookup): New function.
18365 (pointer_set_contains): Use pointer_set_lookup.
18366 (pointer_set_insert): Likewise.
18367 (insert_aux): Remove.
18368 (struct pointer_map_t): Embed a pointer_set_t.
18369 (pointer_map_create): Adjust.
18370 (pointer_map_destroy): Likewise.
18371 (pointer_map_contains): Likewise.
18372 (pointer_map_insert): Likewise.
18373 (pointer_map_traverse): Likewise.
18374 * tree-streamer.h (struct streamer_tree_cache_d): Use a
18375 pointer_map<unsigned> instead of a pointer_map_t.
18376 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
18377 (streamer_tree_cache_lookup): Likewise.
18378 (streamer_tree_cache_create): Likewise.
18379 (streamer_tree_cache_delete): Likewise.
18380 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
18381 pointer_map<unsigned> instead of a pointer_map_t.
18382 (lto_init_tree_ref_encoder): Adjust.
18383 (lto_destroy_tree_ref_encoder): Likewise.
18384 * lto-section-out.c (lto_output_decl_index): Likewise.
18385 (lto_record_function_out_decl_state): Likewise.
18386 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
18388 2013-06-24 Richard Biener <rguenther@suse.de>
18390 PR tree-optimization/57488
18391 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
18393 2013-06-24 Alan Modra <amodra@gmail.com>
18395 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
18396 (gen_easy_altivec_constant): Likewise.
18397 * config/rs6000/predicates.md (easy_vector_constant_add_self,
18398 easy_vector_constant_msb): Likewise.
18400 2013-06-23 Jakub Jelinek <jakub@redhat.com>
18403 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
18404 add missing return true.
18406 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
18409 * config/sh/predicates.md (general_extend_operand): Invoke
18410 general_movsrc_operand for memory operands.
18411 (general_movsrc_operand): Allow reg+reg addressing, do not use
18412 general_operand for memory operands.
18414 2013-06-23 Sriraman Tallam <tmsriram@google.com>
18416 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
18417 when current target options does not apply.
18418 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
18419 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
18420 * config/i386/bmiintrin.h: Pass appropriate target
18421 attributes to header.
18422 * config/i386/mmintrin.h: Ditto.
18423 * config/i386/nmmintrin.h: Ditto.
18424 * config/i386/avx2intrin.h: Ditto.
18425 * config/i386/fxsrintrin.h: Ditto.
18426 * config/i386/tbmintrin.h: Ditto.
18427 * config/i386/xsaveintrin.h: Ditto.
18428 * config/i386/f16cintrin.h: Ditto.
18429 * config/i386/xtestintrin.h: Ditto.
18430 * config/i386/xsaveoptintrin.h: Ditto.
18431 * config/i386/bmi2intrin.h: Ditto.
18432 * config/i386/lzcntintrin.h: Ditto.
18433 * config/i386/smmintrin.h: Ditto.
18434 * config/i386/wmmintrin.h: Ditto.
18435 * config/i386/x86intrin.h: Remove all header include guards.
18436 * config/i386/prfchwintrin.h: Ditto.
18437 * config/i386/pmmintrin.h: Ditto.
18438 * config/i386/tmmintrin.h: Ditto.
18439 * config/i386/xmmintrin.h: Ditto.
18440 * config/i386/popcntintrin.h: Ditto.
18441 * config/i386/rdseedintrin.h: Ditto.
18442 * config/i386/ammintrin.h: Ditto.
18443 * config/i386/emmintrin.h: Ditto.
18444 * config/i386/immintrin.h: Remove all header include guards.
18445 * config/i386/fma4intrin.h: Ditto.
18446 * config/i386/lwpintrin.h: Ditto.
18447 * config/i386/xopintrin.h: Ditto.
18448 * config/i386/ia32intrin.h: Ditto.
18449 * config/i386/avxintrin.h: Ditto.
18450 * config/i386/rtmintrin.h: Ditto.
18451 * config/i386/fmaintrin.h: Ditto.
18452 * config/i386/mm3dnow.h: Ditto.
18454 2013-06-22 Sriraman Tallam <tmsriram@google.com>
18456 * common/config/i386/i386-common.c: Handle LZCNT.
18458 2013-06-22 Andi Kleen <ak@linux.intel.com>
18460 * doc/extend.texi: Use __atomic_store_n instead of
18461 __atomic_store in HLE example.
18463 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
18465 * config/sh/sh.c: Remove <cstdlib> workaround.
18467 2013-06-21 Andi Kleen <ak@linux.intel.com>
18469 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
18471 2013-06-21 Andi Kleen <ak@linux.intel.com>
18473 * doc/extend.texi: Document that __atomic_clear and
18474 __atomic_test_and_set should only be used with bool.
18476 2013-06-20 Jan Hubicka <jh@suse.cz>
18478 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
18479 types_same_for_odr.
18480 * tree.c (decls_same_for_odr): New function.
18481 (same_for_edr): New function.
18482 (types_same_for_odr): New function.
18483 (get_binfo_at_offset): Use it.
18484 * tree.h (types_same_for_odr): Declare.
18486 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
18487 Jason Merrill <jason@redhat.com>
18489 * system.h: Include <cstdlib> as well as <stdlib.h>.
18491 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
18494 * config/i386/i386.c (construct_container): Report error if
18495 long double is used with disabled x87 float returns.
18497 2013-06-20 Jan Hubicka <jh@suse.cz>
18499 * lto-cgraph.c (input_symtab): Do not set cgraph state.
18501 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
18503 PR rtl-optimization/57425
18504 PR rtl-optimization/57569
18505 * alias.c (write_dependence_p): Remove parameters mem_mode and
18506 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
18507 Changed all callers.
18508 (canon_anti_dependence): Get comments and semantics in sync.
18509 Add parameter mem_canonicalized. Changed all callers.
18510 * rtl.h (canon_anti_dependence): Update prototype.
18512 2013-06-20 Richard Biener <rguenther@suse.de>
18514 * data-streamer-in.c (streamer_read_uhwi): Optimize single
18515 byte case, inline streamer_read_uchar and defer section
18518 2013-06-20 Richard Biener <rguenther@suse.de>
18520 PR tree-optimization/57584
18521 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
18522 SSA names into the expanded expression that take part in
18523 abnormal coalescing.
18525 2013-06-19 Sharad Singhai <singhai@google.com>
18527 * gcov.c (print_usage): Handle new option.
18528 (process_args): Ditto.
18529 (get_gcov_intermediate_filename): New function.
18530 (output_intermediate_file): New function.
18531 (output_gcov_file): New function
18532 (generate_results): Handle new option.
18533 (release_function): Relase demangled name.
18534 (read_graph_file): Handle demangled name.
18535 (output_lines): Ditto.
18536 * doc/gcov.texi: Document gcov intermediate format.
18538 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
18541 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
18542 (lra_emit_add): Use the functions. Add comment about Y as an
18545 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
18548 * collect2.c (collect_atexit): New.
18549 (collect_exit): Delete.
18550 (main): Register collect_atexit with atexit.
18551 (collect_wait): Change collect_exit to exit.
18553 * collect2.h (collect_exit): Delete.
18554 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
18556 2013-06-19 Wei Mi <wmi@google.com>
18558 PR rtl-optimization/57518
18559 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
18560 if regno is used in paradoxical subreg.
18561 (update_equiv_regs): Check pdx_subregs[regno] before
18562 set a reg to be equivalent with a mem.
18564 2013-06-19 Matthias Klose <doko@ubuntu.com>
18567 * file-find.h (find_a_file): Add a mode parameter.
18568 * file-find.c (find_a_file): Likewise.
18569 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
18570 with X_OK for the executables.
18571 * collect2.c (main): Call find_a_file with X_OK.
18573 2013-06-19 Steve Ellcey <sellcey@mips.com>
18576 * config/mips/mips.md (casesi_internal_mips16_<mode>):
18577 Use NEXT_INSN instead of next_real_insn.
18579 2013-06-19 Jan Hubicka <jh@suse.cz>
18581 * cgraph.h (const_value_known_p): Replace by ...
18582 (ctor_for_folding): .. this one.
18583 * cgraphunit.c (process_function_and_variable_attributes): Use it.
18584 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
18585 * expr.c (expand_expr_real_1): Likewise.
18586 (string_constant): Likewise.
18587 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
18588 * ipa.c (process_references): Likewise.
18589 (symtab_remove_unreachable_nodes): Likewise.
18590 * ipa-inline-analysis.c (param_change_prob): Likewise.
18591 * gimple-fold.c (canonicalize_constructor_val): Likewise.
18592 (get_base_constructor): Likwise.
18593 * varpool.c (varpool_remove_node): Likewise.
18594 (varpool_remove_initializer): LIkewise.
18595 (dump_varpool_node): LIkwise.
18596 (const_value_known_p): Rewrite to ...
18597 (ctor_for_folding): ... this one.
18599 2013-06-19 Jakub Jelinek <jakub@redhat.com>
18602 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
18603 PERSONALITY in $PATH derived prefixes.
18605 2013-06-19 Jeff Law <law@redhat.com>
18607 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
18610 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
18611 (simplify_bitwise_binary): Use it to simpify certain binary ops on
18614 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
18616 * config/arm/vfp.md: Move VFP instruction classification documentation
18618 * config/arm/arm.md: ... here. Update instruction classification
18621 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
18623 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
18624 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
18625 pattern. Use more efficient sequences on ARMv5 and Thumb2.
18627 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
18630 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
18631 with NEXT_INSN. Use tablejump_p to check for jump table data
18634 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
18637 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
18638 that now in C++ the value is correct per the C++ standards.
18640 2013-06-19 Richard Biener <rguenther@suse.de>
18642 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
18643 for global context.
18645 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18648 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18651 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
18652 with next_active_insn.
18654 2013-06-18 Sriraman Tallam <tmsriram@google.com>
18656 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
18657 functions are inlined during failures to flag an error.
18658 * tree-inline.c (expand_call_inline): Allow the error to be flagged
18659 in early inline pass.
18661 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
18663 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
18666 2013-06-18 Julian Brown <julian@codesourcery.com>
18668 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
18669 Permit virtual register pre-reload if !strict.
18670 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
18672 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
18674 * config/arm/neon.md (movmisalign<mode>): Use
18675 neon_perm_struct_or_reg_operand instead of
18676 neon_struct_or_register_operand.
18677 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
18678 neon_permissive_struct_operand instead of neon_struct_operand.
18679 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
18680 neon_vector_mem_operand.
18681 * config/arm/predicates.md (neon_struct_operand): Adjust call to
18682 neon_vector_mem_operand.
18683 (neon_permissive_struct_operand): New.
18684 (neon_struct_or_register_operand): Rename to...
18685 (neon_perm_struct_or_reg_operand): This. Adjust call to
18686 neon_vector_mem_operand.
18688 2013-06-18 Richard Biener <rguenther@suse.de>
18690 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
18691 * lto-streamer.h: Include pointer-set.h.
18692 (struct lto_decl_slot): Remove.
18693 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
18694 Remove next_index entry.
18695 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
18696 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
18697 (lto_init_tree_ref_encoder): Adjust.
18698 (lto_destroy_tree_ref_encoder): Likewise.
18699 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
18700 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
18701 (lto_output_decl_index): Adjust.
18702 (lto_new_out_decl_state): Likewise.
18703 (lto_record_function_out_decl_state): Likewise.
18704 * lto-streamer-out.c (copy_function): Likewise.
18706 2013-06-18 Richard Biener <rguenther@suse.de>
18708 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
18709 * cgraphunit.c: Include cfgloop.h.
18710 (init_lowered_empty_function): Initialize the loop tree.
18711 (assemble_thunk): Insert new BBs into loops.
18713 2013-06-18 Richard Biener <rguenther@suse.de>
18715 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
18716 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
18717 the map from cache entry to cache index optional.
18718 (streamer_tree_cache_replace_tree): Adjust accordingly.
18719 (streamer_tree_cache_append): Likewise.
18720 (streamer_tree_cache_delete): Likewise.
18721 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
18722 streamer cache map from cache entry to cache index.
18723 * lto-streamer-out.c (create_output_block): Adjust.
18725 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
18727 * config/arm/arm.md (attribute "insn"): Move multiplication and
18728 division attributes to...
18729 (attribute "type"): ... here. Remove mult.
18730 (attribute "mul32"): New attribute.
18731 (attribute "mul64"): Add umaal.
18732 (*arm_mulsi3): Update attributes.
18733 (*arm_mulsi3_v6): Likewise.
18734 (*thumb_mulsi3): Likewise.
18735 (*thumb_mulsi3_v6): Likewise.
18736 (*mulsi3_compare0): Likewise.
18737 (*mulsi3_compare0_v6): Likewise.
18738 (*mulsi_compare0_scratch): Likewise.
18739 (*mulsi_compare0_scratch_v6): Likewise.
18740 (*mulsi3addsi): Likewise.
18741 (*mulsi3addsi_v6): Likewise.
18742 (*mulsi3addsi_compare0): Likewise.
18743 (*mulsi3addsi_compare0_v6): Likewise.
18744 (*mulsi3addsi_compare0_scratch): Likewise.
18745 (*mulsi3addsi_compare0_scratch_v6): Likewise.
18746 (*mulsi3subsi): Likewise.
18747 (*mulsidi3adddi): Likewise.
18748 (*mulsi3addsi_v6): Likewise.
18749 (*mulsidi3adddi_v6): Likewise.
18750 (*mulsidi3_nov6): Likewise.
18751 (*mulsidi3_v6): Likewise.
18752 (*umulsidi3_nov6): Likewise.
18753 (*umulsidi3_v6): Likewise.
18754 (*umulsidi3adddi): Likewise.
18755 (*umulsidi3adddi_v6): Likewise.
18756 (*smulsi3_highpart_nov6): Likewise.
18757 (*smulsi3_highpart_v6): Likewise.
18758 (*umulsi3_highpart_nov6): Likewise.
18759 (*umulsi3_highpart_v6): Likewise.
18760 (mulhisi3): Likewise.
18761 (*mulhisi3tb): Likewise.
18762 (*mulhisi3bt): Likewise.
18763 (*mulhisi3tt): Likewise.
18764 (maddhisi4): Likewise.
18765 (*maddhisi4tb): Likewise.
18766 (*maddhisi4tt): Likewise.
18767 (maddhidi4): Likewise.
18768 (*maddhidi4tb): Likewise.
18769 (*maddhidi4tt): Likewise.
18770 (divsi3): Likewise.
18771 (udivsi3): Likewise.
18772 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
18773 (thumb2_mulsi_short_compare0): Likewise.
18774 (thumb2_mulsi_short_compare0_scratch): Likewise.
18775 * config/arm/arm1020e.md (1020mult1): Update attribute change.
18776 (1020mult2): Likewise.
18777 (1020mult3): Likewise.
18778 (1020mult4): Likewise.
18779 (1020mult5): Likewise.
18780 (1020mult6): Likewise.
18781 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
18783 (cortex_a15_mult64): Likewise.
18784 (cortex_a15_sdiv): Likewise.
18785 (cortex_a15_udiv): Likewise.
18786 * config/arm/arm1026ejs.md (mult1): Update attribute change.
18792 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
18793 (pj4_ir_div): Likewise.
18794 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
18795 (11_mult2): Likewise.
18796 (11_mult3): Likewise.
18797 (11_mult4): Likewise.
18798 (11_mult5): Likewise.
18799 (11_mult6): Likewise.
18800 (11_mult7): Likewise.
18801 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
18802 (cortex_a8_mla): Likewise.
18803 (cortex_a8_mull): Likewise.
18804 (cortex_a8_smulwy): Likewise.
18805 (cortex_a8_smlald): Likewise.
18806 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
18807 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
18808 (cortex_r4_mul_3): Likewise.
18809 (cortex_r4_mla_4): Likewise.
18810 (cortex_r4_mla_3): Likewise.
18811 (cortex_r4_smlald): Likewise.
18812 (cortex_r4_mull): Likewise.
18813 (cortex_r4_sdiv): Likewise.
18814 (cortex_r4_udiv): Likewise.
18815 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
18816 (cortex_a7_idiv): Likewise.
18817 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
18818 (9_mult2): Likewise.
18819 (9_mult3): Likewise.
18820 (9_mult4): Likewise.
18821 (9_mult5): Likewise.
18822 (9_mult6): Likewise.
18823 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
18824 (cortex_a53_sdiv): Likewise.
18825 (cortex_a53_udiv): Likewise.
18826 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
18827 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
18828 (mp626_mult2): Likewise.
18829 (mp626_mult3): Likewise.
18830 (mp626_mult4): Likewise.
18831 * config/arm/fa526.md (526_mult1): Update attribute change.
18832 (526_mult2): Likewise.
18833 * config/arm/arm-generic.md (mult): Update attribute change.
18834 (mult_ldsched_strongarm): Likewise.
18835 (mult_ldsched): Likewise.
18836 (multi_cycle): Likewise.
18837 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
18838 * config/arm/fa606te.md (606te_mult1): Update attribute change.
18839 (606te_mult2): Likewise.
18840 (606te_mult3): Likewise.
18841 (606te_mult4): Likewise.
18842 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
18843 (cortex_a9_mac16): Likewise.
18844 (cortex_a9_multiply): Likewise.
18845 (cortex_a9_mac): Likewise.
18846 (cortex_a9_multiply_long): Likewise.
18847 * config/arm/fa626te.md (626te_mult1): Update attribute change.
18848 (626te_mult2): Likewise.
18849 (626te_mult3): Likewise.
18850 (626te_mult4): Likewise.
18852 2013-06-18 Richard Biener <rguenther@suse.de>
18855 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
18857 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18860 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
18861 with next_active_insn.
18863 2013-06-18 Alan Modra <amodra@gmail.com>
18865 * config/rs6000/rs6000.h (enum data_align): New.
18866 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
18867 (DATA_ABI_ALIGNMENT): Define.
18868 (CONSTANT_ALIGNMENT): Correct comment.
18869 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
18870 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
18872 2013-06-17 David Malcolm <dmalcolm@redhat.com>
18874 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
18875 ATTRIBUTE_UNUSED marking.
18877 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
18879 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
18880 alternative and update.
18881 (aarch64_dup_lanedi): Delete.
18882 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
18883 * config/aarch64/aarch64-simd-builtins.def: Update.
18885 2013-06-17 Richard Biener <rguenther@suse.de>
18887 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
18888 (lto_input_scc): Declare.
18889 (lto_input_tree_1): Likewise.
18890 (struct lto_stats_d): Add num_tree_bodies_output and
18891 num_pickle_refs_output.
18892 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
18893 (lto_read_tree_1): Split out from ...
18894 (lto_read_tree): ... this.
18895 (lto_input_scc): New function.
18896 (lto_input_tree_1): Split out from ...
18897 (lto_input_tree): ... this. Handle LTO_tree_scc.
18898 (lto_data_in_create): Create the streamer cache without hashes.
18899 * lto-streamer-out.c (create_output_block): Create the streamer
18900 cache with hashes when not doing WPA.
18901 (lto_write_tree_1): Split out from ...
18902 (lto_write_tree): ... this.
18903 (get_symbol_initial_value): New function.
18904 (lto_output_tree_1): Split out from ...
18905 (lto_output_tree): ... this. Write trees as series of SCCs
18906 using a DFS walk via DFS_write_tree.
18907 (struct sccs, struct scc_entry): New types.
18908 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
18909 (DFS_write_tree_body): New function.
18910 (DFS_write_tree): Likewise.
18911 (hash_tree): Likewise.
18912 (scc_entry_compare): Likewise.
18913 (hash_scc): Likewise.
18914 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
18915 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
18916 TREE_CHAIN as regular reference.
18917 (streamer_read_integer_cst): Remove.
18918 (streamer_get_pickled_tree): Adjust.
18919 * tree-streamer-out.c (streamer_write_chain): Disable streaming
18920 of DECL_EXTERNALs in BLOCK_VARS for now.
18921 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
18923 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
18924 Add hash value argument and record that if hashes are recorded
18926 (streamer_tree_cache_insert_1): Adjust.
18927 (streamer_tree_cache_insert): Likewise.
18928 (streamer_tree_cache_insert_at): Rename to ...
18929 (streamer_tree_cache_replace_tree): ... this and adjust.
18930 (streamer_tree_cache_append): Adjust.
18931 (record_common_node): Likewise.
18932 (streamer_tree_cache_create): Add argument whether to
18933 record hash values together with trees.
18934 (streamer_tree_cache_delete): Adjust.
18935 * tree-streamer.h (struct streamer_tree_cache_d): Add
18937 (streamer_read_integer_cst): Remove.
18938 (streamer_tree_cache_insert): Adjust.
18939 (streamer_tree_cache_append): Likewise.
18940 (streamer_tree_cache_insert_at): Rename to ...
18941 (streamer_tree_cache_replace_tree): ... this and adjust.
18942 (streamer_tree_cache_create): Add argument whether to record hashes.
18943 (streamer_tree_cache_get): Rename to ...
18944 (streamer_tree_cache_get_tree): ... this.
18945 (streamer_tree_cache_get_hash): New function.
18946 * tree.c (cache_integer_cst): New function.
18947 * tree.h (cache_integer_cst): Declare.
18948 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
18949 * lto-symtab.c (lto_varpool_replace_node): Only release
18950 DECL_INITIAL of non-prevailing decls.
18951 * varpool.c (varpool_remove_initializer): Do not release
18952 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
18954 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
18956 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
18957 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
18958 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
18959 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
18960 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
18961 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
18962 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
18963 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
18964 instead of TARGET_64BIT.
18965 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
18966 Require ISA_HAS_<D>DIV.
18968 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
18970 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
18971 (mips*-*-linux*): Move default with_llsc setting to where other
18973 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
18975 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
18976 Likewise. Remove default with_tune setting. Move default float
18977 setting to its own block. Handle with_llsc in the same block as above.
18979 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
18981 PR rtl-optimization/57425
18982 PR rtl-optimization/57569
18983 * alias.c (write_dependence_p): Add new parameters mem_mode,
18984 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
18985 Changed all callers.
18986 (canon_anti_dependence): New function.
18987 * cse.c (check_dependence): Use canon_anti_dependence.
18988 * cselib.c (cselib_invalidate_mem): Likewise.
18989 * rtl.h (canon_anti_dependence): Declare.
18991 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
18993 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
18994 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
18995 ".set mips3" for 64-bit targets.
18997 2013-06-15 Dehao Chen <dehao@google.com>
18999 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
19000 * gimple-low.c (gimple_check_call_matching_types): Likewise.
19001 (gimple_check_call_args): Likewise.
19002 * value-prof.c (check_ic_target): Likewise.
19003 * ipa-inline.c (early_inliner): Likewise.
19004 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
19005 * cgraph.c (cgraph_create_edge_1): Likewise.
19006 (cgraph_make_edge_direct): Likewise.
19008 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
19011 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
19012 rs6000_output_move_128bit to handle emitting quad memory
19013 operations. Set attribute length to 8 bytes.
19015 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
19017 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
19019 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
19020 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
19021 (aarch64_<su>mlsl<mode>): Likewise.
19023 2013-06-14 Mike Stump <mikestump@comcast.net>
19025 * Makefile.in (TARGET_H): Add insn-codes.h.
19027 2013-06-14 Alan Modra <amodra@gmail.com>
19029 PR middle-end/57134
19030 PR middle-end/57586
19031 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
19032 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
19033 bitfield expansion when EXPAND_MEMORY.
19034 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
19036 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
19038 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
19039 test for clearing quad memory on 32-bit later.
19041 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
19043 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
19044 (fold_negate_expr): Likewise.
19045 (fold_real_zero_addition_p): Handle vectors.
19046 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
19048 2013-06-14 Alan Modra <amodra@gmail.com>
19050 * varasm.c (force_const_mem): Revert 2013-06-07 change.
19052 2013-06-13 Jan Hubicka <jh@suse.cz>
19054 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
19055 Local comdats are not externally visible.
19056 * symtab.c (dump_symtab_base): Dump externally visible.
19057 (verify_symtab_base): Verify back links in the symtab hash.
19059 2013-06-13 Bin Cheng <bin.cheng@arm.com>
19061 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
19062 CONVERT_EXPR as equal nodes.
19064 2013-06-13 Bin Cheng <bin.cheng@arm.com>
19066 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
19068 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
19070 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
19071 Generalize to complex and vector.
19072 * tree.c (build_all_ones_cst): New function.
19073 * tree.h (build_all_ones_cst): Declare it.
19075 2013-06-13 Alan Modra <amodra@gmail.com>
19077 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
19078 * config/rs6000/rs6000.md (signbittf2): New insn.
19079 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
19080 (abstf2_internal, cmptf_internal2): Likewise.
19081 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
19083 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
19084 Pat Haugen <pthaugen@us.ibm.com>
19085 Peter Bergner <bergner@vnet.ibm.com>
19087 * config/rs6000/rs6000.c (emit_load_locked): Add support for
19088 power8 byte, half-word, and quad-word atomic instructions.
19089 (emit_store_conditional): Likewise.
19090 (rs6000_expand_atomic_compare_and_swap): Likewise.
19091 (rs6000_expand_atomic_op): Likewise.
19093 * config/rs6000/sync.md (larx): Add new modes for power8.
19095 (AINT): New mode iterator to include TImode as well as normal
19096 integer modes on power8.
19097 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
19098 that VSX registers are not considered. Use AINT mode iterator
19099 instead of INT1 to allow inclusion of quad word atomic operations
19101 (load_locked<mode>): Likewise.
19102 (store_conditional<mode>): Likewise.
19103 (atomic_compare_and_swap<mode>): Likewise.
19104 (atomic_exchange<mode>): Likewise.
19105 (atomic_nand<mode>): Likewise.
19106 (atomic_fetch_<fetchop_name><mode>): Likewise.
19107 (atomic_nand_fetch<mode>): Likewise.
19108 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
19110 (ATOMIC): On power8, add QImode, HImode modes.
19111 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
19112 modes that promote to SImode.
19113 (load_lockedti): Convert TImode arguments to PTImode, so that we
19114 get a guaranteed even/odd register pair.
19115 (load_lockedpti): Likewise.
19116 (store_conditionalti): Likewise.
19117 (store_conditionalpti): Likewise.
19119 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
19120 atomic load/store instructions.
19123 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
19125 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
19127 (insn_count): New attribute, with most cases extracted from...
19128 (length): ...here. Redefine most cases in terms of insn_count.
19129 (single_insn): Delete.
19130 (can_delay): Use insn_count to check for single instructions.
19131 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
19132 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
19133 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
19134 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
19135 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
19136 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
19137 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
19138 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
19139 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
19140 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
19141 rather than "length".
19142 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
19143 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
19144 Use "insn_count" rather than "length".
19145 * config/mips/mips-dsp.md
19146 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
19147 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
19150 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
19152 PR tree-optimization/57361
19153 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
19155 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
19157 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
19159 (aarch64_simd_combine<mode>): New instruction expansion.
19160 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
19161 function prototype.
19162 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
19163 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
19165 2013-06-12 Jan Hubicka <jh@suse.cz>
19167 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
19168 decl has when in streaming stage.
19169 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
19170 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
19172 2013-06-12 Roland Stigge <stigge@antcom.de>
19175 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
19177 2013-06-12 Jakub Jelinek <jakub@redhat.com>
19179 PR tree-optimization/57537
19180 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
19181 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
19183 2013-06-12 Richard Biener <rguenther@suse.de>
19185 * data-streamer.h (streamer_write_char_stream): CSE
19186 obs->current_pointer.
19187 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
19188 streamer_write_char_stream manually and optimize the resulting loop.
19189 (streamer_write_hwi_stream): Likewise.
19191 2013-06-12 Jan Hubicka <jh@suse.cz>
19193 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
19194 * cgraph.h (varpool_create_empty_node): Declare.
19195 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
19197 * symtab.c (symtab_unregister_node): Be lax about missin entries
19199 (symtab_get_node): Update comment.
19200 * varpool.c (varpool_create_empty_node): Break out from ...
19201 (varpool_node_for_decl): ... here.
19202 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
19204 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
19206 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
19207 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
19208 part. Use straight-line flow at the end.
19209 <COMPONENT_REF>: Remove superfluous else.
19210 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
19212 2013-06-12 Jakub Jelinek <jakub@redhat.com>
19215 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
19216 target hook even for !TREE_PUBLIC decls. If no resolution info
19217 is available, return false for common and external decls.
19219 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19221 * config/rl78/constraints.md (U): New constraint.
19222 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
19225 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
19228 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
19229 to allow returning address to AT_PLATFORM name.
19231 2013-06-11 Jan Hubicka <jh@suse.cz>
19233 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
19234 * cgraph.h (symtab_node_base): Add weakref flag.
19235 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
19236 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
19237 (output_weakrefs): Use weakref flag.
19238 * fold-const.c (simple_operand_p): Handle WEAK.
19239 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
19240 * ipa.c (varpool_externally_visible_p): Drop weakref.
19241 (function_and_variable_visibility): Update comment; fix weakref
19242 sanity checks; do not clear DECL_WEAK on them.
19243 * lto-cgraph.c (lto_output_node): update.
19244 (lto_output_varpool_node): Update.
19245 (input_overwrite_node): Update.
19246 (input_node): Update.
19247 (input_varpool_node): Update.
19248 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
19249 (lto_symtab_merge_symbols): Add sanity check.
19250 (lto_symtab_prevailing_decl): Do not special case weakrefs.
19251 * passes.c (rest_of_decl_compilation): Set static flag, too.
19252 * symtab.c (dump_symtab_base): Dump weakref.
19253 (verify_symtab_base): Sanity check weakrefs.
19254 (symtab_make_decl_local): Remove duplicated code.
19255 (symtab_alias_ultimate_target): Simplify.
19256 * varpool.c (varpool_create_variable_alias): Set weakref flag.
19258 2013-06-11 Tom de Vries <tom@codesourcery.com>
19260 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
19261 sequence_vect == NULL.
19263 2013-06-11 DJ Delorie <dj@redhat.com>
19265 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
19266 (rl78_unwind_word_mode): New.
19268 2013-06-11 David Malcolm <dmalcolm@redhat.com>
19270 * final.c (debug_prefix_maps): Make static.
19272 2013-06-11 David Malcolm <dmalcolm@redhat.com>
19274 * function.c (initial_trampoline): Remove stray copy.
19276 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
19278 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
19280 2013-06-11 Martin Jambor <mjambor@suse.cz>
19282 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
19283 within bounds at the beginning of the function.
19285 2013-06-11 Alan Modra <amodra@gmail.com>
19287 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
19289 (get_named_section): Don't NULL !DECL_P decl.
19291 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
19293 * doc/invoke.texi (core-avx2): Document.
19295 (atom): Updated with MOVBE.
19297 2013-06-11 Richard Biener <rguenther@suse.de>
19299 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
19301 2013-06-11 Anton Blanchard <anton@samba.org>
19303 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
19304 correct shift value in little-endian mode.
19306 2013-06-11 Jakub Jelinek <jakub@redhat.com>
19309 * varasm.c (get_variable_align): Move #endif to the right place.
19311 2013-06-10 Cary Coutant <ccoutant@google.com>
19313 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
19314 for hash so that hash table traversal order is deterministic.
19316 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
19317 Pat Haugen <pthaugen@us.ibm.com>
19318 Peter Bergner <bergner@vnet.ibm.com>
19320 * config/rs6000/vector.md (GPR move splitter): Do not split moves
19321 of vectors in GPRS if they are direct moves or quad word load or
19324 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
19326 (direct_move_p): Likewise.
19327 (quad_load_store_p): Likewise.
19329 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
19330 classes into bins based on the physical register type.
19331 (reg_class_to_reg_type): Likewise.
19332 (IS_STD_REG_TYPE): Likewise.
19333 (IS_FP_VECT_REG_TYPE): Likewise.
19334 (reload_fpr_gpr): Arrays to determine what insn to use if we can
19335 use direct move instructions.
19336 (reload_gpr_vsx): Likewise.
19337 (reload_vsx_gpr): Likewise.
19338 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
19339 information that is a simplification of register classes. Also
19340 precalculate direct move reload helpers.
19341 (direct_move_p): New function to return true if the operation can
19342 be done as a direct move instruciton.
19343 (quad_load_store_p): New function to return true if the operation
19344 is a quad memory operation.
19345 (rs6000_legitimize_address): If quad memory, only allow register
19346 indirect for TImode addresses.
19347 (rs6000_legitimate_address_p): Likewise.
19348 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
19349 (rs6000_reload_register_type): Likewise.
19350 (register_to_reg_type): Return register type.
19351 (rs6000_secondary_reload_simple_move): New helper function for
19352 secondary reload and secondary memory needed to identify anything
19353 that is a simple move, and does not need reloading.
19354 (rs6000_secondary_reload_direct_move): New helper function for
19355 secondary reload to identify cases that can be done with several
19356 instructions via the direct move instructions.
19357 (rs6000_secondary_reload_move): New helper function for secondary
19358 reload to identify moves between register types that can be done.
19359 (rs6000_secondary_reload): Add support for quad memory operations
19360 and for direct move.
19361 (rs6000_secondary_memory_needed): Likewise.
19362 (rs6000_debug_secondary_memory_needed): Change argument names.
19363 (rs6000_output_move_128bit): New function to return the move to
19364 use for 128-bit moves, including knowing about the various
19365 limitations of quad memory operations.
19367 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
19368 memory operations. call rs6000_output_move_128bit for the actual
19369 instruciton(s) to generate.
19370 (vsx_movti_64bit): Likewise.
19372 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
19373 (UNSPEC_P8V_MTVSRWZ): Likewise.
19374 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
19375 (UNSPEC_P8V_MTVSRD): Likewise.
19376 (UNSPEC_P8V_XXPERMDI): Likewise.
19377 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
19378 (UNSPEC_FUSION_GPR): Likewise.
19379 (FMOVE128_GPR): New iterator for direct move.
19380 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19381 (f32_sv): Likewise.
19382 (f32_dm): Likewise.
19383 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
19384 loads and direct move instructions.
19385 (zero_extendsidi2_lfiwzx): Likewise.
19386 (extendsidi2_lfiwax): Likewise.
19387 (extendsidi2_nocell): Likewise.
19388 (floatsi<mode>2_lfiwax): Likewise.
19389 (lfiwax): Likewise.
19390 (floatunssi<mode>2_lfiwzx): Likewise.
19391 (lfiwzx): Likewise.
19392 (fix_trunc<mode>_stfiwx): Likewise.
19393 (fixuns_trunc<mode>_stfiwx): Likewise.
19394 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
19395 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
19396 (parity<mode>2_cmpb): Set length/type attr.
19397 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
19398 for 'mr.' to fast_compare.
19399 (bpermd_<mode>): Change type attr to popcnt.
19400 (p8_fmrgow_<mode>): New insns for power8 direct move support.
19401 (p8_mtvsrwz_1): Likewise.
19402 (p8_mtvsrwz_2): Likewise.
19403 (reload_fpr_from_gpr<mode>): Likewise.
19404 (p8_mtvsrd_1): Likewise.
19405 (p8_mtvsrd_2): Likewise.
19406 (p8_xxpermdi_<mode>): Likewise.
19407 (reload_vsx_from_gpr<mode>): Likewise.
19408 (reload_vsx_from_gprsf): Likewise.
19409 (p8_mfvsrd_3_<mode>): LIkewise.
19410 (reload_gpr_from_vsx<mode>): Likewise.
19411 (reload_gpr_from_vsxsf): Likewise.
19412 (p8_mfvsrd_4_disf): Likewise.
19413 (multi-word GPR splits): Do not split direct moves or quad memory
19416 2013-06-10 David Malcolm <dmalcolm@redhat.com>
19418 * tree-into-ssa.c (interesting_blocks): Make static.
19420 2013-06-10 Jakub Jelinek <jakub@redhat.com>
19423 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
19424 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
19425 Use DATA_ABI_ALIGNMENT for that case instead if defined.
19426 (get_variable_align): New function.
19427 (get_variable_section, emit_bss, emit_common,
19428 assemble_variable_contents, place_block_symbol): Use
19429 get_variable_align instead of DECL_ALIGN.
19430 (assemble_noswitch_variable): Add align argument, use it
19431 instead of DECL_ALIGN.
19432 (assemble_variable): Adjust caller. Use get_variable_align
19433 instead of DECL_ALIGN.
19434 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
19436 (DATA_ABI_ALIGNMENT): Define.
19437 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
19438 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
19439 opt is false, only return the psABI mandated alignment increase.
19440 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
19441 (DATA_ABI_ALIGNMENT): ... this.
19442 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
19443 (DATA_ABI_ALIGNMENT): ... this.
19444 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
19445 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
19446 (DATA_ABI_ALIGNMENT): ... this.
19447 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
19448 * doc/tm.texi: Regenerated.
19450 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
19452 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
19453 cmp_code to construct REG_EQUAL note.
19455 2013-06-09 Jakub Jelinek <jakub@redhat.com>
19458 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
19459 that operands[2] doesn't overlap with operands[0].
19461 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
19462 Jan Hubicka <jh@suse.cz>
19464 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
19465 hack to mark symbols as used.
19467 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
19469 PR rtl-optimization/57559
19470 * lra-constraints.c (process_alt_operands): Don't discourage
19471 memory with known offset for offsetable memory constraint.
19472 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
19474 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
19476 * varasm.c (struct oc_local_state): Reorder fields.
19477 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
19478 and adjust accordingly.
19479 (output_constructor): Reorder initialization code and adjust call to
19480 output_constructor_bitfield.
19482 2013-06-07 Jan Hubicka <jh@suse.cz>
19484 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
19486 2013-06-07 David Malcolm <dmalcolm@redhat.com>
19488 * tree-object-size.c (unknown): Make const.
19490 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19492 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
19493 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
19494 for last alternative in the cpu_facility attribute.
19496 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19499 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
19500 (xordi3): Change operand 2 constraint to arm_xordi_operand.
19501 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
19502 * config/arm/constraints.md (Dg): New constraint.
19503 * config/arm/neon.md (xordi3_neon): Remove.
19504 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
19505 * config/arm/predicates.md (arm_xordi_operand): New predicate.
19507 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19509 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
19510 Clean up alternatives.
19512 2013-06-07 Alan Modra <amodra@gmail.com>
19514 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
19517 2013-06-07 Alan Modra <amodra@gmail.com>
19519 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
19521 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19523 * config/arm/constraints.md (Df): New constraint.
19524 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
19525 Correct length attribute for last two alternatives.
19527 2013-06-07 Alan Modra <amodra@gmail.com>
19529 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
19530 override user -mfp-in-toc.
19531 (offsettable_ok_by_alignment): Consider just the current access
19532 rather than the whole object, unless BLKmode. Handle
19533 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
19534 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
19535 for -mcmodel=medium.
19536 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
19537 override user -mfp-in-toc or -msum-in-toc. Default to
19538 -mno-fp-in-toc for -mcmodel=medium.
19540 2013-06-06 DJ Delorie <dj@redhat.com>
19542 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
19543 TARGET_VALID_POINTER_MODE.
19545 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
19546 Pat Haugen <pthaugen@us.ibm.com>
19547 Peter Bergner <bergner@vnet.ibm.com>
19549 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
19550 Document new power8 builtins.
19552 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
19553 condition code register, to allow 128-bit logical operations to be
19554 done in the VSX or GPR registers.
19555 (nor<mode>3): Use the canonical form for nor.
19556 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
19557 vclz*, and vpopcnt* vector instructions.
19558 (nand<mode>3): Likewise.
19559 (orc<mode>3): Likewise.
19560 (clz<mode>2): LIkewise.
19561 (popcount<mode>2): Likewise.
19563 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
19564 that only the GPRs are recognized.
19566 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
19567 support for new power8 builtins.
19569 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
19571 (xscvdpspn): Likewise.
19577 (vpopcnt): Likewise.
19578 (vpopcntb): Likewise.
19579 (vpopcnth): Likewise.
19580 (vpopcntw): Likewise.
19581 (vpopcntd): Likewise.
19583 (vmrgew): Likewise.
19584 (vmrgow): Likewise.
19586 (eqv_v16qi3): Likewise.
19587 (eqv_v8hi3): Likewise.
19588 (eqv_v4si3): Likewise.
19589 (eqv_v2di3): Likewise.
19590 (eqv_v4sf3): Likewise.
19591 (eqv_v2df3): Likewise.
19593 (nand_v16qi3): Likewise.
19594 (nand_v8hi3): Likewise.
19595 (nand_v4si3): Likewise.
19596 (nand_v2di3): Likewise.
19597 (nand_v4sf3): Likewise.
19598 (nand_v2df3): Likewise.
19600 (orc_v16qi3): Likewise.
19601 (orc_v8hi3): Likewise.
19602 (orc_v4si3): Likewise.
19603 (orc_v2di3): Likewise.
19604 (orc_v4sf3): Likewise.
19605 (orc_v2df3): Likewise.
19607 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
19608 allow power8 quad mode in 64-bit.
19609 (rs6000_builtin_vectorized_function): Add support to vectorize
19610 ISA 2.07 count leading zeros, population count builtins.
19611 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
19612 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
19613 (builtin_function_type): Add vgbbd builtin function which takes an
19615 (altivec_expand_vec_perm_const): Add support for new power8 merge
19618 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
19619 that does not include TImdoe for use with 32-bit.
19620 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
19622 (UNSPEC_VSX_CVDPSPN): Likewise.
19623 (vsx_xscvdpspn): Likewise.
19624 (vsx_xscvspdpn): Likewise.
19625 (vsx_xscvdpspn_scalar): Likewise.
19626 (vsx_xscvspdpn_directmove): Likewise.
19627 (vsx_and<mode>3): Split logical operations into 32-bit and
19628 64-bit. Add support to do logical operations on TImode as well as
19629 VSX vector types. Allow logical operations to be done in either
19630 VSX registers or in general purpose registers in 64-bit mode. Add
19631 splitters if GPRs were used. For AND, add clobber of CCmode to
19632 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
19634 (vsx_and<mode>3_32bit): Likewise.
19635 (vsx_and<mode>3_64bit): Likewise.
19636 (vsx_ior<mode>3): Likewise.
19637 (vsx_ior<mode>3_32bit): Likewise.
19638 (vsx_ior<mode>3_64bit): Likewise.
19639 (vsx_xor<mode>3): Likewise.
19640 (vsx_xor<mode>3_32bit): Likewise.
19641 (vsx_xor<mode>3_64bit): Likewise.
19642 (vsx_one_cmpl<mode>2): Likewise.
19643 (vsx_one_cmpl<mode>2_32bit): Likewise.
19644 (vsx_one_cmpl<mode>2_64bit): Likewise.
19645 (vsx_nor<mode>3): Likewise.
19646 (vsx_nor<mode>3_32bit): Likewise.
19647 (vsx_nor<mode>3_64bit): Likewise.
19648 (vsx_andc<mode>3): Likewise.
19649 (vsx_andc<mode>3_32bit): Likewise.
19650 (vsx_andc<mode>3_64bit): Likewise.
19651 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
19652 and xxlorc instructions.
19653 (vsx_eqv<mode>3_64bit): Likewise.
19654 (vsx_nand<mode>3_32bit): Likewise.
19655 (vsx_nand<mode>3_64bit): Likewise.
19656 (vsx_orc<mode>3_32bit): Likewise.
19657 (vsx_orc<mode>3_64bit): Likewise.
19659 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
19661 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
19663 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
19664 (p8_vmrgow): Likewise.
19665 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
19666 GPRs to be split under VSX.
19667 (p8v_clz<mode>2): Add power8 count leading zero support.
19668 (p8v_popcount<mode>2): Add power8 population count support.
19669 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
19672 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
19675 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
19677 (vec_nand): Likewise.
19678 (vec_vclz): Likewise.
19679 (vec_vclzb): Likewise.
19680 (vec_vclzd): Likewise.
19681 (vec_vclzh): Likewise.
19682 (vec_vclzw): Likewise.
19683 (vec_vgbbd): Likewise.
19684 (vec_vmrgew): Likewise.
19685 (vec_vmrgow): Likewise.
19686 (vec_vpopcnt): Likewise.
19687 (vec_vpopcntb): Likewise.
19688 (vec_vpopcntd): Likewise.
19689 (vec_vpopcnth): Likewise.
19690 (vec_vpopcntw): Likewise.
19692 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
19694 PR rtl-optimization/57468
19695 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
19698 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
19700 PR rtl-optimization/57459
19701 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
19702 type when setting live regs.
19704 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
19706 * config/s390/s390.opt (mlra): New option.
19707 * config/s390/s390.c (s390_decompose_address): Check displacement
19708 for all registers for LRA.
19709 (s390_secondary_reload): Don't used secondary reloads for LRA.
19710 (s390_lra_p): New function.
19711 (TARGET_LRA_P): Define.
19712 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
19713 of attribute cpu_facility to zarch for the last alternative.
19714 (*cmpmem_short): Ditto.
19716 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
19718 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
19719 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
19720 (arm_expand_prologue): Likewise.
19722 2013-06-06 Teresa Johnson <tejohnson@google.com>
19725 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
19726 as this is now done by redirect_edge_and_branch_force.
19727 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
19728 barriers, and fix interaction with splitting.
19729 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
19730 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
19731 reflect changes made in the routine.
19732 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
19733 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
19734 since this is called in cfglayout mode, and replace partition fixup
19735 with assert as that is now done by force_nonfallthru_and_redirect.
19736 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
19737 already be marked with region crossing note.
19738 (insert_section_boundary_note): Make non-static, gate on flag
19739 has_bb_partition, rewrite to also check for multiple partitions.
19740 (rest_of_handle_reorder_blocks): Remove call to
19741 insert_section_boundary_note, now done later during free_cfg.
19742 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
19743 * bb-reorder.h (insert_section_boundary_note): Declare.
19744 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
19745 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
19746 invoke insert_section_boundary_note.
19747 (try_redirect_by_replacing_jump): Remove unnecessary
19748 check for region crossing note.
19749 (fixup_partition_crossing): New function.
19750 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
19751 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
19752 in non-cfglayout mode.
19753 (force_nonfallthru_and_redirect): Fixup partition boundaries,
19754 remove old code that tried to do this. Emit barrier correctly
19755 when we are in cfglayout mode.
19756 (last_bb_in_partition): New function.
19757 (rtl_split_edge): Correctly fixup partition boundaries.
19758 (commit_one_edge_insertion): Remove old code that tried to
19759 fixup region crossing edge since this is now handled in
19760 split_block, and set up insertion point correctly since
19761 block may now end in a jump.
19762 (verify_hot_cold_block_grouping): Guard against checking when not in
19763 linearized RTL mode.
19764 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
19766 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
19767 rtl_verify_flow_info, so not called in cfglayout mode.
19768 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
19769 (fixup_reorder_chain): Remove old code that attempted to fixup region
19770 crossing note as this is now handled in force_nonfallthru_and_redirect.
19771 (duplicate_insn_chain): Don't duplicate switch section notes.
19772 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
19774 * basic-block.h (emit_barrier_after_bb): Declare.
19776 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19778 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
19779 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
19780 arm_usatsihi): Adjust alternatives for arm_restrict_it.
19782 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19784 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
19786 * config/arm/ldmstm.md: Regenerate.
19788 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19790 * config/arm/sync.md (atomic_loaddi_1):
19791 Disable predication for arm_restrict_it.
19792 (arm_load_exclusive<mode>): Likewise.
19793 (arm_load_exclusivesi): Likewise.
19794 (arm_load_exclusivedi): Likewise.
19795 (arm_load_acquire_exclusive<mode>): Likewise.
19796 (arm_load_acquire_exclusivesi): Likewise.
19797 (arm_load_acquire_exclusivedi): Likewise.
19798 (arm_store_exclusive<mode>): Likewise.
19799 (arm_store_exclusive<mode>): Likewise.
19800 (arm_store_release_exclusivedi): Likewise.
19801 (arm_store_release_exclusive<mode>): Likewise.
19803 2013-06-06 Richard Biener <rguenther@suse.de>
19805 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
19807 (lto_tag_is_tree_code_p): Adjust.
19808 (lto_tag_is_gimple_code_p): Likewise.
19809 (lto_gimple_code_to_tag): Likewise.
19810 (lto_tag_to_gimple_code): Likewise.
19811 (lto_tree_code_to_tag): Likewise.
19812 (lto_tag_to_tree_code): Likewise.
19813 * data-streamer.h (streamer_write_hwi_in_range): Use
19814 uhwi streaming to stream the normalized range.
19815 (streamer_read_hwi_in_range): Likewise.
19817 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19819 * config/arm/arm.md (enabled_for_depr_it): New attribute.
19820 (predicable_short_it): Likewise.
19821 (predicated): Likewise.
19822 (enabled): Handle above.
19823 (define_cond_exec): Set predicated attribute to yes.
19825 2013-06-05 Mike Stump <mikestump@comcast.net>
19827 * gdbinit.in (__FUNCTION__): Add.
19829 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
19831 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
19832 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
19834 2013-06-05 Jan Hubicka <jh@suse.cz>
19836 * varasm.c (mark_decl_referenced): Revert the removal until targets
19839 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
19841 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
19842 instead of mark_decl_referenced.
19844 2013-06-05 Jan Hubicka <jh@suse.cz>
19846 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
19847 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
19848 and symtab_used_from_object_file_p.
19849 (cgraph_make_node_local_1): Clear forced_by_abi.
19850 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
19851 * cgraph.h (symtab_node_base): Add forced_by_abi.
19852 (decide_is_variable_needed): Remove.
19853 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
19854 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
19855 (decide_is_symbol_needed): ... this one; handle symbols in general;
19856 always analyze virtuals; honnor forced_by_abi.
19857 (cgraph_finalize_function): Update.
19858 (varpool_finalize_decl): Update.
19859 (symbol_defined_and_needed): Remove.
19860 (analyze_functions): Update.
19861 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
19862 output_refs, input_overwrite_node): Handle forced_by_abi.
19863 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
19864 (address_taken_from_non_vtable_p): ... this one.
19865 (comdat_can_be_unshared_p_1): New function.
19866 (cgraph_comdat_can_be_unshared_p): Rename to ...
19867 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
19868 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
19869 (function_and_variable_visibility): Clear forced_by_abi as needed.
19870 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
19871 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
19872 * symtab.c (dump_symtab_base): Dump forced_by_abi.
19873 * varpool.c (decide_is_variable_needed): Remove.
19875 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19877 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
19878 (arm_option_override): Override arm_restrict_it where appropriate.
19879 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
19880 * config/arm/arm.opt (mrestrict-it): New command-line option.
19881 * doc/invoke.texi: Document -mrestrict-it.
19883 2013-06-05 David Malcolm <dmalcolm@redhat.com>
19885 * tsan.c (tsan_atomic_table): Make const.
19887 2013-06-05 Richard Biener <rguenther@suse.de>
19889 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
19890 index associated with the tree we are supposed to replace.
19891 * tree-streamer-out.c (pack_ts_base_value_fields): Output
19892 TREE_ASM_WRITTEN as zero for everything but SSA names.
19894 2013-06-05 David Malcolm <dmalcolm@redhat.com>
19896 * tree-ssa-structalias.c (call_stmt_vars): Make static.
19898 2013-06-04 Jan Hubicka <jh@suse.cz>
19900 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
19901 (input_node, input_varpool_node): Handle correctly external same
19903 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
19904 nodes at ltrans stage.
19906 2013-06-04 Jan Hubicka <jh@suse.cz>
19908 * ipa-inline.c (update_caller_keys): Fix availability test.
19909 (update_callee_keys): Likewise.
19910 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
19911 to follow ELF standard.
19913 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
19915 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
19916 (mips64r5900el-*-elf*): New configurations.
19917 * config/mips/mips-cpus.def (r5900): New processor.
19918 * config/mips/mips-tables.opt: Regenerate.
19919 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
19920 (mips_issue_rate): Handle PROCESSOR_R5900.
19921 (mips_reorg_process_insns): Force reorder mode for the R5900.
19922 * config/mips/mips.h (TARGET_MIPS5900): Define.
19923 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
19925 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
19927 * config/mips/mips.md (processor): Add r5900.
19928 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
19930 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19932 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
19933 into function to generate MOVI instruction.
19934 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
19935 (aarch64_preferred_simd_mode): Turn into wrapper.
19936 (aarch64_output_scalar_simd_mov_immediate): New function.
19937 * config/aarch64/aarch64-protos.h: Add prototype for above.
19939 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19941 * config/aarch64/aarch64.c (simd_immediate_info): Remove
19942 element_char member.
19943 (sizetochar): Return signed char.
19944 (aarch64_simd_valid_immediate): Remove elchar and other
19945 unnecessary variables.
19946 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
19947 Calculate element_char as required.
19948 * config/aarch64/aarch64-protos.h: Update and move prototype
19949 for aarch64_output_simd_mov_immediate.
19950 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
19953 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19955 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
19956 information completed by aarch64_simd_valid_immediate.
19957 (aarch64_legitimate_constant_p): Update arguments.
19958 (aarch64_simd_valid_immediate): Work with struct rather than many
19960 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
19961 (aarch64_simd_make_constant): Update arguments.
19962 (aarch64_output_simd_mov_immediate): Work with struct rather than
19963 many pointers. Output immediate directly rather than as operand.
19964 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
19966 * config/aarch64/constraints.md (Dn): Update arguments.
19968 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19970 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
19972 (aarch64_simd_immediate_valid_for_move): Remove.
19973 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
19974 (aarch64_simd_make_constant): Update call.
19975 (aarch64_output_simd_mov_immediate): Update call.
19976 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
19978 * config/aarch64/constraints.md (Dn): Update call.
19980 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19982 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
19983 return type to bool for prototype.
19984 (aarch64_legitimate_constant_p): Check for true instead of not -1.
19985 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
19986 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
19988 2013-06-04 Catherine Moore <clm@codesourcery.com>
19990 * config/mips/mips.opt (meva): New.
19991 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
19992 (ASM_SPEC): Handle -meva.
19993 * doc/invoke.texi (meva): Document.
19995 2013-06-04 Alan Modra <amodra@gmail.com>
19997 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
20000 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20002 * rtl.def: Add extra fourth optional field to define_cond_exec.
20003 * gensupport.c (process_one_cond_exec): Process attributes from
20005 * doc/md.texi: Document fourth field in define_cond_exec.
20007 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
20009 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
20010 out the processing order as in store_bit_field_1.
20012 2013-06-04 Jan Hubicka <jh@suse.cz>
20014 PR middle-end/57500
20015 * cgraphunit.c (cgraph_process_same_body_aliases): Create
20016 non-VAR_DECL node if it does not exist yet.
20018 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
20020 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
20021 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
20022 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
20023 target_cpu_default setting.
20025 2013-06-03 Teresa Johnson <tejohnson@google.com>
20027 * dumpfile.c (opt_info_switch_p): Change -fopt-info
20028 default to -fopt-info=optimized instead of all.
20029 * doc/invoke.texi: Ditto.
20030 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
20031 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
20032 (execute_vect_slp): Emit BB vectorization success under
20033 MSG_OPTIMIZED_LOCATIONS.
20034 * tree-vect-slp.c (vect_slp_transform_bb): Change
20035 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
20036 * tree-vect-loop.c (vect_transform_loop): Ditto.
20038 2013-06-03 Jason Merrill <jason@redhat.com>
20041 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
20042 Use TARGET_EXPR for C++.
20044 2013-06-03 Jakub Jelinek <jakub@redhat.com>
20046 PR rtl-optimization/57268
20047 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
20048 if DEBUG_INSN_P (insn).
20051 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
20053 PR rtl-optimization/57268
20054 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20055 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20057 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20059 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
20060 (ix86_avoid_lea_for_addr): Likewise.
20061 (exact_dependency_1): Likewise.
20062 (ix86_adjust_cost): Likewise.
20063 (swap_top_of_ready_list): Fix formatting and !reload_completed check
20065 (do_reorder_for_imul): Fix typo, formatting and
20066 !reload_completed check removed.
20067 (ix86_sched_reorder): Fix typo and formatting.
20068 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
20071 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
20073 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
20075 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
20077 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
20078 <STRING_CST>: Likewise.
20079 <VECTOR_CST>: Likewise.
20081 2013-06-01 Janus Weil <janus@gcc.gnu.org>
20082 Mikael Morin <mikael@gcc.gnu.org>
20084 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
20085 * config.in: Regenerated.
20086 * configure: Regenerated.
20088 2013-06-01 Jan Hubicka <jh@suse.cz>
20090 PR middle-end/57366
20091 * cgraphunit.c (compile): When weakref is not supported,
20092 set up transparent aliases before final output pass.
20093 * varasm.c (assemble_alias): Do not try to do it here.
20095 2013-06-01 Jan Hubicka <jh@suse.cz>
20097 PR middle-end/57467
20098 * passes.c (for_per_function): Skip unanalyzed functions.
20100 2013-06-01 Jan Hubicka <jh@suse.cz>
20102 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
20103 (lto_symtab_merge_symbols_1): ... this one.
20104 (lto_symtab_merge_cgraph_nodes): Rename to ...
20105 (lto_symtab_merge_symbols): ... this one; simplify.
20106 * cgraph.c (same_body_aliases_done): Rename to ...
20107 (cpp_implicit_aliases_done): ... this one.
20108 (cgraph_create_function_alias): Update.
20109 (cgraph_same_body_alias): Update.
20110 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
20111 (verify_edge_corresponds_to_fndecl): Simplify.
20112 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
20113 (cgraph_node): Remove same_body_alias.
20114 (varpool_node): Remove alias_of and extra_name_alias.
20115 (same_body_aliases_done): Rename to ..
20116 (cpp_implicit_aliases_done): ... this one.
20117 (symtab_alias_ultimate_target): Add default parameter.
20118 (symtab_resolve_alias): New function.
20119 (fixup_same_cpp_alias_visibility): Declare.
20120 (cgraph_function_node): Add default parameter.
20121 (cgraph_node_asm_name): Likewise.
20122 (cgraph_function_or_thunk_node): Add default parameter; do
20123 not ICE when it is NULL.
20124 (varpool_variable_node): Likewise.
20125 * tree-emutls.c (create_emultls_var): Update.
20126 (ipa_lower_emutls): Update.
20127 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
20128 (cgraph_reset_node): Reset alias info.
20129 (cgraph_finalize_function): Update.
20130 (fixup_same_cpp_alias_visibility): Move to symtab.c.
20131 (analyze_function): Simplify.
20132 (cgraph_process_same_body_aliases): Simplify.
20133 (analyze_functions): Fixup same body aliases.
20134 (handle_alias_pairs): Simplify.
20135 (assemble_thunk): Update.
20136 (assemble_thunks_and_aliases): Update.
20137 (output_weakrefs): Rewrite.
20138 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
20139 (lto_output_varpool_node): Likewise.
20140 (compute_ltrans_boundary): Remve assert.
20141 (get_alias_symbol): New functoin.
20142 (input_node): Rewrite alias handling.
20143 (input_varpool_node): Likewise.
20144 * ipa-pure-const.c (propagate_pure_const): Fix formating.
20145 * ipa.c (process_references): Handle weakrefs correctly.
20146 (symtab_remove_unreachable_nodes): Likewise.
20147 * trans-mem.c (get_cg_data): Update.
20148 (ipa_tm_create_version_alias): Update.
20149 (ipa_tm_execute): Update.
20150 * symtab.c (dump_symtab_base): Dump aliases.
20151 (verify_symtab_base): Verify aliases.
20152 (symtab_node_availability): New function.
20153 (symtab_alias_ultimate_target): Simplify.
20154 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
20155 handle all the fixup cases.
20156 (symtab_resolve_alias): New function.
20157 * passes.c (ipa_write_summaries): Handle weakrefs.
20158 * varpool.c (varpool_analyze_node): Simplify.
20159 (assemble_aliases): Update.
20160 (varpool_create_variable_alias): Simplify.
20161 (varpool_extra_name_alias): Simplify.
20162 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
20163 (lto_symtab_merge_symbols): ... this one.
20165 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
20168 PR rtl-optimization/57268
20169 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20170 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20172 2013-06-01 Tobias Burnus <burnus@net-b.de>
20174 Partially reverted:
20175 2013-05-31 Tobias Burnus <burnus@net-b.de>
20177 PR middle-end/57073
20178 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
20181 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
20183 PR rtl-optimization/57268
20184 * sched-deps.c (sched_analyze_2): Flush dependence lists if
20185 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
20187 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
20189 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
20190 unordered comparison operators when -fno-trapping-math is in effect
20192 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
20193 and implement unordered comparison operators properly on the e500.
20195 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
20197 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
20198 for constant scalar integers.
20199 (simplify_relational_operation_1): Likewise.
20201 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
20203 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
20204 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
20207 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20208 Igor Zamyatin <igor.zamyatin@intel.com>
20210 Silvermont (SLM) architecture performance tuning.
20211 * config/i386/i386.h (enum ix86_tune_indices): Add
20212 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
20213 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
20215 * config/i386/i386.c (initial_ix86_tune_features)
20216 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
20217 (ix86_lea_outperforms): Handle Silvermont tuning.
20218 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
20220 (ix86_use_lea_for_mov): Likewise.
20221 (ix86_avoid_lea_for_addr): Likewise.
20222 (ix86_lea_for_add_ok): Likewise.
20223 (exact_dependency_1): New function.
20224 (exact_store_load_dependency): Likewise.
20225 (ix86_adjust_cost): Handle Silvermont tuning.
20226 (do_reoder_for_imul): Likewise.
20227 (swap_top_of_ready_list): New function.
20228 (ix86_sched_reorder): Changed to handle Silvermont tuning.
20230 * config/i386/i386.md (peepholes that split memory operand in fp
20233 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
20235 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
20236 Remove un-necessary braces.
20238 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
20240 * config/aarch64/aarch64.c (aarch64_classify_symbol):
20241 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
20243 2013-05-31 Tobias Burnus <burnus@net-b.de>
20245 PR middle-end/57073
20246 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
20248 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20251 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
20252 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
20253 * config/arm/neon.md (iordi3_neon): Remove.
20254 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
20255 * config/arm/predicates.md (imm_for_neon_logic_operand):
20256 Move to earlier in the file.
20257 (neon_logic_op2): Likewise.
20258 (arm_iordi_operand_neon): New predicate.
20260 2013-05-31 Richard Biener <rguenther@suse.de>
20262 PR tree-optimization/57478
20263 PR tree-optimization/57453
20264 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
20267 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20269 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
20270 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
20272 2013-05-30 Tobias Burnus <burnus@net-b.de>
20273 Thomas Koenig <tkoenig@gcc.gnu.org>
20275 PR middle-end/57073
20276 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
20277 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
20279 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
20281 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
20283 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
20285 * target.def (register_usage_leveling_p): New hook.
20286 * targhooks.c (default_register_usage_leveling_p): New.
20287 * targhooks.h (default_register_usage_leveling_p): New prototype.
20288 * lra-assigns.c (register_usage_leveling_p): Use the hook.
20289 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
20290 * doc/tm.texi: Update.
20291 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
20293 2013-05-30 Ian Bolton <ian.bolton@arm.com>
20295 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
20296 (*insv_reg<mode>): New define_insn.
20298 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
20300 PR rtl-optimization/57439
20301 * postreload.c (move2add_valid_value_p): Check that we have
20302 a zero subreg_regno_offset when accessing the register in
20303 the requested mode.
20305 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
20306 Igor Zamyatin <igor.zamyatin@intel.com>
20308 Silvermont (SLM) architecture pipeline model, tuning and
20310 * config.gcc: Add slm config options and target.
20312 * config/i386/slm.md: New.
20314 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
20316 * config/i386/i386-c.c (ix86_target_macros_internal): New case
20318 (ix86_target_macros_internal): Likewise.
20320 * config/i386/i386.c (slm_cost): New cost.
20321 (m_SLM): New macro flag.
20322 (initial_ix86_tune_features): Set m_SLM.
20323 (x86_accumulate_outgoing_args): Likewise.
20324 (x86_arch_always_fancy_math_387): Likewise.
20325 (processor_target_table): Add slm cost.
20326 (cpu_names): Add slm cpu name.
20327 (x86_option_override_internal): Set SLM ISA.
20328 (ix86_issue_rate): New case PROCESSOR_SLM.
20329 (ia32_multipass_dfa_lookahead): Likewise.
20330 (fold_builtin_cpu): Add slm.
20332 * config/i386/i386.h (TARGET_SLM): New target macro.
20333 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
20334 (processor_type): Add PROCESSOR_SLM.
20336 * config/i386/i386.md (cpu): Add new value "slm".
20337 (slm.md): Include slm.md.
20339 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
20340 Zhenqiang Chen <zhenqiang.chen@linaro.org>
20342 * config/arm/arm-protos.h: Add and update function protos.
20343 * config/arm/arm.c (use_simple_return_p): New added.
20344 (thumb2_expand_return): Check simple_return flag.
20345 * config/arm/arm.md: Add simple_return and conditional simple_return.
20346 * config/arm/iterators.md: Add iterator for return and simple_return.
20348 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
20350 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
20351 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
20352 (arm_emit_vfp_multi_reg_pop): Likewise.
20353 (thumb2_emit_ldrd_pop): Likewise.
20354 (arm_expand_epilogue): Add misc REG_CFA notes.
20355 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
20357 2013-05-29 Lawrence Crowl <crowl@google.com>
20359 * config/arm/t-arm: Update for below.
20361 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
20362 Change type to hash_table. Update dependent calls and types.
20364 * config/i386/t-cygming: Update for below.
20366 * config/i386/t-interix: Update for below.
20368 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
20369 Change type to hash_table. Update dependent calls and types.
20370 (i386_find_on_wrapper_list::wrappers): Likewise.
20372 * config/ia64/t-ia64: Update for below.
20374 * config/ia64/ia64.c (bundle_state_table):
20375 Change type to hash_table. Update dependent calls and types.
20377 * config/mips/mips.c (mips_reorg_process_insns::htab):
20378 Change type to hash_table. Update dependent calls and types.
20380 * config/sol2.c (solaris_comdat_htab):
20381 Change type to hash_table. Update dependent calls and types.
20383 * config/t-sol2: Update for above.
20385 2013-05-29 Teresa Johnson <tejohnson@google.com>
20387 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
20388 functions are not yet marked as defined.
20390 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
20391 Pat Haugen <pthaugen@us.ibm.com>
20392 Peter Bergner <bergner@vnet.ibm.com>
20394 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
20398 (vrotl<mode>3): Likewise.
20399 (vashl<mode>3): Likewise.
20400 (vlshr<mode>3): Likewise.
20401 (vashr<mode>3): Likewise.
20403 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
20404 support for power8 V2DI builtins.
20406 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
20407 power8 V2DI builtins.
20408 (vupkhsw): Likewise.
20409 (vupklsw): Likewise.
20410 (vaddudm): Likewise.
20411 (vminsd): Likewise.
20412 (vmaxsd): Likewise.
20413 (vminud): Likewise.
20414 (vmaxud): Likewise.
20415 (vpkudum): Likewise.
20416 (vpksdss): Likewise.
20417 (vpkudus): Likewise.
20418 (vpksdus): Likewise.
20423 (vsubudm): Likewise.
20424 (vcmpequd): Likewise.
20425 (vcmpgtsd): Likewise.
20426 (vcmpgtud): Likewise.
20427 (vcmpequd_p): Likewise.
20428 (vcmpgtsd_p): Likewise.
20429 (vcmpgtud_p): Likewise.
20430 (vupkhsw): Likewise.
20431 (vupklsw): Likewise.
20432 (vaddudm): Likewise.
20433 (vmaxsd): Likewise.
20434 (vmaxud): Likewise.
20435 (vminsd): Likewise.
20436 (vminud): Likewise.
20437 (vpksdss): Likewise.
20438 (vpksdus): Likewise.
20439 (vpkudum): Likewise.
20440 (vpkudus): Likewise.
20445 (vsubudm): Likewise.
20447 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
20448 support for power8 V2DI instructions.
20450 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
20451 power8 V2DI instructions. Combine pack and unpack insns to use an
20452 iterator for each mode. Check whether a particular mode supports
20453 Altivec instructions instead of just checking TARGET_ALTIVEC.
20454 (UNSPEC_VPKUWUM): Likewise.
20455 (UNSPEC_VPKSHSS): Likewise.
20456 (UNSPEC_VPKSWSS): Likewise.
20457 (UNSPEC_VPKUHUS): Likewise.
20458 (UNSPEC_VPKSHUS): Likewise.
20459 (UNSPEC_VPKUWUS): Likewise.
20460 (UNSPEC_VPKSWUS): Likewise.
20461 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
20462 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
20463 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
20464 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
20465 (UNSPEC_VUPKHSB): Likewise.
20466 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
20467 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
20468 (UNSPEC_VUPKHSH): Likewise.
20469 (UNSPEC_VUPKLSB): Likewise.
20470 (UNSPEC_VUPKLSH): Likewise.
20472 (VI_char): Likewise.
20473 (VI_scalar): Likewise.
20474 (VI_unit): Likewise.
20476 (VP_small): Likewise.
20477 (VP_small_lc): Likewise.
20478 (VU_char): Likewise.
20479 (add<mode>3): Likewise.
20480 (altivec_vaddcuw): Likewise.
20481 (altivec_vaddu<VI_char>s): Likewise.
20482 (altivec_vadds<VI_char>s): Likewise.
20483 (sub<mode>3): Likewise.
20484 (altivec_vsubcuw): Likewise.
20485 (altivec_vsubu<VI_char>s): Likewise.
20486 (altivec_vsubs<VI_char>s): Likewise.
20487 (altivec_vavgs<VI_char>): Likewise.
20488 (altivec_vcmpbfp): Likewise.
20489 (altivec_eq<mode>): Likewise.
20490 (altivec_gt<mode>): Likewise.
20491 (altivec_gtu<mode>): Likewise.
20492 (umax<mode>3): Likewise.
20493 (smax<mode>3): Likewise.
20494 (umin<mode>3): Likewise.
20495 (smin<mode>3): Likewise.
20496 (altivec_vpkuhum): Likewise.
20497 (altivec_vpkuwum): Likewise.
20498 (altivec_vpkshss): Likewise.
20499 (altivec_vpkswss): Likewise.
20500 (altivec_vpkuhus): Likewise.
20501 (altivec_vpkshus): Likewise.
20502 (altivec_vpkuwus): Likewise.
20503 (altivec_vpkswus): Likewise.
20504 (altivec_vpks<VI_char>ss): Likewise.
20505 (altivec_vpks<VI_char>us): Likewise.
20506 (altivec_vpku<VI_char>us): Likewise.
20507 (altivec_vpku<VI_char>um): Likewise.
20508 (altivec_vrl<VI_char>): Likewise.
20509 (altivec_vsl<VI_char>): Likewise.
20510 (altivec_vsr<VI_char>): Likewise.
20511 (altivec_vsra<VI_char>): Likewise.
20512 (altivec_vsldoi_<mode>): Likewise.
20513 (altivec_vupkhsb): Likewise.
20514 (altivec_vupkhs<VU_char>): Likewise.
20515 (altivec_vupkls<VU_char>): Likewise.
20516 (altivec_vupkhsh): Likewise.
20517 (altivec_vupklsb): Likewise.
20518 (altivec_vupklsh): Likewise.
20519 (altivec_vcmpequ<VI_char>_p): Likewise.
20520 (altivec_vcmpgts<VI_char>_p): Likewise.
20521 (altivec_vcmpgtu<VI_char>_p): Likewise.
20522 (abs<mode>2): Likewise.
20523 (vec_unpacks_hi_v16qi): Likewise.
20524 (vec_unpacks_hi_v8hi): Likewise.
20525 (vec_unpacks_lo_v16qi): Likewise.
20526 (vec_unpacks_hi_<VP_small_lc>): Likewise.
20527 (vec_unpacks_lo_v8hi): Likewise.
20528 (vec_unpacks_lo_<VP_small_lc>): Likewise.
20529 (vec_pack_trunc_v8h): Likewise.
20530 (vec_pack_trunc_v4si): Likewise.
20531 (vec_pack_trunc_<mode>): Likewise.
20533 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
20535 (vec_vmaxsd): Likewise.
20536 (vec_vmaxud): Likewise.
20537 (vec_vminsd): Likewise.
20538 (vec_vminud): Likewise.
20539 (vec_vpksdss): Likewise.
20540 (vec_vpksdus): Likewise.
20541 (vec_vpkudum): Likewise.
20542 (vec_vpkudus): Likewise.
20543 (vec_vrld): Likewise.
20544 (vec_vsld): Likewise.
20545 (vec_vsrad): Likewise.
20546 (vec_vsrd): Likewise.
20547 (vec_vsubudm): Likewise.
20548 (vec_vupkhsw): Likewise.
20549 (vec_vupklsw): Likewise.
20551 2013-05-29 Jan Hubicka <jh@suse.cz>
20553 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
20554 flags; reorder rest of fields in more consistent way.
20555 (varpool_node): Remove analyzed, finalized and alias.
20556 (cgraph_ndoe): Likewise.
20557 (symtab_alias_ultimate_target): New function.
20558 (cgraph_function_node): Move offline.
20559 (cgraph_reset_node): Declare.
20560 (cgraph_comdat_can_be_unshared_p): Remove.
20561 (varpool_remove_initializer): Declare.
20562 (varpool_first_defined_variable, varpool_next_defined_variable
20563 cgraph_first_defined_function, cgraph_next_defined_function): Update.
20564 (cgraph_function_with_gimple_body_p): Update.
20565 (varpool_all_refs_explicit_p): Update.
20566 (symtab_alias_target): New function.
20567 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
20568 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
20569 (cgraph_function_or_thunk_node): Simplify using
20570 symtab_alias_ultimate_target.
20571 (varpool_variable_node): Likewise.
20572 * cgraph.c (cgraph_create_function_alias): Update.
20573 (cgraph_add_thunk): Update.
20574 (cgraph_remove_node): Update.
20575 (dump_cgraph_node): Do not dump removed flags.
20576 (cgraph_function_body_availability): Update.
20577 (cgraph_propagate_frequency): Update.
20578 (verify_cgraph_node): Check sanity of local flag.
20579 (cgraph_function_node): Move here from cgraph.h; revamp for
20580 cgraph_function_or_thunk_node.
20581 * lto-symtab.c (lto_varpool_replace_node): Update.
20582 (lto_symtab_resolve_can_prevail_p): Update.
20583 (lto_symtab_merge_cgraph_nodes): Update.
20584 * ipa-cp.c (determine_versionability, initialize_node_lattices,
20585 propagate_constants_accross_call, devirtualization_time_bonus,
20586 ipcp_propagate_stage): Update.
20587 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
20588 * ipa-inline-transform.c (clone_inlined_nodes,
20589 preserve_function_body_p): Update.
20590 * ipa-reference.c (propagate): Update.
20591 (write_node_summary_p): Update.
20592 * toplev.c (wrapup_global_declaration_2): Update.
20593 * cgraphunit.c (cgraph_analyze_function): Rename to ...
20594 (analyze_function) ... this one.
20595 (cgraph_process_new_functions): Update.
20596 (cgraph_reset_node): Export.
20597 (cgraph_finalize_function): Update.
20598 (cgraph_add_new_function): Update.
20599 (process_function_and_variable_attributes): Update.
20600 (varpool_finalize_decl): Update.
20601 (symbol_finalized): Remove.
20602 (symbol_finalized_and_needed): Rename to ...
20603 (symbol_defined_and_needed): ... update.
20604 (cgraph_analyze_functions): Update.
20605 (handle_alias_pairs): Update.
20606 (mark_functions_to_output): Update.
20607 (assemble_thunk): Update.
20608 (output_in_order): Update.
20609 (output_weakrefs): Update.
20610 (finalize_compilation_unit): Update.
20611 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
20612 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
20613 input_node, input_varpool_node): Update.
20614 * dbxout.c (dbxout_expand_expr): Update.
20615 * cgraphclones.c (cgraph_clone_node): Update.
20616 (cgraph_copy_node_for_versioning): Update.
20617 (cgraph_materialize_clone): Update.
20618 (cgraph_materialize_all_clones): Update.
20619 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
20620 propagate_pure_const, propagate_nothrow): Update.
20621 * lto-streamer-out.c (lto_output, write_symbol): Update.
20622 * ipa-utils.c (ipa_reverse_postorder): Update.
20623 * ipa-inline.c (can_inline_edge_p): Update.
20624 (update_caller_keys, ipa_inline): Update.
20625 * dwarf2out.c (reference_to_unused,
20626 premark_types_used_by_global_vars_helper): Update.
20627 * tree-eh.c (tree_could_trap_p): Update.
20628 * ipa-split.c (consider_split, execute_split_functions): Update.
20629 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
20630 has_addr_references_p): Update; move ahead in file for better
20632 (process_references): Simplify.
20633 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
20634 bodies are removed.
20635 (cgraph_comdat_can_be_unshared_p): Make static.
20636 (cgraph_externally_visible_p): Update.
20637 (varpool_externally_visible_p): Update.
20638 (function_and_variable_visibility): Update.
20639 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
20640 ipa_tm_mark_force_output_node): Update.
20641 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
20642 estimate_edge_devirt_benefit, inline_generate_summary,
20643 inline_write_summary): Update.
20644 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
20645 * ipa-prop.c (ipa_compute_jump_functions): Update.
20646 (ipa_print_node_params, ipa_prop_read_section,
20647 ipa_update_after_lto_read, read_replacements_section): Update.
20648 * varasm.c (mark_decl_referenced): Update.
20649 (assemble_alias, dump_tm_clone_pairs): Update.
20650 * tree-inline.c (copy_bb): Update.
20651 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
20653 * symtab.c (dump_symtab_base): Print new flags.
20654 (verify_symtab_base): Verify new flags.
20655 (symtab_alias_ultimate_target): New function.
20656 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
20657 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
20659 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
20661 * i386.c (ix86_get_function_versions_dispatcher,
20662 ix86_generate_version_dispatcher_body): Update.
20663 (fold_builtin_cpu): Use varpool_add_new_variable.
20664 * varpool.c (varpool_remove_initializer): Break out from ...
20665 (varpool_remove_node): ... this one.
20666 (dump_varpool_node, varpool_node_for_asm,
20667 cgraph_variable_initializer_availability, varpool_analyze_node,
20668 varpool_assemble_decl, varpool_remove_unreferenced_decls,
20669 varpool_finalize_named_section_flags, varpool_create_variable_alias):
20672 2013-05-29 Jan Hubicka <jh@suse.cz>
20674 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
20676 2013-05-29 Easwaran Raman <eraman@google.com>
20678 PR tree-optimization/57442
20679 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
20680 when control exits the main loop.
20682 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
20684 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
20686 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
20687 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
20688 * rx/t-rx: Add rx100 under multi library matches option for nofpu
20691 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20693 PR tree-optimization/57441
20694 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
20695 Don't limit size of incr_vec to number of candidates.
20697 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
20699 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
20700 and mips16 directories.
20701 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
20702 (MULTILIB_DIRNAMES): Ditto.
20703 (MULTILIB_EXCEPTIONS): Add new exceptions.
20704 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
20705 (MULTILIB_DIRNAMES): Ditto.
20706 (MULTILIB_EXCEPTIONS): Add new exceptions.
20708 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20709 Marcus Shawcroft <marcus.shawcroft@arm.com>
20711 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
20712 SYMBOL_TINY_ABSOLUTE.
20713 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
20714 SYMBOL_TINY_ABSOLUTE.
20715 (aarch64_expand_mov_immediate): Likewise.
20716 (aarch64_classify_symbol): Likewise.
20717 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
20718 Permit SYMBOL_TINY_ABSOLUTE.
20719 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
20721 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20722 Marcus Shawcroft <marcus.shawcroft@arm.com>
20724 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
20725 Refactor if/switch. Replace gcc_assert with if.
20727 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20729 * config/i386/i386.c (initial_ix86_tune_features): Enable
20730 FP Reassociation for AMD bdver1 and bdver2.
20732 2013-05-29 Martin Jambor <mjambor@suse.cz>
20734 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
20735 and IMAGPART_EXPR do not occur within other handled_components.
20737 2013-05-29 Richard Biener <rguenther@suse.de>
20739 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
20740 access on whether the use is in the BB we currently try to
20742 (vect_bb_vectorization_profitable_p): Pass the BB we currently
20743 vectorize to vect_bb_slp_scalar_cost.
20745 2013-05-29 Richard Biener <rguenther@suse.de>
20747 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
20748 computing scalar cost offsetted by stmts that are kept live
20750 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
20751 for computation of scalar cost.
20753 2013-05-28 Steve Ellcey <sellcey@mips.com>
20755 * config/mips/mips-cpus.def (mips32r2): Change processor type.
20757 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
20759 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
20760 array notation built-in reduction functions.
20761 * doc/passes.texi (Passes): Added documentation about changes done
20763 * doc/invoke.texi (C Dialect Options): Added documentation about
20764 the -fcilkplus flag.
20765 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
20766 (BUILTINS_DEF): Depend on cilkplus.def.
20767 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
20768 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
20769 * cilkplus.def: New file.
20771 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
20773 PR rtl-optimization/57439
20774 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
20776 2013-05-28 Easwaran Raman <eraman@google.com>
20778 PR tree-optimization/57337
20779 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
20780 (find_insert_point): Correctly identify the insertion point
20781 when two statements with the same UID is compared.
20783 2013-05-28 Richard Biener <rguenther@suse.de>
20785 PR tree-optimization/56787
20786 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
20787 from the list of data references.
20788 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
20790 (vect_analyze_loop_operations): Likewise.
20791 (vect_transform_loop): Remove clobbers.
20793 2013-05-28 Martin Jambor <mjambor@suse.cz>
20795 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
20796 and REALPART_EXPRs have scalar type.
20798 2013-05-28 Richard Biener <rguenther@suse.de>
20800 PR tree-optimization/57411
20801 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
20803 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
20804 virtual operand propagation.
20806 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
20808 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
20809 destination register for bmasksi_vis.
20810 (vector_init_bshuffle): Likewise.
20811 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
20813 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
20815 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
20816 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
20817 mode if the instruction isn't available in the original mode.
20818 * config/sparc/sparc.opt (mfix-ut699): New option.
20819 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
20820 (divdf3): Turn into expander.
20821 (divdf3_nofix): New insn.
20822 (divdf3_fix): Likewise.
20823 (divsf3): Disable if -mfix-ut699.
20824 (sqrtdf2): Turn into expander.
20825 (sqrtdf2_nofix): New insn.
20826 (sqrtdf2_fix): Likewise.
20827 (sqrtsf2): Disable if -mfix-ut699.
20829 2013-05-27 Richard Biener <rguenther@suse.de>
20831 PR middle-end/57412
20832 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
20833 block for the new loop.
20835 2013-05-27 Richard Biener <rguenther@suse.de>
20837 PR tree-optimization/57343
20838 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
20839 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
20840 (number_of_iterations_cond): Do not build the folded tree.
20842 2013-05-27 Richard Biener <rguenther@suse.de>
20845 PR middle-end/57381
20846 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
20847 OEP_CONSTANT_ADDRESS_OF retained.
20849 PR tree-optimization/57417
20850 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
20851 for unchanged base.
20852 (set_ssa_val_to): Compare addresses using
20853 get_addr_base_and_unit_offset.
20855 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
20857 PR rtl-optimization/56833
20858 * postreload.c (move2add_record_mode): New function.
20859 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
20860 (move2add_use_add2_insn): Use move2add_record_sym_value.
20861 (move2add_use_add3_insn): Likewise.
20862 (reload_cse_move2add): Use move2add_valid_value_p and
20863 move2add_record_mode. Invalidate call-clobbered and REG_INC
20864 affected regs by setting reg_mode to VOIDmode.
20865 (move2add_note_store): Don't pretend the inside of a SUBREG is
20866 the actual destination. Invalidate single/leading registers by
20867 setting reg_mode to VOIDmode.
20868 Use move2add_record_sym_value, move2add_valid_value_p and
20869 move2add_record_mode.
20871 2013-05-27 Richard Biener <rguenther@suse.de>
20873 PR tree-optimization/57396
20874 * tree-affine.c (double_int_constant_multiple_p): Properly
20875 return false for val == 0 and div != 0.
20877 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
20879 * config/mips/mips.h: Use #elif in preprocessor conditions.
20881 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
20884 * config/mips/constraints.md (kl): New constraint.
20885 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
20886 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
20887 constraint for operand 0. Split after CSE for MIPS16. Emit a move
20888 from LO for MIPS16.
20889 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
20891 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
20894 * config/mips/mips.c (mips_can_inline_p): New function.
20895 (TARGET_CAN_INLINE_P): Define.
20897 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
20899 * sched-int.h (ds_t, dw_t): Make unsigned int.
20900 Fix documentation that describes how all the ds_t bits are used.
20901 Reserve the last bit for delayed-branch scheduling.
20902 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
20903 (BITS_PER_DEP_WEAK): Fix definition and documentation.
20904 (gen_dep_weak_1): Remove prototype.
20905 * sched-deps.c (get_dep_weak_1): Make static.
20906 * target.def (speculate_insn, needs_block_p, gen_spec_check,
20907 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
20908 * doc/tm.texi: Regenerate.
20909 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
20911 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
20914 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
20916 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
20917 Sandra Loosemore <sandra@codesourcery.com>
20919 * config.gcc (powerpc-*): Allow native for with-cpu.
20921 2013-05-24 Jeff Law <law@redhat.com>
20923 PR tree-optimization/57124
20924 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
20925 conversion feeding a condition if the range has an overflow
20926 if -fstrict-overflow. Add warnings for when we do make the
20929 2013-05-24 Dehao Chen <dehao@google.com>
20931 * tree-cfg.c (locus_discrim_map): Fix the typo.
20932 (locus_discrim_hasher): Likewise.
20933 (locus_discrim_hasher::hash): Likewise.
20934 (locus_discrim_hasher::equal): Likewise.
20936 2013-05-24 Martin Jambor <mjambor@suse.cz>
20938 PR tree-optimization/57294
20939 * cgraph.h (ipa_record_stmt_references): Declare.
20940 * cgraphbuild.c (ipa_record_stmt_references): New function.
20941 (build_cgraph_edges): Use ipa_record_stmt_references.
20942 (rebuild_cgraph_edges): Likewise.
20943 (cgraph_rebuild_references): Likewise.
20944 * ipa-prop.c (ipa_modify_call_arguments): Discard references
20945 associated with the old statement and build references from the
20946 newly built statements.
20947 * ipa-ref.c (ipa_remove_stmt_references): New function.
20948 * ipa-ref.h (ipa_remove_stmt_references): Declare.
20950 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
20952 * lra-constraints.c (emit_spill_move): Use smaller mode for
20954 (check_and_process_move): Consider mem-reg moves for secondary
20956 (curr_insn_transform): Don't lose insns emitted before for
20957 secondary memory moves.
20958 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
20959 reg set up in the current insn.
20961 2013-05-24 Dehao Chen <dehao@google.com>
20963 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
20965 (locus_descrim_hasher::equal): Likewise.
20966 (build_gimple_cfg): New discriminator assignment algorithm.
20967 (make_edges): Likewise.
20968 (next_discriminator_for_locus): Likewise.
20969 (same_line_p): Likewise.
20970 (assign_discriminators): Likewise.
20971 (make_cond_expr_edges): Likewise.
20972 (make_gimple_switch_edges): Likewise.
20973 (make_goto_expr_edges): Likewise.
20974 (make_gimple_asm_edges): Likewise.
20976 2013-05-24 Ian Bolton <ian.bolton@arm.com>
20978 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
20979 X format specifier to only display bottom 16 bits.
20980 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
20981 immediate to match for operand 2, since it will be masked.
20983 2013-05-24 Richard Biener <rguenther@suse.de>
20985 PR tree-optimization/57287
20986 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
20987 all SSA names that occur in abnormal PHIs.
20989 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
20992 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
20993 that index is not negative.
20995 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
20997 PR rtl-optimization/55177
20998 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
20999 (simplify_byte_swapping_operation): New.
21000 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
21001 (simplify_relational_operation_1): Deal with BSWAP.
21003 2013-05-23 Richard Henderson <rth@redhat.com>
21006 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
21007 (ix86_reorg): Call it.
21009 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
21012 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
21013 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
21014 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
21016 2013-05-23 Christian Bruel <christian.bruel@st.com>
21019 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
21021 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21022 Marcus Shawcroft <marcus.shawcroft@arm.com>
21024 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
21025 * config/aarch64/constraints.md (Usa): Remove.
21026 * doc/md.texi (AArch64 Usa): Remove.
21028 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
21029 Marcus Shawcroft <marcus.shawcroft@arm.com>
21031 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
21032 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
21033 * config/aarch64/predicates.md (aarch64_const_address): Remove.
21034 (aarch64_mov_operand): Use aarch64_mov_operand_p.
21036 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
21038 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
21039 instruction (AdvSIMD).
21040 * config/aarch64/aarch64-builtins.c
21041 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
21042 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
21044 2013-05-23 Martin Jambor <mjambor@suse.cz>
21046 PR middle-end/57347
21047 * tree.h (contains_bitfld_component_ref_p): Declare.
21048 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
21049 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
21051 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
21052 not access a bit-field. Assert all final offsets are byte-aligned.
21054 2013-05-23 Richard Biener <rguenther@suse.de>
21056 PR tree-optimization/57380
21057 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
21058 least one invariant or re-used load.
21059 * passes.c (init_optimization_passes): Move pass_phiprop before
21062 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
21064 * config/aarch64/aarch64-simd.md
21065 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
21067 2013-05-23 Richard Biener <rguenther@suse.de>
21069 PR middle-end/57381
21070 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
21071 OEP_CONSTANT_ADDRESS_OF retained.
21073 2013-05-23 Jakub Jelinek <jakub@redhat.com>
21075 PR middle-end/57344
21076 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
21077 don't lower unit. Handle unit not being always BITS_PER_WORD.
21079 2013-05-23 Richard Biener <rguenther@suse.de>
21081 PR rtl-optimization/57341
21082 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
21083 instead of true_dependence.
21085 2013-05-22 David Malcolm <dmalcolm@redhat.com>
21087 * bb-reorder.c (branch_threshold): Make const.
21088 (exec_threshold): Ditto.
21090 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
21091 Pat Haugen <pthaugen@us.ibm.com>
21092 Peter Bergner <bergner@vnet.ibm.com>
21094 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
21095 documentation for the power8 crypto builtins.
21097 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
21099 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
21100 macros for defining power8 builtin functions.
21101 (BU_P8V_AV_2): Likewise.
21102 (BU_P8V_AV_P): Likewise.
21103 (BU_P8V_VSX_1): Likewise.
21104 (BU_P8V_OVERLOAD_1): Likewise.
21105 (BU_P8V_OVERLOAD_2): Likewise.
21106 (BU_CRYPTO_1): Likewise.
21107 (BU_CRYPTO_2): Likewise.
21108 (BU_CRYPTO_3): Likewise.
21109 (BU_CRYPTO_OVERLOAD_1): Likewise.
21110 (BU_CRYPTO_OVERLOAD_2): Likewise.
21111 (XSCVSPDP): Fix typo, point to the correct instruction.
21112 (VCIPHER): Add power8 crypto builtins.
21113 (VCIPHERLAST): Likewise.
21114 (VNCIPHER): Likewise.
21115 (VNCIPHERLAST): Likewise.
21116 (VPMSUMB): Likewise.
21117 (VPMSUMH): Likewise.
21118 (VPMSUMW): Likewise.
21119 (VPERMXOR_V2DI): Likewise.
21120 (VPERMXOR_V4SI: Likewise.
21121 (VPERMXOR_V8HI: Likewise.
21122 (VPERMXOR_V16QI: Likewise.
21123 (VSHASIGMAW): Likewise.
21124 (VSHASIGMAD): Likewise.
21125 (VPMSUM): Likewise.
21126 (VPERMXOR): Likewise.
21127 (VSHASIGMA): Likewise.
21129 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
21130 __CRYPTO__ if the crypto instructions are available.
21131 (altivec_overloaded_builtins): Add support for overloaded power8
21134 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
21135 support for power8 crypto builtins.
21136 (builtin_function_type): Likewise.
21137 (altivec_init_builtins): Add support for builtins that take vector
21138 long long (V2DI) arguments.
21140 * config/rs6000/crypto.md: New file, define power8 crypto
21143 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
21144 Pat Haugen <pthaugen@us.ibm.com>
21145 Peter Bergner <bergner@vnet.ibm.com>
21147 * doc/invoke.texi (Option Summary): Add power8 options.
21148 (RS/6000 and PowerPC Options): Likewise.
21150 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
21151 constraints.md instead of rs6000.h. Reorder w* constraints. Add
21152 wm, wn, wr documentation.
21154 * config/rs6000/constraints.md (wm): New constraint for VSX
21155 registers if direct move instructions are enabled.
21156 (wn): New constraint for no registers.
21157 (wq): New constraint for quad word even GPR registers.
21158 (wr): New constraint if 64-bit instructions are enabled.
21159 (wv): New constraint if power8 vector instructions are enabled.
21160 (wQ): New constraint for quad word memory locations.
21162 * config/rs6000/predicates.md (const_0_to_15_operand): New
21163 constraint for 0..15 for crypto instructions.
21164 (gpc_reg_operand): If VSX allow registers in VSX registers as well
21165 as GPR and floating point registers.
21166 (int_reg_operand): New predicate to match only GPR registers.
21167 (base_reg_operand): New predicate to match base registers.
21168 (quad_int_reg_operand): New predicate to match even GPR registers
21169 for quad memory operations.
21170 (vsx_reg_or_cint_operand): New predicate to allow vector logical
21171 operations in both GPR and VSX registers.
21172 (quad_memory_operand): New predicate for quad memory operations.
21173 (reg_or_indexed_operand): New predicate for direct move support.
21175 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
21176 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
21177 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
21178 (POWERPC_MASKS): Add power8 options.
21179 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
21182 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
21183 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
21185 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
21186 (-mpower8-fusion): New power8 options.
21187 (-mpower8-fusion-sign): Likewise.
21188 (-mpower8-vector): Likewise.
21189 (-mcrypto): Likewise.
21190 (-mdirect-move): Likewise.
21191 (-mquad-memory): Likewise.
21193 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
21194 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
21196 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
21197 (rs6000_debug_vector_unit): Add p8_vector.
21198 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
21199 definitions. Also print fusion state.
21200 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
21201 (rs6000_builtin_mask_calculate): Add power8 builtin support.
21202 (rs6000_option_override_internal): Add support for power8.
21203 (rs6000_common_init_builtins): Add debugging for skipped builtins
21204 if -mdebug=builtin.
21205 (rs6000_adjust_cost): Add power8 support.
21206 (rs6000_issue_rate): Likewise.
21207 (insn_must_be_first_in_group): Likewise.
21208 (insn_must_be_last_in_group): Likewise.
21209 (force_new_group): Likewise.
21210 (rs6000_register_move_cost): Likewise.
21211 (rs6000_opt_masks): Likewise.
21213 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
21214 power8 capable assembler, default to power7 options.
21215 (TARGET_DIRECT_MOVE): Likewise.
21216 (TARGET_CRYPTO): Likewise.
21217 (TARGET_P8_VECTOR): Likewise.
21218 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
21219 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
21220 (VECTOR_MEM_P8_VECTOR_P): Likewise.
21221 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
21222 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
21223 (TARGET_XSCVDPSPN): Likewise.
21224 (TARGET_XSCVSPDPN): Likewsie.
21225 (TARGET_SYNC_HI_QI): Likewise.
21226 (TARGET_SYNC_TI): Likewise.
21227 (MASK_CRYPTO): Likewise.
21228 (MASK_DIRECT_MOVE): Likewise.
21229 (MASK_P8_FUSION): Likewise.
21230 (MASK_P8_VECTOR): Likewise.
21231 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
21232 temporary used by some of the direct move instructions to get two FP
21233 temporary registers does not force creation of a stack frame.
21234 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
21235 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
21236 that any VSX registers are tieable, even if they are also an
21237 Altivec vector mode.
21238 (r6000_reg_class_enum): Add wm, wr, wv constraints.
21239 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
21240 (RS6000_BTM_CRYPTO): Likewise.
21241 (RS6000_BTM_COMMON): Likewise.
21243 * config/rs6000/rs6000.md (cpu attribute): Add power8.
21244 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
21245 (enum rs6000_vector): Add power8 vector support.
21247 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21251 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
21252 (any_sibcall_could_use_r3): this and handle indirect calls.
21253 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
21255 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21257 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
21259 2013-05-22 Richard Biener <rguenther@suse.de>
21261 PR middle-end/57349
21262 * profile.c (branch_prob): Do not split blocks that are
21263 abnormally receiving from ECF_RETURNS_TWICE functions.
21265 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21267 * recog.c (offsettable_address_addr_space_p): Fix calculation of
21268 address mode. Move pointer mode initialization to the same place.
21270 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
21272 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
21273 while it has any effect.
21275 2013-05-21 Easwaran Raman <eraman@google.com>
21277 PR tree-optimization/57322
21278 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
21279 UID of the statement added to the BB to be 1.
21281 2013-05-21 Jakub Jelinek <jakub@redhat.com>
21283 PR tree-optimization/57331
21284 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
21285 of conversion from pointer type to integral type with integer.
21287 2013-05-21 Martin Jambor <mjambor@suse.cz>
21290 * ipa-prop.c (ipa_read_node_info): Process param_used and
21291 controlled_uses in the same order as when writing.
21293 2013-05-21 Magnus Granberg <baldrick@free.fr>
21296 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
21298 2013-05-21 Richard Biener <rguenther@suse.de>
21300 PR tree-optimization/57318
21301 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
21302 estimate stmts with side-effects as likely eliminated.
21304 2013-05-21 Richard Biener <rguenther@suse.de>
21306 PR tree-optimization/57330
21307 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
21308 preserve the call stmts fntype.
21310 2013-05-21 Richard Biener <rguenther@suse.de>
21312 PR tree-optimization/57303
21313 * tree-ssa-sink.c (statement_sink_location): Improve killing
21314 stmt detection and properly handle self-assignments.
21316 2013-05-21 Christian Bruel <christian.bruel@st.com>
21318 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
21319 spanning registers. LEAF_REG_REMAP is supported only for contiguous
21320 registers. Set register size out of the PARALLEL loop.
21322 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
21325 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
21326 (*fmasf4, *fmasf4_media): New insns.
21328 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21330 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
21331 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
21332 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
21333 (mips_idiv_insns): Update the comments to say that the returned
21334 instruction counts are in units of BASE_INSN_LENGTH.
21335 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
21336 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
21337 using 2 rather than 4 as the length of indirect MIPS16 and
21338 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
21339 length of a NOP. Don't divide MIPS16 lengths by 2.
21340 (mips16_split_long_branches): Assume a branch is long if the
21341 length is greater than 4 rather than 8.
21342 * config/mips/mips.md (length): Give MIPS16 lengths directly,
21343 rather than multiplying them by 2. Multiply instruction counts
21344 by BASE_INSN_LENGTH rather than 4.
21345 (*jump_mips16, tls_get_tp_mips16_<mode>)
21346 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
21348 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21350 * config/mips/mips.md (extended_mips16): Remove branch case.
21351 (length): Remove duplicated extended_mips16 test.
21353 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
21355 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
21357 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
21359 * recog.h (Recog_data): Rename to...
21360 (recog_data_d): ...this.
21361 (recog_data): Update accordingly.
21362 * recog.c (recog_data): Likewise.
21363 * reload.c (save_recog_data): Likewise.
21364 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
21365 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
21367 2013-05-17 Julian Brown <julian@codesourcery.com>
21369 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
21370 found in a REG_EQUAL note, invalidate it.
21372 2013-05-17 Easwaran Raman <eraman@google.com>
21374 * tree-ssa-reassoc.c (find_insert_point): New function.
21375 (insert_stmt_after): Likewise.
21376 (get_def_stmt): Likewise.
21377 (ensure_ops_are_available): Likewise.
21378 (not_dominated_by): Likewise.
21379 (rewrite_expr_tree): Do not move statements beyond what is
21380 necessary. Remove call to swap_ops_for_binary_stmt...
21381 (reassociate_bb): ... and move it here.
21382 (build_and_add_sum): Assign UIDs for new statements.
21383 (linearize_expr): Likewise.
21384 (do_reassoc): Renumber gimple statement UIDs.
21386 2013-05-17 Jan Hubicka <jh@suse.cz>
21388 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
21390 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
21391 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
21393 (output_weakrefs): Update.
21395 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
21396 Martin Jambor <mjambor@suse.cz>
21398 PR middle-end/57276
21399 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
21400 value that corresponds to the given aggval is found in values vector.
21402 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
21404 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
21405 sse, sse2, sse3, ssse3 and sse4a flags to options.
21407 2013-05-17 David Malcolm <dmalcolm@redhat.com>
21409 * gengtype-state.c: (s_expr_writer): New class, to handle
21410 prettifying of output layout of s-expressions.
21411 (state_writer): New class, to write out gtype.state.
21412 (state_written_type_count): Move this variable into member data of
21414 (s_expr_writer::s_expr_writer): New code: constructor for new class
21415 (state_writer::state_writer(): ditto
21416 (s_expr_writer::write_new_line): New function
21417 (s_expr_writer::write_any_indent): ditto
21418 (s_expr_writer::begin_s_expr): ditto
21419 (s_expr_writer::end_s_expr): ditto
21420 (write_state_fileloc): convert to method of state_writer...
21421 (state_writer:: write_state_fileloc): ...and use methods of
21422 s_expr_writer to write indentation into the gtype.state output file
21423 to visually represent the hierarchical structure of the list
21425 (write_state_fields): ditto, renaming to...
21426 (state_writer::write_state_fields)
21427 (write_state_a_string): ditto, renaming to...
21428 (state_writer::write_state_a_string)
21429 (write_state_string_option): ditto, renaming to...
21430 (state_writer::write_state_string_option)
21431 (write_state_type_option): ditto, renaming to...
21432 (state_writer::write_state_type_option)
21433 (write_state_nested_option): ditto, renaming to...
21434 (state_writer::write_state_nested_option)
21435 (write_state_option): ditto, renaming to...
21436 (state_writer::write_state_option)
21437 (write_state_options): ditto, renaming to...
21438 (state_writer::write_state_options)
21439 (write_state_lang_bitmap): ditto, renaming to...
21440 (state_writer::write_state_lang_bitmap)
21441 (write_state_version): ditto, renaming to...
21442 (state_writer::write_state_version)
21443 (write_state_scalar_type): ditto, renaming to...
21444 (state_writer::write_state_scalar_type)
21445 (write_state_string_type): ditto, renaming to...
21446 (state_writer::write_state_string_type)
21447 (write_state_undefined_type): ditto, renaming to...
21448 (state_writer::write_state_undefined_type)
21449 (write_state_struct_union_type): ditto, renaming to...
21450 (state_writer::write_state_struct_union_type)
21451 (write_state_struct_type): ditto, renaming to...
21452 (state_writer::write_state_struct_type)
21453 (write_state_user_struct_type): ditto, renaming to...
21454 (state_writer::write_state_user_struct_type)
21455 (write_state_lang_struct_type): ditto, renaming to...
21456 (state_writer::write_state_lang_struct_type)
21457 (write_state_param_struct_type): ditto, renaming to...
21458 (state_writer::write_state_param_struct_type)
21459 (write_state_pointer_type): ditto, renaming to...
21460 (state_writer::write_state_pointer_type)
21461 (write_state_array_type): ditto, renaming to...
21462 (state_writer::write_state_array_type)
21463 (write_state_gc_used): ditto, renaming to...
21464 (state_writer::write_state_gc_used)
21465 (write_state_common_type_content): ditto, renaming to...
21466 (state_writer::write_state_common_type_content)
21467 (write_state_type): ditto, renaming to...
21468 (state_writer::write_state_type)
21469 (write_state_pair_list): ditto, renaming to...
21470 (state_writer::write_state_pair_list)
21471 (write_state_pair): ditto, renaming to...
21472 (state_writer::write_state_pair)
21473 (write_state_typedefs): ditto, renaming to...
21474 (state_writer::write_state_typedefs)
21475 (write_state_structures): ditto, renaming to...
21476 (state_writer::write_state_structures)
21477 (write_state_param_structs): ditto, renaming to...
21478 (state_writer::write_state_param_structs)
21479 (write_state_variables): ditto, renaming to...
21480 (state_writer::write_state_variables)
21481 (write_state_srcdir): ditto, renaming to...
21482 (state_writer::write_state_srcdir)
21483 (write_state_files_list): ditto, renaming to...
21484 (state_writer::write_state_files_list)
21485 (write_state_languages): ditto, renaming to...
21486 (state_writer::write_state_languages)
21487 (write_state): create a state_writer instance and use it when
21488 writing out the state file
21490 2013-05-17 Mike Stump <mikestump@comcast.net>
21492 PR rtl-optimization/57304
21493 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
21494 accessing DF_REF_REAL_LOC.
21496 2013-05-17 Jakub Jelinek <jakub@redhat.com>
21498 PR rtl-optimization/57281
21499 PR rtl-optimization/57300
21500 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
21501 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
21502 what the other splitter did if the registers are dead.
21504 2013-05-17 Richard Biener <rguenther@suse.de>
21506 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
21509 2013-05-17 Jakub Jelinek <jakub@redhat.com>
21511 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
21514 2013-05-17 Marek Polacek <polacek@redhat.com>
21516 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
21517 length when doing non-zero store of storing '\0' to '\0'.
21519 2013-05-17 Jakub Jelinek <jakub@redhat.com>
21521 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
21522 vect_external_def oprnd1 with loop_vinfo, try to emit
21523 optional cast, negation and and stmts on the loop preheader
21524 edge instead of into the pattern def seq.
21526 PR tree-optimization/57051
21527 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
21528 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
21530 2013-05-16 Nick Clifton <nickc@redhat.com>
21532 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
21533 (rl78_is_naked_func): New function.
21534 (rl78_expand_prologue): Skip prologue generation for naked functions.
21535 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
21536 * doc/extend.texi (naked): Add RL78 to the list of processors
21537 that supports this attribute.
21539 2013-05-16 Jeff Law <law@redhat.com>
21541 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
21543 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
21545 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
21546 cache parameters using detect_caches_amd also for CYRIX,
21547 NSC and TM2 signatures.
21549 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
21550 Dzianis Kahanovich <mahatma@eu.by>
21554 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
21555 VIA/Centaur processors and determine their cache parameters
21556 using detect_caches_amd.
21558 2013-05-16 Teresa Johnson <tejohnson@google.com>
21560 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
21561 (rtl_verify_edges): New function.
21562 (rtl_verify_bb_insns): Ditto.
21563 (rtl_verify_bb_pointers): Ditto.
21564 (rtl_verify_bb_insn_chain): Ditto.
21565 (rtl_verify_fallthru): Ditto.
21566 (rtl_verify_bb_layout): Ditto.
21567 (rtl_verify_flow_info_1): Outline checks into new functions.
21568 (rtl_verify_flow_info): Ditto.
21570 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
21572 * cfghooks.c (copy_bbs): Add update_dominance argument.
21573 * cfghooks.h (copy_bbs): Update prototype.
21574 * tree-cfg.c (gimple_duplicate_sese_region):
21575 Add update_dominance argument.
21576 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
21577 * tree-ssa-loop-ch.c (copy_loop_headers): Update
21578 gimple_duplicate_sese_region call.
21579 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
21580 Update copy_bbs call.
21581 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
21582 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
21584 2013-05-16 Jakub Jelinek <jakub@redhat.com>
21586 * tree-vectorizer.h (NUM_PATTERNS): Increment.
21587 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
21588 vect_recog_rotate_pattern.
21589 (vect_recog_rotate_pattern): New function.
21591 2013-05-16 Jason Merrill <jason@redhat.com>
21593 * Makefile.in (LLINKER): New variable.
21594 (mostlyclean): Remove link mutex.
21595 * configure.ac: Handle --enable-link-mutex.
21596 * lock-and-run.sh: New script.
21598 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21601 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
21604 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21606 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
21608 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
21610 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
21611 * config/arm/arm.c (next_consecutive_mem): New function.
21612 (gen_movmem_ldrd_strd): Likewise.
21613 * config/arm/arm.md (movmemqi): Update condition and code.
21614 (unaligned_loaddi, unaligned_storedi): New patterns.
21616 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21618 * config.gcc: Obsolete *-*-solaris2.9*.
21619 * doc/install.texi (Specific, *-*-solaris2*): Document it.
21621 2013-05-16 Richard Biener <rguenther@suse.de>
21623 * passes.c (init_optimization_passes): Move pass_parallelize_loops
21624 earlier, after GRAPHITE transforms and IV canonicalization.
21626 2013-05-16 Jakub Jelinek <jakub@redhat.com>
21628 * omp-low.c (extract_omp_for_data): For collapsed loops,
21629 if at least one of the loops is known at compile time to
21630 iterate zero times, set count to 0.
21631 (expand_omp_regimplify_p): New function.
21632 (expand_omp_for_generic): For collapsed loops, if at least
21633 one of the loops isn't known to iterate at least once,
21634 add runtime check with setting count to 0.
21635 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
21636 For unsigned types if it isn't known at compile time that
21637 the loop will iterate at least once, add runtime check to bypass
21638 the whole loop if initial condition isn't true.
21640 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
21642 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
21644 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
21646 PR middle-end/57286
21647 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
21648 transformations to avoid an infinite loop.
21650 2013-05-16 Marek Polacek <polacek@redhat.com>
21652 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
21654 2013-05-15 Leif Ekblad <leif@rdos.net>
21656 * config/i386/i386.c (ix86_decompose_address): Use
21657 DEFAULT_TLS_SEG_REG to access TLS segment register.
21658 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
21659 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
21660 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
21662 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
21665 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
21666 sibling calls to functions that would normally be lazily bound,
21667 unless $gp is call-clobbered.
21669 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
21671 * config/i386/i386.c (ix86_option_override_internal): Update
21672 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
21673 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
21674 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
21675 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
21677 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
21679 2013-05-15 Andreas Schwab <schwab@suse.de>
21681 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
21682 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
21685 2013-05-15 Teresa Johnson <tejohnson@google.com>
21687 * loop-unroll.c (report_unroll_peel): Check decision before
21688 emitting unroll/peel message.
21690 2013-05-15 Teresa Johnson <tejohnson@google.com>
21692 * function.h (has_bb_partition): New rtl_data flag.
21693 (bb_reorder_complete): Ditto.
21694 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
21695 instead of flag_reorder_blocks_and_partition.
21696 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
21697 with some enhancements.
21698 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
21699 * bb-reorder.c (connect_traces): Check for has_bb_partition
21700 instead of flag_reorder_blocks_and_partition.
21701 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
21702 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
21703 verify_hot_cold_block_grouping.
21704 (partition_hot_cold_basic_blocks): Set has_bb_partition.
21706 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21709 * config/arm/predicates.md (call_insn_operand): New predicate.
21710 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
21711 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
21712 if insn is not a tail call.
21713 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
21715 * config/arm/arm.h (enum reg_class): New caller save register class.
21716 (REG_CLASS_NAMES): Likewise.
21717 (REG_CLASS_CONTENTS): Likewise.
21718 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
21721 2013-05-15 Richard Biener <rguenther@suse.de>
21723 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
21724 of MSG_OPTIMIZED_LOCATIONS.
21725 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
21726 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
21728 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
21729 of MSG_OPTIMIZED_LOCATIONS.
21730 (execute_vect_slp): Likewise.
21731 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
21732 (vect_create_cond_for_alias_checks): Likewise.
21733 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
21734 (vect_recog_widen_mult_pattern): Likewise.
21735 (vect_recog_widen_sum_pattern): Likewise.
21736 (vect_recog_over_widening_pattern): Likewise.
21737 (vect_recog_widen_shift_pattern): Likewise.
21738 (vect_recog_vector_vector_shift_pattern): Likewise.
21739 (vect_recog_divmod_pattern): Likewise.
21740 (vect_recog_mixed_size_cond_pattern): Likewise.
21741 (vect_recog_bool_pattern): Likewise.
21742 (vect_pattern_recog_1): Likewise.
21744 2013-05-15 Martin Jambor <mjambor@suse.cz>
21746 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
21747 non-functions to builtin_unreachable.
21748 * ipa-inline-transform.c (inline_call): Do not assert estimates were
21749 correct when new direct edges were discovered.
21751 2013-05-15 Martin Jambor <mjambor@suse.cz>
21753 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
21754 header, print symbol order instead of node uid, print more information
21755 about indirect edge targets.
21756 (ipa_make_edge_direct_to_target): Print symbol order instead of node
21758 (ipa_make_edge_direct_to_target): Likewise.
21759 (remove_described_reference): Likewise.
21760 (propagate_controlled_uses): Likewise.
21761 (ipa_print_node_params): Also print symbol order.
21762 (ipcp_transform_function): Print symbol order instead of node uids.
21763 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
21764 (cgraph_get_create_real_symbol_node): Likewise.
21765 * ipa-cp.c (print_lattice): Likewise.
21766 (print_all_lattices): Likewise.
21767 (determine_versionability): Likewise.
21768 (initialize_node_lattices): Likewise.
21769 (estimate_local_effects): Likewise.
21770 (update_profiling_info): Likewise.
21771 (create_specialized_node): Likewise.
21772 (perhaps_add_new_callers): Likewise.
21773 (decide_about_value): Likewise.
21774 (decide_whether_version_node): Likewise.
21775 (identify_dead_nodes): Likewise.
21776 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
21777 (dump_inline_summary): Likewise.
21778 (estimate_node_size_and_time): Likewise.
21779 (inline_analyze_function): Likewise.
21780 * ipa-inline.c (report_inline_failed_reason): Likewise.
21781 (want_early_inline_function_p): Likewise.
21782 (edge_badness): Likewise.
21783 (update_edge_key): Likewise.
21784 (inline_small_functions): Likewise. Add dumping of order to two other
21786 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
21787 instead of node uids.
21788 (propagate_pure_const): Likewise.
21789 (propagate_pure_const): Likewise.
21790 * ipa-utils.c (dump_cgraph_node_set): Likewise.
21791 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
21792 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
21794 * tree-pretty-print.c (dump_function_header): Likewise.
21795 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
21796 Print symbol order instead of node uids.
21798 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21800 * config/s390/s390.c (s390_register_move_cost): Don't impose the
21801 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
21803 2013-05-15 Richard Biener <rguenther@suse.de>
21805 PR tree-optimization/57275
21806 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
21807 return value for fail to do runtime alias checks for gather loads.
21809 2013-05-15 Jan Hubicka <jh@suse.cz>
21813 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
21814 weakrefs are not external.
21815 (lto_symtab_merge_decls): Fix thinko when dealing with
21816 non-lto_symtab decls.
21817 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
21818 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
21819 * varpool.c (dump_varpool_node): Dump more flags.
21821 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
21823 * config/i386/i386.c (processor_alias_table): Add instruction
21824 FSGSBASE for AMD bdver3 architecture.
21826 2013-05-14 Jakub Jelinek <jakub@redhat.com>
21828 * tree.c (warn_deprecated_use): Print file:line using locus color.
21829 * diagnostic.c (diagnostic_report_current_module): Print file:line
21830 and file:line:column using locus color.
21832 2013-05-14 Mike Stump <mikestump@comcast.net>
21834 * gdbinit.in: Add __null.
21836 2013-05-14 Mike Stump <mikestump@comcast.net>
21838 * recog.h: Rename struct recog_data to Recog_data.
21839 * recog.c: Likewise.
21840 * reload.c (can_reload_into): Likewise.
21841 * config/picochip/picochip.c: Likewise.
21843 2013-05-14 Mike Stump <mikestump@comcast.net>
21845 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
21847 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
21849 * resource.h (struct resources): Remove unch_memory member.
21850 (CLEAR_RESOURCE): Don't clear unch_memory.
21851 * resource.c (mark_referenced_resources): Don't set it.
21852 (mark_set_resources): Likewise.
21853 (mark_target_live_regs): Don't clear it.
21854 (init_resource_info): Likewise.
21855 * reorg.c (resource_conflicts_p): Don't compare it.
21856 (redundant_insn): Don't set it.
21858 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
21860 * emit-rtl.c (next_label): Remove unused function.
21861 (skip_consecutive_labels, link_cc0_insns): Move to ...
21862 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
21863 only place where these functions are used, and make them static.
21865 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
21867 * fold-const.c (fold_negate_expr): Handle vectors.
21868 (fold_truth_not_expr): Make it static.
21869 (fold_invert_truthvalue): New static function.
21870 (invert_truthvalue_loc): Handle vectors. Do not call
21871 fold_truth_not_expr directly.
21872 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
21873 <TRUTH_NOT_EXPR>: Do not cast to boolean.
21874 (fold_comparison): Handle vector constants.
21875 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
21876 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
21877 * tree.h (fold_truth_not_expr): Remove declaration.
21879 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
21881 * config/aarch64/aarch64-simd.md
21882 (aarch64_vcond_internal<mode>): Rename to...
21883 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
21884 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
21885 float modes. Clarify all iterator modes.
21886 (vcond<mode><mode>): Use new name for vcond expanders.
21887 (vcond<v_cmp_result><mode>): Likewise.
21888 (vcondu<mode><mode>: Likewise.
21889 * config/aarch64/iterators.md (VDQF_COND): New.
21891 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
21894 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
21895 variable for the shift amount. Check that we shift by non-negative
21898 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
21901 * config/arm/arm.h (EPILOGUE_USES): Only return true
21902 for LR_REGNUM after epilogue_completed.
21904 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
21906 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
21907 is error_mark_node.
21909 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21912 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
21913 and Solaris 11+/x86 with gld.
21914 * configure: Regenerate.
21916 2013-05-14 Jakub Jelinek <jakub@redhat.com>
21918 * expmed.c (expand_shift_1): Canonicalize rotates by
21919 constant bitsize / 2 to bitsize - 1.
21920 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
21921 case ROTATERT>: Likewise.
21924 2013-05-10 Jakub Jelinek <jakub@redhat.com>
21926 * config/i386/i386.md (rotateinv): New code attr.
21927 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
21928 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
21929 roll $31, %eax, etc.
21931 2013-05-14 Richard Biener <rguenther@suse.de>
21933 PR middle-end/57235
21934 * tree-eh.c (sink_clobbers): Give up for successors with
21935 multiple predecessors and no virtual uses.
21937 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
21939 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
21940 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
21942 2013-05-14 Jakub Jelinek <jakub@redhat.com>
21944 PR middle-end/57251
21945 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
21946 the case when both op0 and op1 have VOIDmode.
21948 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
21950 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
21951 in multiply-accumulate mode.
21953 2013-05-13 Guozhi Wei <carrot@google.com>
21955 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
21957 2013-05-13 Kai Tietz <ktietz@redhat.com>
21960 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
21961 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
21962 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
21963 * config/i386/i386.c (ix86_option_override_internal): Likewise.
21964 (ix86_expand_prologue): Likewise.
21965 (ix86_expand_split_stack_prologue): Likewise.
21966 (legitimate_pic_address_disp_p): Likewise.
21967 (legitimize_pic_address): Likewise.
21968 (legitimize_tls_address): Likewise.
21969 (legitimize_pe_coff_symbol): Likewise.
21970 (output_pic_addr_const): Likewise.
21971 (construct_plt_address): Likewise.
21972 (ix86_expand_call): Likewise.
21973 (x86_output_mi_thunk): Likewise.
21974 (x86_function_profiler): Likewise.
21976 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
21978 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
21979 similar switch cases.
21980 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
21981 (aarch64_simd_mov_to_<mode>low): Delete.
21982 (aarch64_simd_mov_to_<mode>high): Delete.
21983 (move_lo_quad_<mode>): Add w<-r alternative.
21984 (aarch64_simd_move_hi_quad_<mode>): Likewise.
21985 (aarch64_simd_mov_from_*): Update type attribute.
21986 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
21989 2013-05-13 Jan Hubicka <jh@suse.cz>
21991 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
21992 * config/i386/i386.c (ix86_compute_frame_layout,
21993 ix86_expand_epilogue, emit_i387_cw_initialization,
21994 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
21995 ix86_local_alignment): Fix use of size/speed predicates.
21997 2013-05-13 Jakub Jelinek <jakub@redhat.com>
21999 PR tree-optimization/45216
22000 PR tree-optimization/57157
22001 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
22002 the (-Y) & (B - 1) variant if OP is |.
22003 * expmed.c (expand_shift_1): For rotations by const0_rtx just
22004 return shifted. Use (-op1) & (prec - 1) as other_amount
22005 instead of prec - op1.
22007 2013-05-13 Martin Jambor <mjambor@suse.cz>
22009 PR middle-end/42371
22010 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
22011 (ipa_constant_data): New type.
22012 (ipa_jump_func): Use ipa_constant_data to hold information about
22013 constant jump functions.
22014 (ipa_get_jf_constant): Adjust to jump function type changes.
22015 (ipa_get_jf_constant_rdesc): New function.
22016 (ipa_param_descriptor): New field controlled_uses.
22017 (ipa_get_controlled_uses): New function.
22018 (ipa_set_controlled_uses): Likewise.
22019 * ipa-ref.h (ipa_find_reference): Declare.
22020 * ipa-prop.c (ipa_cst_ref_desc): New type.
22021 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
22023 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
22024 New parameter cs. Adjust all callers.
22025 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
22026 (remove_described_reference): New function.
22027 (jfunc_rdesc_usable): Likewise.
22028 (try_make_edge_direct_simple_call): Decrement controlled use count,
22029 attempt to remove reference if it hits zero.
22030 (combine_controlled_uses_counters): New function.
22031 (propagate_controlled_uses): Likewise.
22032 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
22033 (ipa_edge_duplication_hook): Duplicate reference descriptions.
22034 (ipa_print_node_params): Print described use counter.
22035 (ipa_write_jump_function): Adjust to jump function type changes.
22036 (ipa_read_jump_function): New parameter CS, pass it to
22037 ipa_set_jf_constant. Adjust caller.
22038 (ipa_write_node_info): Stream controlled use count
22039 (ipa_read_node_info): Likewise.
22040 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
22042 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
22043 count. Remove cloning-added reference if it reaches zero.
22044 * ipa-ref.c (ipa_find_reference): New function.
22046 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
22048 * config/i386/i386.c (processor_target_table): Modified default
22049 alignment values for AMD BD and BT architectures.
22051 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
22053 * tree-vect-generic.c (uniform_vector_p): Move ...
22054 * tree.c (uniform_vector_p): ... here.
22055 * tree.h (uniform_vector_p): Declare it.
22056 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
22059 2013-05-13 Jakub Jelinek <jakub@redhat.com>
22061 PR tree-optimization/57230
22062 * tree-ssa-strlen.c (handle_char_store): Record length for
22063 array store from STRING_CST.
22065 PR tree-optimization/57230
22066 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
22069 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
22071 * config/epiphany/epiphany.c (epiphany_init): Check size of
22072 NUM_MODES_FOR_MODE_SWITCHING.
22073 (epiphany_expand_prologue):
22074 Remove CONFIG_REGNUM initial value handling code.
22075 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
22076 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
22077 (emit_set_fp_mode, epiphany_mode_after): Likewise.
22078 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
22079 Don't return 1 for FP_MODE_NONE.
22080 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
22081 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
22082 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
22083 * config/epiphany/epiphany.md (save_config): New pattern.
22085 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
22087 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
22089 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
22091 * config/i386/i386.md (memory): Handle sseishft1.
22092 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
22093 (*vec_extractv2di_1): Ditto.
22095 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
22097 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
22100 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
22102 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
22103 Add mthumb/march=armv7-a multilib.
22104 Add mthumb/march=armv7-r multilib.
22105 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
22107 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
22109 * config/v850/t-rtems: Add more multilibs.
22111 2013-05-10 Richard Biener <rguenther@suse.de>
22113 PR tree-optimization/57214
22114 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
22115 not propagate from SSA names that occur in abnormal PHI nodes.
22117 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
22119 * stor-layout.c (element_precision): New function.
22120 * machmode.h (element_precision): Declare it.
22121 * tree.c (build_minus_one_cst): New function.
22122 (element_precision): Likewise.
22123 * tree.h (build_minus_one_cst): Declare new function.
22124 (element_precision): Likewise.
22125 * fold-const.c (operand_equal_p): Use element_precision.
22126 (fold_binary_loc): Handle vector types.
22127 * convert.c (convert_to_integer): Use element_precision.
22128 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
22131 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
22133 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
22134 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
22135 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
22136 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
22137 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
22138 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
22139 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
22140 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
22141 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
22142 (Uuw8): New constraints.
22143 (Usb4): Move into alphabetical order.
22144 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
22145 (sd8_operand, ub8_operand, uw8_operand): New predicates.
22146 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
22147 previously unnamed patterns.
22148 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
22149 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
22150 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
22151 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
22152 of set_attr_alternative/if_then_else. Use extended_mips16 instead
22153 of specific lengths.
22155 2013-05-10 Jakub Jelinek <jakub@redhat.com>
22157 * config/i386/i386.md (rotateinv): New code attr.
22158 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
22159 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
22160 roll $31, %eax, etc.
22162 PR tree-optimization/45216
22163 PR tree-optimization/57157
22164 * tree-ssa-forwprop.c (simplify_rotate): New function.
22165 (ssa_forward_propagate_and_combine): Call it.
22167 2013-05-10 Richard Biener <rguenther@suse.de>
22169 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
22170 disable peeling when we version for aliasing.
22171 (vector_alignment_reachable_p): Honor explicit user alignment.
22172 (vect_supportable_dr_alignment): Likewise.
22173 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
22174 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
22175 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
22176 then peeling to arrange for the cost-model check to come first.
22178 2013-05-10 Alan Modra <amodra@gmail.com>
22180 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
22181 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
22182 * configure: Regenerate.
22184 2013-05-10 Alan Modra <amodra@gmail.com>
22187 * varasm.c (default_elf_select_section): Move !DECL_P check..
22188 (get_named_section): ..to here before calling get_section_name.
22190 (default_section_type_flags): Add DECL_P check.
22191 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
22192 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
22194 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
22196 * config/epiphany/epiphany.c (epiphany_expand_prologue):
22197 When using gen_stack_adjust_str with a register offset, add a
22198 REG_FRAME_RELATED_EXPR note.
22200 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
22202 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
22203 (*vec_extractv4si_zext_mem): Ditto.
22204 (*vec_extractv2di): Add 0->x and x->x alternatives.
22205 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
22206 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
22208 2013-05-09 Jason Merrill <jason@redhat.com>
22210 N3639 C++1y VLA support
22211 * gimplify.c (gimplify_vla_decl): Don't touch an existing
22214 * tree.c (build_constructor_va): New.
22215 * tree.h: Declare it.
22217 2013-05-09 Martin Jambor <mjambor@suse.cz>
22220 * gimple-fold.c (canonicalize_constructor_val): Call
22221 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
22223 2013-05-09 Jan Hubicka <jh@suse.cz>
22224 Richard Biener <rguenther@suse.de>
22227 * symtab.c (symtab_make_decl_local): Do not add private names.
22229 2013-05-09 Jan Hubicka <jh@suse.cz>
22232 * symtab.c (insert_to_assembler_name_hash): Handle clones.
22233 (unlink_from_assembler_name_hash): Likewise.
22234 (symtab_prevail_in_asm_name_hash, symtab_register_node,
22235 symtab_unregister_node, symtab_initialize_asm_name_hash,
22236 change_decl_assembler_name): Update.
22238 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
22240 * config/aarch64/aarch64.md: New movtf split.
22241 (*movtf_aarch64): Update.
22242 (aarch64_movdi_tilow): Handle TF modes and rename to
22243 aarch64_movdi_<mode>low.
22244 (aarch64_movdi_tihigh): Handle TF modes and rename to
22245 aarch64_movdi_<mode>high
22246 (aarch64_movtihigh_di): Handle TF modes and rename to
22247 aarch64_mov<mode>high_di
22248 (aarch64_movtilow_di): Handle TF modes and rename to
22249 aarch64_mov<mode>low_di
22250 (aarch64_movtilow_tilow): Remove spurious whitespace.
22251 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
22253 (aarch64_print_operand): Update.
22255 2013-05-09 Alan Modra <amodra@gmail.com>
22257 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
22259 * configure: Regenerate.
22261 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
22263 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
22264 splitter preparation statements.
22265 * config/i386/sse.md (*vec_extract* splitters): Ditto.
22266 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
22269 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22271 * gimple-ssa-strength-reduction.c (count_candidates): Change
22272 return value to int.
22273 (analyze_candidates_and_replace): Change type of length to int.
22275 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
22277 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
22278 (*vec_extract<mode>): Use VI12_128 mode iterator.
22279 (*vec_extract<mode>_mem): Ditto.
22280 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
22283 2013-05-08 Diego Novillo <dnovillo@google.com>
22288 2012-08-17 Diego Novillo <dnovillo@google.com>
22291 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
22292 * config.in: Regenerate.
22293 * configure: Regenerate.
22294 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
22296 2013-05-08 Jan Hubicka <jh@suse.cz>
22299 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
22300 * cgraph.h (symtab_node_base): Add unique_name.
22301 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
22302 input_overwrite_node, input_varpool_node): Stream unique_name.
22303 * cgraphclones.c (cgraph_create_virtual_clone,
22304 cgraph_function_versioning): Set unique_name.
22305 * ipa.c (function_and_variable_visibility): Set unique_name.
22307 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22309 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
22310 (alloc_cand_and_find_basis): Restrict conditional candidate
22311 processing to CAND_MULTs.
22313 2013-05-08 Jan Hubicka <jh@suse.cz>
22316 lto-symtab.c (lto_symtab_symbol_p): New function.
22317 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
22318 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
22319 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
22320 Skip static symbols.
22322 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
22324 PR tree-optimization/57200
22325 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
22326 Only call inform if the preceding warning_at returns true.
22328 2013-05-07 Han Shen <shenhan@google.com>
22330 * cfgexpand.c (record_or_union_type_has_array_p): New function.
22331 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
22332 * common.opt (fstack-protector-strong): New option.
22333 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
22334 * doc/invoke.texi (Optimization Options): Document
22335 "-fstack-protector-strong".
22336 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
22338 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
22340 * config/mips/mips.c (mips_machine_reorg2): Return 0.
22342 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
22344 * ira.c (update_equiv_regs): Add insn having equiv memory even if
22345 it is not lhs of the insn.
22346 (setup_reg_equiv): Remove insn having equiv memory which it is not
22348 * lra-constraints.c (process_address): Try to improve generation
22349 code for address base + disp.
22350 (lra_constraints): Make correct the code for checking insn setting
22351 up backward equivalence. Remove insn only if it is in the init
22353 * lra-eliminations.c (update_reg_eliminate): Change return value.
22354 (lra_eliminate): Use the result.
22356 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
22358 * config/i386/sse.md (ssescalarnummask): New mode attribute.
22359 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
22360 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
22361 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
22362 register target operands.
22363 (*vec_extractv8hi_sse2): New pattern.
22364 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
22365 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
22366 (*vec_extract<mode>_mem): New insn and split pattern.
22368 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
22370 * config/arm/arm.c (arm_asan_shadow_offset): New function.
22371 (TARGET_ASAN_SHADOW_OFFSET): Define.
22372 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
22373 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
22375 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22377 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
22378 (incr_vec_index): Return -1 if increment not found.
22379 (create_add_on_incoming_edge): Assert if increment not found.
22380 (record_increment): Limit number of increments recorded.
22381 (all_phi_incrs_profitable): Return false if an increment not found.
22382 (replace_profitable_candidates): Don't process increments that were
22384 (analyze_candidates_and_replace): Limit size of incr_vec.
22386 2013-05-07 Richard Biener <rguenther@suse.de>
22388 * calls.c (special_function_p): setjmp-like functions are leaf.
22389 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
22390 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
22392 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
22394 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
22395 (aarch64_simd_mov<mode>): New expander.
22396 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
22397 (aarch64_simd_mov_to_<mode>high): Likewise.
22398 (aarch64_simd_mov_from_<mode>low): Likewise.
22399 (aarch64_simd_mov_from_<mode>high): Likewise.
22400 (aarch64_dup_lane<mode>): Update.
22401 (aarch64_dup_lanedi): New instruction pattern.
22402 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
22403 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
22405 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22407 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
22408 (replace_mult_candidate): Remove unnecessary argument; remove
22409 unnecessary parameter from call to introduce_cast_before_cand.
22410 (replace_unconditional_candidate): Remove unnecessary parameter
22411 from call to replace_mult_candidate.
22412 (replace_conditional_candidate): Likewise.
22413 (insert_initializers): Use make_temp_ssa_name.
22414 (introduce_cast_before_cand): Remove unnecessary argument; use
22415 make_temp_ssa_name.
22416 (replace_one_candidate): Remove unnecessary argument; remove
22417 unnecessary parameter from calls to introduce_cast_before_cand.
22418 (replace_profitable_candidates): Remove unnecessary parameters
22419 from calls to replace_one_candidate.
22421 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22423 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
22424 phi def as possibly hiding a basis for a CAND_ADD whose operands
22425 have been commuted in the analysis.
22426 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
22428 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22430 * config/aarch64/aarch64.md
22431 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
22432 shift value between 0-4.
22434 2013-05-07 Richard Biener <rguenther@suse.de>
22436 * double-int.h (rshift): New overload.
22437 * double-int.c (rshift): New function.
22438 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
22439 (create_reference_ops_from_ref): Remove.
22440 (vn_reference_insert): Use shared ops for constructing the
22441 reference and copy it.
22443 2013-05-07 Richard Biener <rguenther@suse.de>
22445 PR middle-end/57190
22446 * tree-eh.c (sink_clobbers): Properly propagate
22447 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
22449 2013-05-07 Jakub Jelinek <jakub@redhat.com>
22451 PR tree-optimization/57149
22452 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
22453 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
22454 collect_phi_def_edges, execute_late_warn_uninitialized): Use
22455 uninit_undefined_value_p instead of ssa_undefined_value_p.
22458 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
22459 for modifier == EXPAND_INITIALIZER.
22461 2013-05-07 Anton Blanchard <anton@samba.org>
22463 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
22464 for powerpc64 little endian.
22465 * configure: Regenerate.
22467 2013-05-06 Graham Stott <grahams@btinternet.com>
22469 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
22470 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
22471 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
22472 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
22474 2013-05-06 Graham Stott <grahams@btinternet.com>
22476 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
22477 codes which allow non-lvalues.
22479 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
22481 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
22482 components are all 1s.
22483 (integer_minus_onep): New function.
22484 * tree.h (integer_minus_onep): Declare it.
22485 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
22486 integer_minus_onep instead of integer_all_onesp.
22488 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
22491 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
22492 variations of these patterns.
22494 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
22496 * config/i386/i386.md (isa): Add x64_sse4 member.
22497 (enabled): Handle x64_sse4.
22498 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
22499 instruction for 64bit SSE4_1 targets. Update insn attributes.
22500 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
22501 instruction for SSE4_1 targets. Update insn attributes.
22502 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
22503 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
22504 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
22506 (*vec_extractv4si): Rename from *sse4_1_pextrd.
22507 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
22508 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
22510 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
22513 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
22515 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
22517 * final.c (do_assembler_dialects): Don't handle curly braces and
22518 vertical bar escaped by % as dialect delimiters.
22519 (output_asm_insn): Print curly braces and vertical bar if escaped
22520 by % and ASSEMBLER_DIALECT defined.
22521 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
22522 * doc/tm.texi: Regenerated.
22524 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
22526 * config/mips/mips.c: Include tree-pass.h.
22527 (mips_reorg): Split in pre- and post-dbr_schedule parts.
22528 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
22529 (pass_mips_machine_reorg2): New machine specific pass.
22530 (insert_pass_mips_machine_reorg2): New pass plugin definition.
22531 (mips_option_override): Register the new pass.
22532 * rtl.h (cleanup_barriers): Remove prototype.
22533 (dbr_schedule): Likewise.
22534 * jump.c (cleanup_barriers): Make static.
22535 * reorg.c (dbr_schedule): Likewise.
22537 2013-05-06 Richard Biener <rguenther@suse.de>
22539 PR tree-optimization/57185
22540 * tree-parloops.c (add_field_for_reduction): Handle anonymous
22541 SSA names properly.
22543 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
22546 * config/i386/i386.c (add_parameter_dependencies): Add dependence
22547 between "first_arg" and "insn", not "last" and "insn".
22549 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22551 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
22552 (find_candidates_in_block): Re-enable slsr_process_phi.
22553 (create_phi_basis): Fix double counting of candidate adjustment.
22555 2013-05-06 Richard Biener <rguenther@suse.de>
22557 PR middle-end/57147
22558 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
22559 the edge is also fallthru, preserve it and just clear the
22561 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
22562 also complex, preserve that and just clear the fallthru flag.
22563 * tree-inline.c (update_ssa_across_abnormal_edges): Also
22564 update virtual operands.
22566 2013-05-06 Alan Modra <amodra@gmail.com>
22568 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
22569 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
22570 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
22571 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
22572 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
22573 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
22575 2013-05-06 Alan Modra <amodra@gmail.com>
22577 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
22578 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
22579 (DEFAULT_ASM_ENDIAN): Define.
22580 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
22581 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
22582 Update -K PIC clause from sysv4.h.
22583 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
22584 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
22586 2013-05-06 Alan Modra <amodra@gmail.com>
22588 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
22589 twice for little-endian.
22590 (ashrdi3_no_power, ashrdi3): Support little-endian.
22592 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
22595 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
22596 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
22598 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
22599 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
22600 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
22603 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
22604 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22606 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
22607 * config/pa/pa.opt: Make mbig-switch a no-op.
22608 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
22609 (CASE_VECTOR_MODE): Always return SImode.
22610 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
22611 for the !TARGET_BIG_SWITCH case.
22612 * config/pa/pa-linux.h: Likewise.
22613 * config/pa/pa-openbsd.h: Likewise.
22614 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
22615 * config/pa/pa.md (short_jump): Remove define_insn.
22616 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
22617 (casesi0): Remove define_insn.
22618 (type): Remove btable_branch.
22619 (pa_combine_type): Likewise.
22620 (in_nullified_branch_delay): Likewise.
22621 (in_call_delay): Likewise.
22622 (define_delay): Likewise.
22623 (define_insn_reservation "Z3"): Likewise.
22624 (define_insn_reservation "Z4"): Likewise.
22625 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
22626 (pa_adjust_insn_length): Remove adjustment for btable branches.
22627 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
22630 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
22632 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
22633 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
22634 Add m->r,x alternatives.
22635 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
22636 splitters using SWI48x mode iterator.
22637 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
22638 TARGET_64BIT. Add m->x alternative.
22639 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
22640 Add o->x alternative. Enable for TARGET_SSE.
22641 (sse_storeq): Remove expander.
22642 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
22643 with memory input operand.
22644 (*vec_extractv2di_1 splitter): New.
22645 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
22646 * config/i386/i386.md (ssevecmodelower): New mode attribute.
22648 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
22650 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
22651 (INT_LOWPART): Delete.
22652 (extract_MB): Adjust.
22653 (extract_ME): Adjust.
22654 (print_operand): Adjust.
22656 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
22658 * config/rs6000/predicates.md (reg_or_add_cint_operand,
22659 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
22660 (reg_or_logical_cint_operand, easy_fp_constant,
22661 logical_const_operand): Delete "CONST_DOUBLE" case.
22662 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
22663 "HOST_BITS_PER_WIDE_INT == 64" test.
22664 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
22665 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
22666 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
22667 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
22669 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
22670 CONST_DOUBLE DImode/VOIDmode case.
22671 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
22672 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
22673 CONST_DOUBLE VOIDmode case.
22674 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
22675 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
22676 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
22677 Delete CONST_DOUBLE case.
22678 (splitters for mov FMOVE64 const_double): Delete
22679 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
22680 "HOST_BITS_PER_WIDE_INT >= 64" test.
22681 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
22683 (mov DI const_double): Delete.
22685 2013-05-04 Jakub Jelinek <jakub@redhat.com>
22687 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
22688 on op shows all bits zero in mode of a lowpart subreg, return zero.
22690 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
22693 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
22694 to save TFmode registers and DImode to save TImode registers for
22695 caller save operations.
22696 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
22697 mark being partially clobbered since they only use the first
22700 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
22701 and TDmode only use the upper 64-bits of each VSX register.
22703 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22705 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
22706 (find_candidates_in_block): Disable slsr_process_phi.
22708 2013-05-03 Guozhi Wei <carrot@google.com>
22710 * coverage.c (coverage_obj_init): Move the construction of gcov
22712 (build_init_ctor): ... here.
22714 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22716 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
22717 (slsr_cand_d): Redefine def_phi.
22718 (stride_status, phi_adjust_status, count_phis_status): New enums.
22719 (find_phi_def): New.
22720 (find_basis_for_base_expr): New.
22721 (find_basis_for_candidate): Handle hidden bases.
22722 (alloc_cand_and_find_basis): Handle phi candidates.
22723 (slsr_process_phi): New.
22724 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
22725 (create_mul_imm_cand): Likewise.
22726 (create_add_ssa_cand): Exclude phi base candidates.
22727 (create_add_imm_cand): Likewise.
22728 (slsr_process_cast): Likewise.
22729 (slsr_process_copy): Likewise.
22730 (find_candidates_in_block): Handle phi candidates.
22731 (dump_candidate): Likewise.
22732 (unconditional_cands): Delete.
22733 (unconditional_cands_with_known_stride_p): Delete.
22734 (phi_dependent_cand_p): New.
22735 (cand_increment): Handle phi-dependent candidates.
22736 (replace_dependent): Delete.
22737 (replace_mult_candidate): New.
22738 (replace_unconditional_candidate): New.
22739 (incr_vec_index): Move to avoid forward reference.
22740 (create_add_on_incoming_edge): New.
22741 (create_phi_basis): New.
22742 (replace_dependents): Delete.
22743 (replace_conditional_candidate): New.
22744 (phi_add_costs): New.
22745 (replace_uncond_cands_and_profitable_phis): New.
22746 (record_increment): Handle phi adjustments.
22747 (record_phi_increments): New.
22748 (record_increments): Handle phi adjustments.
22749 (phi_incr_cost): New.
22750 (lowest_cost_path): Handle phis.
22751 (total_savings): Likewise.
22752 (analyze_increments): Likewise.
22753 (ncd_with_phi): New.
22754 (ncd_of_cand_and_phis): New.
22755 (nearest_common_dominator_for_cands): Handle phi increments.
22756 (all_phi_incrs_profitable): New.
22757 (replace_profitable_candidates): Handle phi-dependent candidates.
22758 (analyze_candidates_and_replace): Likewise.
22760 2013-05-03 Teresa Johnson <tejohnson@google.com>
22763 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
22764 do not exceed REG_BR_PROB_BASE.
22766 2013-05-03 Jeff Law <law@redhat.com>
22768 PR tree-optimization/57144
22769 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
22770 operand of the condition will bit into the new type when eliminating
22771 a cast feeding a condition.
22773 2013-05-03 Jakub Jelinek <jakub@redhat.com>
22775 PR rtl-optimization/57130
22776 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
22777 of COMPARE as in_code to the recursive call if needed.
22779 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
22781 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
22782 (enabled): Handle new members.
22783 * config/i386/sse.md (*vec_concatv2si): Merge from
22784 *vec_concatv2si_sse2 and vec_concatv2si_sse.
22785 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
22787 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
22789 PR tree-optimization/57027
22790 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
22791 for fnms opportunity, check we got the prerequisite kind
22792 of tree / gimple before using accessor functions.
22794 2013-05-03 Richard Biener <rguenther@suse.de>
22796 * double-int.h (lshift): New overload without precision
22797 and arith argument.
22798 (operator *=, operator +=, operator -=): Move ...
22799 * double-int.c (operator *=, operator +=, operator -=): ... here
22800 and implement more efficiently.
22801 (mul_double_with_sign): Remove.
22802 (lshift_double): Adjust to take unsinged shift argument, push
22803 dispatching code to callers.
22804 (mul_double_wide_with_sign): Add early out for callers that
22805 are not interested in high parts or overflow.
22806 (lshift): New function.
22807 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
22808 dispatch code here.
22809 (lrotate, rrotate): Use logical shifts.
22810 * expr.c (get_inner_reference): Use lshift.
22811 * fixed-value.c (do_fixed_divide): Likewise.
22812 * tree-dfa.c (get_ref_base_and_extent): Likewise.
22813 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
22814 (indirect_refs_may_alias_p): Likewise.
22815 (stmt_kills_ref_p_1): Likewise.
22817 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
22819 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
22821 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
22823 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
22824 scalar form of FABD instruction.
22826 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
22828 * lra-constraints.c (process_alt_operands): Add checking alt
22829 number to choose the best alternative.
22831 2013-05-02 Richard Biener <rguenther@suse.de>
22833 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
22834 bitmap and its handling.
22835 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
22837 2013-05-02 Richard Biener <rguenther@suse.de>
22839 PR middle-end/57140
22840 * tree-inline.c (copy_loops): Properly handle removed loops.
22841 (copy_cfg_body): Mark destination loops for fixup if source
22842 loops needed fixup.
22844 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
22847 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
22848 generating simple_return for naked functions.
22850 2013-05-02 Martin Jambor <mjambor@suse.cz>
22852 PR middle-end/56988
22853 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
22854 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
22856 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
22857 ipa_agg_replacement_value structures.
22858 (known_aggs_to_agg_replacement_list): Likewise.
22859 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
22860 (read_agg_replacement_chain): Likewise.
22861 (ipcp_transform_function): Also check that by_ref flags match.
22863 2013-05-02 Richard Biener <rguenther@suse.de>
22865 * graphds.h (struct graph): Add obstack member.
22866 * graphds.c (new_graph): Initialize obstack and allocate
22868 (add_edge): Allocate edge from the obstack.
22869 (free_graph): Free the obstack instead of all edges and vertices.
22871 2013-05-02 Teresa Johnson <tejohnson@google.com>
22873 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
22875 * cfg.c (update_bb_profile_for_threading): Ditto.
22876 * tree-inline.c (copy_bb): Ditto.
22877 (copy_edges_for_bb): Ditto.
22878 (initialize_cfun): Ditto.
22879 (copy_cfg_body): Ditto.
22880 (expand_call_inline): Ditto.
22881 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
22882 (estimate_node_size_and_time): Ditto.
22883 (inline_merge_summary): Ditto.
22884 * cgraphclones.c (cgraph_clone_edge): Ditto.
22885 (cgraph_clone_node): Ditto.
22886 * sched-rgn.c (compute_dom_prob_ps): Ditto.
22887 (compute_trg_info): Ditto.
22889 2013-05-02 Ian Bolton <ian.bolton@arm.com>
22891 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
22892 S reg when fp attribute set.
22893 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
22895 2013-05-02 Ian Bolton <ian.bolton@arm.com>
22897 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
22899 (*and_one_cmplsi3_compare0_uxtw): Likewise.
22900 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
22901 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
22903 2013-05-02 Richard Biener <rguenther@suse.de>
22905 * tree-scalar-evolution.c (scev_info_hasher): Remove.
22906 (struct instantiate_cache_entry): New type.
22907 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
22908 (struct instantiate_cache_type): New type.
22909 (set_instantiated_value, get_instantiated_value): Remove.
22910 (get_instantiated_value_entry): New function.
22911 (instantiate_scev_name): Use the new cache and adjust.
22912 (instantiate_scev_poly): Adjust.
22913 (instantiate_scev_binary): Likewise.
22914 (instantiate_array_ref): Likewise.
22915 (instantiate_scev_convert): Likewise.
22916 (instantiate_scev_not): Likewise.
22917 (instantiate_scev_3): Likewise.
22918 (instantiate_scev_2): Likewise.
22919 (instantiate_scev_r): Likewise.
22920 (instantiate_scev): Likewise.
22921 (resolve_mixers): Likewise.
22923 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
22926 * lra-constraints.c (best_small_class_operands_num): Remove.
22927 (process_alt_operands): Remove small_class_operands_num. Take
22928 small classes operands into losers and only if the operand is not
22929 matched. Modify debugging output.
22930 (curr_insn_transform): Remove best_small_class_operands_num.
22933 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22935 * config/aarch64/aarch64-builtins.c
22936 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
22937 * config/aarch64/aarch64-simd-builtins.def
22938 (reduc_splus_): Add new modes.
22939 (reduc_uplus_): New.
22940 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
22941 (reduc_uplus_v4sf): Likewise.
22942 (reduc_splus_v4sf): Likewise.
22943 (aarch64_addv<mode>): Likewise.
22944 (reduc_uplus_<mode>): Likewise.
22945 (reduc_splus_<mode>): Likewise.
22946 (aarch64_addvv2di): Likewise.
22947 (reduc_uplus_v2di): Likewise.
22948 (reduc_splus_v2di): Likewise.
22949 (aarch64_addvv2si): Likewise.
22950 (reduc_uplus_v2si): Likewise.
22951 (reduc_splus_v2si): Likewise.
22952 (reduc_<sur>plus_<mode>): New.
22953 (reduc_<sur>plus_v2di): Likewise.
22954 (reduc_<sur>plus_v2si): Likewise.
22955 (reduc_<sur>plus_v4sf): Likewise.
22956 (aarch64_addpv4sf): Likewise.
22957 * config/aarch64/arm_neon.h
22958 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
22959 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
22960 add UNSPEC_SADDV, UNSPEC_UADDV.
22962 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
22964 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22966 * config/aarch64/arm_neon.h
22967 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
22969 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22971 * config/aarch64/aarch64-builtins
22972 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
22974 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22976 * config/aarch64/aarch64-simd-builtins.def
22977 (reduc_smax_): New.
22978 (reduc_smin_): Likewise.
22979 (reduc_umax_): Likewise.
22980 (reduc_umin_): Likewise.
22981 (reduc_smax_nan_): Likewise.
22982 (reduc_smin_nan_): Likewise.
22985 (smax): Update for V2SF, V4SF and V2DF modes.
22988 (smin_nan): Likewise.
22989 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
22990 (<su><maxmin><mode>3): ...This, refactor.
22991 (s<maxmin><mode>3): New.
22992 (<maxmin_uns><mode>3): Likewise.
22993 (reduc_<maxmin_uns>_<mode>): Refactor.
22994 (reduc_<maxmin_uns>_v4sf): Likewise.
22995 (reduc_<maxmin_uns>_v2si): Likewise.
22996 (aarch64_<fmaxmin><mode>: Remove.
22997 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
22999 (vmin<q>_f<32,64>): Likewise.
23000 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
23002 (su): Add mappings for smax, smin, umax, umin.
23004 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
23005 (FMAXMIN): Rename as...
23006 (FMAXMIN_UNS): ...This.
23008 (fmaxminv): Likewise.
23009 (fmaxmin): Likewise.
23011 (maxmin_uns_op): Likewise.
23013 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23015 * config/aarch64/arm_neon.h
23016 (vac<ge, gt><sd>_f<32, 64>): Rename to...
23017 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
23018 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
23020 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23022 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
23023 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
23025 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23027 * config/aarch64/aarch64-simd.md
23028 (vcond<mode>_internal): Handle special cases for constant masks.
23029 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
23030 (vcondu<mode><mode>): Likewise.
23031 (vcond<v_cmp_result><mode>): New.
23033 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23035 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
23036 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
23037 * config/aarch64/aarch64-simd-builtins.def
23038 (cmeq): Update to BUILTIN_VALLDI.
23043 * config/aarch64/arm_neon.h
23044 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
23045 to builtins or C as appropriate.
23047 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
23049 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
23051 (cmhi): Rename to...
23053 * config/aarch64/aarch64-simd.md
23054 (simd_mode): Add SF.
23055 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
23056 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
23057 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
23058 (cstore<mode>_neg): ...This.
23059 * config/aarch64/iterators.md
23061 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
23062 (COMPARISONS): New.
23063 (UCOMPARISONS): Likewise.
23064 (optab): Add missing comparisons.
23070 (VCMP_S): Likewise.
23071 (VCMP_U): Likewise.
23072 (V_cmp_result): Add DF, SF modes.
23073 (v_cmp_result): Likewise.
23075 (vmtype): Likewise.
23076 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
23078 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
23080 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
23081 define_insn to define_insn_and_split.
23082 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
23083 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
23084 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
23085 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
23086 (thumb2_negscc): Likewise.
23088 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
23090 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
23092 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
23094 * config/arm/thumb2.md: Remove trailing whitespaces.
23096 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
23098 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
23099 Use gen_int_mode rather than GEN_INT.
23101 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
23103 * value-prof.c (stream_in_histogram_value): Remove the strayed
23106 2013-04-30 Richard Biener <rguenther@suse.de>
23108 PR middle-end/57122
23109 * cfghooks.c (split_edge): Properly check for the loop latch edge.
23111 2013-04-30 Richard Biener <rguenther@suse.de>
23113 PR middle-end/57107
23114 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
23116 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
23118 PR rtl-optimization/56957
23119 PR rtl-optimization/57105
23120 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
23121 variable. Use just INSN_UID for determining whether an insn
23122 should be only disconnected from the insn stream.
23123 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
23125 2013-04-30 Jakub Jelinek <jakub@redhat.com>
23127 PR tree-optimization/57104
23128 * tsan.c (instrument_expr): Don't instrument accesses to
23129 DECL_HARD_REGISTER VAR_DECLs.
23131 2013-04-30 Richard Biener <rguenther@suse.de>
23133 * function.h (loops_for_fn): New inline function.
23134 (set_loops_for_fn): Likewise.
23135 * cfgloop.h (place_new_loop): Add struct function parameter.
23136 (get_loop): Likewise.
23137 (get_loops): Likewise.
23138 (number_of_loops): Likewise.
23139 (fel_next): Adjust.
23140 (fel_init): Likewise.
23141 * cfg.c (get_loop_copy): Adjust.
23142 * cfgloop.c (flow_loops_dump): Likewise.
23143 (record_loop_exits): Likewise.
23144 (verify_loop_structure): Likewise.
23145 * cfgloopanal.c (mark_irreducible_loops): Likewise.
23146 (estimate_reg_pressure_cost): Likewise.
23147 (mark_loop_exit_edges): Likewise.
23148 * cfgloopmanip.c (place_new_loop): Likewise.
23149 (add_loop): Likewise.
23150 (duplicate_loop): Likewise.
23151 * graph.c (draw_cfg_nodes): Likewise.
23152 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
23153 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
23154 (extract_affine_chrec): Likewise.
23155 (build_scop_iteration_domain): Likewise.
23156 * graphite.c (graphite_initialize): Likewise.
23157 * ira-build.c (create_loop_tree_nodes): Likewise.
23158 (more_one_region_p): Likewise.
23159 (rebuild_regno_allocno_maps): Likewise.
23160 (mark_loops_for_removal): Likewise.
23161 (mark_all_loops_for_removal): Likewise.
23162 (remove_unnecessary_regions): Likewise.
23163 (ira_build): Likewise.
23164 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
23165 * loop-init.c (fix_loop_structure): Likewise.
23166 (gate_rtl_move_loop_invariants): Likewise.
23167 (gate_rtl_unswitch): Likewise.
23168 (gate_rtl_unroll_and_peel_loops): Likewise.
23169 (rtl_doloop): Likewise.
23170 * lto-streamer-in.c (input_cfg): Likewise.
23171 * lto-streamer-out.c (output_cfg): Likewise.
23172 * modulo-sched.c (sms_schedule): Likewise.
23173 * predict.c (tree_estimate_probability): Likewise.
23174 (tree_estimate_probability_driver): Likewise.
23175 (estimate_loops): Likewise.
23176 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
23177 (move_sese_region_to_fn): Likewise.
23178 (debug_loop_num): Likewise.
23179 * tree-chrec.c (chrec_evaluate): Likewise.
23180 (hide_evolution_in_other_loops_than_loop): Likewise.
23181 (chrec_component_in_loop_num): Likewise.
23182 (reset_evolution_in_loop): Likewise.
23183 (evolution_function_is_invariant_rec_p): Likewise.
23184 * tree-if-conv.c (main_tree_if_conversion): Likewise.
23185 * tree-inline.c (copy_loops): Likewise.
23186 (copy_cfg_body): Likewise.
23187 (tree_function_versioning): Likewise.
23188 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
23189 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
23191 (add_to_evolution_1): Likewise.
23192 (scev_const_prop): Likewise.
23193 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
23194 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
23195 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
23196 (tree_ssa_lim_initialize): Likewise.
23197 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
23198 (verify_loop_closed_ssa): Likewise.
23199 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
23200 (tree_ssa_loop_im): Likewise.
23201 (tree_ssa_loop_unswitch): Likewise.
23202 (tree_vectorize): Likewise.
23203 (check_data_deps): Likewise.
23204 (tree_ssa_loop_ivcanon): Likewise.
23205 (tree_ssa_loop_bounds): Likewise.
23206 (tree_complete_unroll): Likewise.
23207 (tree_complete_unroll_inner): Likewise.
23208 (tree_parallelize_loops): Likewise.
23209 (tree_ssa_loop_prefetch): Likewise.
23210 (tree_ssa_loop_ivopts): Likewise.
23211 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
23212 * tree-vectorizer.c (vectorize_loops): Likewise.
23214 2013-04-29 Mike Frysinger <vapier@gentoo.org>
23216 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
23217 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
23218 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
23219 with EABI_LINK_SPEC.
23221 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23224 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
23227 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
23230 * lra-constraints.c (process_alt_operands): Discourage a bit more
23231 using memory for pseudos. Print cost dump for alternatives.
23232 Modify cost values for conflicts with early clobbers.
23233 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
23235 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23238 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
23240 2013-04-29 Ian Bolton <ian.bolton@arm.com>
23242 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
23243 from/to S register.
23244 (movdi_aarch64): Support LDR/STR from/to D register.
23246 2013-04-29 Ian Bolton <ian.bolton@arm.com>
23248 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
23249 or higher by default.
23251 2013-04-29 Richard Biener <rguenther@suse.de>
23253 PR middle-end/57075
23254 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
23255 even if not adding abnormal edges for calls that can make
23258 2013-04-29 Richard Biener <rguenther@suse.de>
23260 PR middle-end/57103
23261 * tree-cfg.c (move_stmt_op): Fix condition under which to update
23263 (move_stmt_r): Remove redundant checking.
23265 2013-04-29 Teresa Johnson <tejohnson@google.com>
23268 * basic-block.h (apply_scale): New function.
23269 (apply_probability): Use apply_scale.
23270 * gimple-streamer-in.c (input_bb): Ditto.
23271 * lto-streamer-in.c (input_cfg): Ditto.
23272 * lto-cgraph.c (merge_profile_summaries): Ditto.
23273 * tree-optimize.c (execute_fixup_cfg): Ditto.
23274 * tree-inline.c (copy_bb): Update comment to use apply_scale.
23275 (copy_edges_for_bb): Ditto.
23276 (copy_cfg_body): Ditto.
23278 2013-04-29 Tom de Vries <tom@codesourcery.com>
23280 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
23281 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
23282 (tail_merge_optimize): Handle current_loops == NULL.
23284 2013-04-26 Jeff Law <law@redhat.com>
23286 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
23287 (simplify_cond_using_ranges): Generalize code to simplify
23288 COND_EXPRs where one argument is a constant and the other
23289 is an SSA_NAME created by an integral type conversion.
23291 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23293 * config/arm/arm.md (store_minmaxsi): Use only when
23294 optimize_insn_for_size_p.
23296 2013-04-29 Christian Bruel <christian.bruel@st.com>
23299 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
23301 2013-04-29 Richard Biener <rguenther@suse.de>
23303 PR middle-end/57089
23304 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
23305 loop tree make sure to schedule a fixup for the child as well.
23306 (expand_omp_for_generic): Properly add loops.
23307 (expand_omp_for_static_nochunk): Likewise.
23308 (expand_omp_for_static_chunk): Likewise.
23309 (expand_omp_for): For the degenerate case fixup loops.
23310 (expand_omp_sections): Fix default bb placement in loops.
23311 (expand_omp_atomic_pipeline): Properly add loops.
23313 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23315 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
23317 2013-04-29 Tom de Vries <tom@codesourcery.com>
23319 * tree-ssa-tail-merge.c: Update header comment.
23321 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23323 * config/aarch64/arm_neon.h
23324 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
23325 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
23326 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
23327 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
23328 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
23329 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
23330 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
23331 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
23333 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23335 * config/aarch64/aarch64-simd.md
23336 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
23337 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
23338 fix_trunc, fixuns_trunc.
23339 (ftrunc<VDQF:mode>2): New.
23340 * config/aarch64/iterators.md (optab): Add fix, fixuns.
23341 (fix_trunc_optab): New.
23343 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23345 * config/aarch64/aarch64-builtins.c
23346 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
23347 iceilf, lround, iroundf.
23349 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
23352 * config/i386/i386.h (enum ix86_tune_indices)
23353 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
23354 New, split from X86_TUNE_INTER_UNIT_MOVES.
23355 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
23356 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
23357 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
23358 (TARGET_INTER_UNIT_MOVES): Remove.
23359 * config/i386/i386.c (initial_ix86_tune_features): Update.
23360 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
23361 (ix86_expand_convert_uns_didf_sse): Use
23362 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
23363 (ix86_expand_vector_init_one_nonzero): Ditto.
23364 (ix86_expand_vector_init_interleave): Ditto.
23365 (inline_secondary_memory_needed): Return true for moves from SSE class
23366 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
23367 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
23368 * config/i386/constraints.md (Yi, Ym): Depend on
23369 TARGET_INTER_UNIT_MOVES_TO_VEC.
23370 (Yj, Yn): New constraints.
23371 * config/i386/i386.md (*movdi_internal): Change constraints of
23372 operand 1 from Yi to Yj and from Ym to Yn.
23373 (*movsi_internal): Ditto.
23374 (*movdf_internal): Ditto.
23375 (*movsf_internal): Ditto.
23376 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
23377 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
23378 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
23379 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
23380 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
23381 * config/i386/sse.md (movdi_to_sse): Ditto.
23382 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
23383 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
23384 TARGET_INTER_UNIT_MOVES.
23385 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
23386 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
23387 instead of TARGET_INTER_UNIT_MOVES.
23388 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
23389 operand 1 from Yi to Yj and from Ym to Yn.
23391 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23393 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
23394 (float_truncate_hi_): Likewise.
23395 (float_extend_lo_): Likewise.
23396 (float_truncate_lo_): Likewise.
23397 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
23398 (aarch64_float_extend_lo_v2df): Likewise.
23399 (vec_unpacks_hi_v4sf): Likewise.
23400 (aarch64_float_truncate_lo_v2sf): Likewise.
23401 (aarch64_float_truncate_hi_v4sf): Likewise.
23402 (vec_pack_trunc_v2df): Likewise.
23403 (vec_pack_trunc_df): Likewise.
23405 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23407 * config/aarch64/aarch64-builtins.c
23408 (aarch64_fold_builtin): Fold float conversions.
23409 * config/aarch64/aarch64-simd-builtins.def
23410 (floatv2si, floatv4si, floatv2di): New.
23411 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
23412 * config/aarch64/aarch64-simd.md
23413 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
23414 * config/aarch64/iterators.md (FLOATUORS): New.
23415 (optab): Add float, floatuns.
23416 (su_optab): Likewise.
23418 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23420 * config/aarch64/aarch64-builtins.c
23421 (aarch64_builtin_vectorized_function): Use new names for
23423 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
23424 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
23425 (fcvtzu): Split as...
23426 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
23427 (fcvtas): Split as...
23428 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
23429 (fcvtau): Split as...
23430 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
23431 (fcvtps): Split as...
23432 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
23433 (fcvtpu): Split as...
23434 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
23435 (fcvtms): Split as...
23436 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
23437 (fcvtmu): Split as...
23438 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
23439 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
23440 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
23441 (lfrintnusf, lfrintnudf): Likewise.
23442 * config/aarch64/aarch64-simd.md
23443 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
23445 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
23446 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
23447 (fcvt_pattern): Likewise.
23449 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23451 * config/aarch64/aarch64-simd.md
23452 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
23453 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
23455 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23457 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
23458 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
23459 (vrnd<a,m,n,p>_f32): Implement using builtins.
23460 (vrnd<i,x><q>_f<32, 64>): New.
23462 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
23464 * config/aarch64/aarch64-builtins.c
23465 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
23466 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
23467 (frintz): Rename to...
23469 (frintp): Rename to...
23471 (frintm): Rename to...
23473 (frinti): Rename to...
23474 (nearbyint): ...this.
23475 (frintx): Rename to...
23477 (frinta): Rename to...
23479 * config/aarch64/aarch64-simd.md
23480 (aarch64_frint<frint_suffix><mode>): Delete.
23481 (<frint_pattern><mode>2): Convert to insn.
23482 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
23483 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
23484 (frint_pattern): Likewise.
23485 (frint_suffix): Likewise.
23487 2013-04-29 Richard Biener <rguenther@suse.de>
23489 PR tree-optimization/57081
23490 * loop-init.c: Include tree-flow.h.
23491 (loop_optimizer_finalize): Free number of iteration estimates.
23492 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
23494 2013-04-29 Jakub Jelinek <jakub@redhat.com>
23496 PR tree-optimization/57083
23497 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
23498 non-singleton shift count range, zero extend low_bound for uns case.
23500 * config/i386/predicates.md (general_vector_operand): New predicate.
23501 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
23502 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
23503 if they aren't nonimmediate operands. If their original values
23504 satisfy const_vector_equal_evenodd_p, don't shift them.
23505 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
23506 predicates. For the SSE4.1 case force operands[{1,2}] into registers
23507 if not nonimmediate_operand.
23508 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
23509 instead of register_operand.
23510 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
23512 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
23514 * stor-layout.c (finalize_size_functions): Allocate a structure and
23515 reset cfun before dumping the functions.
23517 2013-04-27 Jakub Jelinek <jakub@redhat.com>
23519 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
23522 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
23523 use xop_pmacsdqh if uns_p.
23524 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
23525 the immediate rotate count.
23527 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
23529 * rtl.h (struct rtx_def): Add comment for field jump.
23530 (LRA_SUBREG_P): New macro.
23531 * recog.c (register_operand): Check LRA_SUBREG_P.
23532 * lra.c (lra): Add note at the end of RTL code. Align non-empty
23534 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
23535 (lra_final_code_change): Skip subreg change for operators.
23536 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
23537 if there are no operand changes.
23538 * lra-constraints.c (curr_insn_set): New.
23539 (match_reload): Set LRA_SUBREG_P.
23540 (emit_spill_move): Ditto.
23541 (check_and_process_move): Use curr_insn_set. Process only single
23542 set insns. Don't initialize sec_mem_p and change_p.
23543 (simplify_operand_subreg): Use LRA_SUBREG_P.
23544 (reg_in_class_p): New function.
23545 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
23546 of #ifdef. Add code to remove cycling.
23547 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
23548 non-null disp. Reload inner instead of disp when base and index
23549 are null. Try to put lo_sum into register.
23550 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
23551 (check_and_process_move): Move code for move cost check to
23552 simple_move_p. Remove equiv_substitution.
23553 (simple_move_p): New function.
23554 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
23555 curr_insn_set. Call check_and_process_move only for single set
23556 insns. Use the new function. Move call of check_and_process_move
23557 after operand equiv substitution and address process.
23559 2013-04-26 Jakub Jelinek <jakub@redhat.com>
23562 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
23563 with nonlocal goto receivers or returns twice calls, ignore
23564 unininitialized values from abnormal edges to nl goto receiver
23565 or returns twice call.
23567 2013-04-26 Jakub Jelinek <jakub@redhat.com>
23569 PR tree-optimization/57051
23570 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
23571 and VEC_RSHIFT_EXPR if shift count is a multiple of element
23574 2013-04-26 Richard Biener <rguenther@suse.de>
23576 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
23577 (expand_omp_taskreg): Likewise. Mark loops for fixup.
23578 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
23579 (fixup_loop_arrays_after_move): New function.
23580 (move_sese_region_to_fn): Properly outline the loop tree parts
23581 of the SESE region.
23583 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
23585 * config/i386/i386.md (type, unit): Fix long lines.
23587 2013-04-26 Richard Biener <rguenther@suse.de>
23589 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
23590 (lto-streamer-out.o): Likewise.
23591 * cfgloop.c (init_loops_structure): Export, add struct function
23592 argument and adjust.
23593 (flow_loops_find): Adjust.
23594 * cfgloop.h (enum loop_estimation): Add EST_LAST.
23595 (init_loops_structure): Declare.
23596 * lto-streamer-in.c: Include cfgloop.h.
23597 (input_cfg): Input the loop tree.
23598 * lto-streamer-out.c: Include cfgloop.h.
23599 (output_cfg): Output the loop tree.
23600 (output_struct_function_base): Do not drop PROP_loops.
23602 2013-03-26 Richard Biener <rguenther@suse.de>
23604 * tree-cfg.c (execute_build_cfg): Build the loop tree.
23605 (pass_build_cfg): Provide PROP_loops.
23606 (move_sese_region_to_fn): Remove loops that are outlined into fn
23608 * tree-inline.c: Include cfgloop.h.
23609 (initialize_cfun): Do not drop PROP_loops.
23610 (copy_loops): New function.
23611 (copy_cfg_body): Copy loop structure.
23612 (tree_function_versioning): Initialize destination loop tree.
23613 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
23614 (pass_parallelize_loops): Do IL verification.
23615 * loop-init.c (loop_optimizer_init): Fixup loops if required.
23616 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
23617 the CFG make sure we fixup loops as well.
23618 * tree-ssa-tail-merge.c: Include cfgloop.h.
23619 (replace_block_by): When merging loop latches mark loops for fixup.
23620 * lto-streamer-out.c (output_struct_function_base): Drop
23621 PROP_loops for now.
23622 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
23623 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
23624 * ipa-split.c: Include cfgloop.h.
23625 (split_function): Add the new return block to the loop tree root.
23626 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
23627 whether we have removed the forwarder block.
23628 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
23629 * cfgloop.h (place_new_loop): Declare.
23630 * cfgloopmanip.c (place_new_loop): Export.
23631 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
23632 (tree-switch-conversion.o): Likewise.
23633 (tree-complex.o): Likewise.
23634 (tree-inline.o): Likewise.
23635 (tree-ssa-tailmerge.o): Likewise.
23636 (ipa-split.o): Likewise.
23637 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
23638 (tree-ssa-copy.o): Likewise.
23639 * tree-switch-conversion.c: Include cfgloop.h
23640 (process_switch): If we emit a bit-test cascade, schedule loops
23642 * tree-complex.c: Include cfgloop.h.
23643 (expand_complex_div_wide): Properly add new basic-blocks to loops.
23644 * asan.c: Include cfgloop.h.
23645 (create_cond_insert_point): Properly add new basic-blocks to
23646 loops, schedule loop fixup.
23647 * cfgloop.c (verify_loop_structure): Check that looks are not
23649 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
23651 (expand_omp_for_generic): Likewise.
23652 (expand_omp_sections): Likewise.
23653 (expand_omp_atomic_pipeline): Schedule loops for fixup.
23654 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
23655 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
23656 is initialized, not when loops are present.
23657 * tree-parloops.c (parallelize_loops): Remove checking here.
23658 * passes.c (init_optimization_passes): Schedule a copy-propagation
23659 pass before complete unrolling of inner loops.
23661 2013-04-26 Jakub Jelinek <jakub@redhat.com>
23663 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
23664 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
23665 (colorize_init): Add argument to _WIN32 version.
23666 * toplev.c: Include diagnostic-color.h.
23667 (process_options): Default to -fdiagnostics-color=auto if
23668 GCC_COLORS env var is in the environment.
23669 * common.opt (fdiagnostics-color=): Add Var and Init.
23670 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
23671 env var is in the environment, the default is auto rather than never.
23673 * diagnostic.h (file_name_as_prefix): Add context argument.
23674 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
23675 the string as locus.
23676 * langhooks.c (lhd_print_error_function): Adjust caller.
23678 2013-04-25 Lawrence Crowl <crowl@google.com>
23680 * var-tracking.c (shared_hash_def::htab):
23681 Change type to hash_table. Update dependent calls and types.
23683 2013-04-25 Lawrence Crowl <crowl@google.com>
23685 * Makefile.in: Update as needed below.
23687 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
23688 Move declaration to after the type's method definitons.
23690 * attribs.c (htab_t scoped_attributes::attribute_hash):
23691 Change type to hash_table. Update dependent calls and types.
23693 * bitmap.c (htab_t bitmap_desc_hash):
23694 Change type to hash_table. Update dependent calls and types.
23696 * cselib.c (htab_t cselib_hash_table):
23697 Change type to hash_table. Update dependent calls and types.
23699 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
23700 (hash_string_slot_node): Move implementation into lto-streamer.h
23701 struct string_slot_hasher.
23702 (eq_string_slot_node): Likewise.
23704 * data-streamer-out.c: Update output_block::string_hash_table
23705 dependent calls and types.
23707 * dwarf2cfi.c (htab_t trace_index):
23708 Change type to hash_table. Update dependent calls and types.
23710 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
23711 Change type to hash_table. Update dependent calls and types.
23712 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
23713 (htab_t optimize_external_refs::map): Likewise.
23714 (htab_t output_comp_unit::extern_map): Likewise.
23715 (htab_t output_comdat_type_unit::extern_map): Likewise.
23716 (htab_t output_macinfo::macinfo_htab): Likewise.
23717 (htab_t optimize_location_lists::htab): Likewise.
23718 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
23720 * except.c (htab_t ehspec_hash_type):
23721 Change type to hash_table. Update dependent calls and types.
23722 (assign_filter_values::ttypes): Likewise.
23723 (assign_filter_values::ehspec): Likewise.
23724 (sjlj_assign_call_site_values::ar_hash): Likewise.
23725 (convert_to_eh_region_ranges::ar_hash): Likewise.
23727 * gcse.c (htab_t pre_ldst_table):
23728 Change type to hash_table. Update dependent calls and types.
23730 * ggc-common.c (htab_t saving_htab):
23731 Change type to hash_table. Update dependent calls and types.
23732 (htab_t loc_hash): Likewise.
23733 (htab_t ptr_hash): Likewise.
23734 (call_count): Rename ggc_call_count.
23735 (call_alloc): Rename ggc_call_alloc.
23736 (loc_descriptor): Rename make_loc_descriptor.
23737 (add_statistics): Rename ggc_add_statistics.
23739 * ggc-common.c (saving_htab):
23740 Change type to hash_table. Update dependent calls and types.
23742 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
23743 (push_gimplify_context): Likewise.
23744 (pop_gimplify_context): Likewise.
23745 (struct gimple_temp_hash_elt): Added.
23746 (struct gimplify_hasher): Likewise.
23747 (struct gimplify_ctx.temp_htab):
23748 Change type to hash_table. Update dependent calls and types.
23750 * gimple-fold.c: Include gimplify-ctx.h.
23752 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
23753 Change type to hash_table. Update dependent calls and types.
23754 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
23755 avoid potential global name collision.
23757 * gimplify.c: Include gimplify-ctx.h.
23758 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
23759 (htab_t gimplify_ctx::temp_htab):
23760 Update dependent calls and types for new type hash_table.
23761 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
23762 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
23764 * gimplify-ctx.h: New.
23765 (struct gimple_temp_hash_elt): Move from gimplify.c.
23766 (class gimplify_hasher): New.
23767 (struct gimplify_ctx): Move from gimple.h.
23768 (htab_t gimplify_ctx::temp_htab):
23769 Change type to hash_table. Update dependent calls and types.
23771 * graphite-clast-to-gimple.c: Include graphite-htab.h.
23772 (htab_t ivs_params::newivs_index):
23773 Change type to hash_table. Update dependent calls and types.
23774 (htab_t ivs_params::params_index): Likewise.
23775 (htab_t print_generated_program::params_index): Likewise.
23776 (htab_t gloog::newivs_index): Likewise.
23777 (htab_t gloog::params_index): Likewise.
23779 * graphite.c: Include graphite-htab.h.
23780 4htab_t graphite_transform_loops::bb_pbb_mapping):
23781 Change type to hash_table. Update dependent calls and types.
23783 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
23784 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
23785 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
23787 * graphite-dependences.c: Include graphite-htab.h.
23788 (loop_is_parallel_p): Change hash table type of parameter.
23790 * graphite-htab.h: New.
23791 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
23792 (extern find_pbb_via_hash): Move from graphite-poly.h.
23793 (extern loop_is_parallel_p): Move from graphite-poly.h.
23794 (extern get_loop_body_pbbs): Move from graphite-poly.h.
23796 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
23797 (extern loop_is_parallel_p): Move to graphite-htab.h.
23798 (extern get_loop_body_pbbs): Move to graphite-htab.h.
23800 * haifa-sched.c (htab_t delay_htab):
23801 Change type to hash_table. Update dependent calls and types.
23802 (htab_t delay_htab_i2): Likewise.
23804 * ira-color.c (htab_t allocno_hard_regs_htab):
23805 Change type to hash_table. Update dependent calls and types.
23807 * ira-costs.c (htab_t cost_classes_htab):
23808 Change type to hash_table. Update dependent calls and types.
23810 * loop-invariant.c (htab_t merge_identical_invariants::eq):
23811 Change type to hash_table. Update dependent calls and types.
23813 * loop-iv.c (htab_t bivs):
23814 Change type to hash_table. Update dependent calls and types.
23816 * loop-unroll.c (htab_t opt_info::insns_to_split):
23817 Change type to hash_table. Update dependent calls and types.
23818 (htab_t opt_info::insns_with_var_to_expand): Likewise.
23820 * lto-streamer.h (struct string_slot): Move from data-streamer.h
23821 (struct string_slot_hasher): New.
23822 (htab_t output_block::string_hash_table):
23823 Change type to hash_table. Update dependent calls and types.
23825 * lto-streamer-in.c (freeing_string_slot_hasher): New.
23826 (htab_t file_name_hash_table):
23827 Change type to hash_table. Update dependent calls and types.
23829 * lto-streamer-out.c: Update output_block::string_hash_table dependent
23832 * lto-streamer.c (htab_t tree_htab):
23833 Change type to hash_table. Update dependent calls and types.
23835 * omp-low.c: Include gimplify-ctx.h.
23837 * passes.c (htab_t name_to_pass_map):
23838 Change type to hash_table. Update dependent calls and types.
23839 (pass_traverse): Rename to passes_pass_traverse.
23841 * plugin.c (htab_t event_tab):
23842 Change type to hash_table. Update dependent calls and types.
23844 * postreload-gcse.c (htab_t expr_table):
23845 Change type to hash_table. Update dependent calls and types.
23846 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
23848 * sese.c (debug_rename_map_1): Make extern.
23849 (htab_t copy_bb_and_scalar_dependences::rename_map):
23850 Change type to hash_table. Update dependent calls and types.
23852 * sese.h (extern debug_rename_map): Move to .c file.
23854 * store-motion.c (htab_t store_motion_mems_table):
23855 Change type to hash_table. Update dependent calls and types.
23857 * trans-mem.c (htab_t tm_new_mem_hash):
23858 Change type to hash_table. Update dependent calls and types.
23860 * tree-browser.c (htab_t TB_up_ht):
23861 Change type to hash_table. Update dependent calls and types.
23863 * tree-cfg.c (htab_t discriminator_per_locus):
23864 Change type to hash_table. Update dependent calls and types.
23866 * tree-complex.c: Include tree-hasher.h
23867 (htab_t complex_variable_components):
23868 Change type to hash_table. Update dependent calls and types.
23870 * tree-eh.c (htab_t finally_tree):
23871 Change type to hash_table. Update dependent calls and types.
23873 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
23874 struct int_tree_hasher.
23875 (extern int_tree_map_eq): Likewise.
23876 (uid_decl_map_hash): Removed.
23877 (extern decl_tree_map_eq): Likewise.
23879 * tree-hasher.h: New.
23880 (struct int_tree_hasher): New.
23881 (typedef int_tree_htab_type): New.
23883 * tree-inline.c: Include gimplify-ctx.h.
23885 * tree-mudflap.c: Include gimplify-ctx.h.
23887 * tree-parloops.c: Include tree-hasher.h.
23888 (htab_t eliminate_local_variables_stmt::decl_address):
23889 Change type to hash_table. Update dependent calls and types.
23890 (htab_t separate_decls_in_region::decl_copies): Likewise.
23892 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
23893 Change type to hash_table. Update dependent calls and types.
23895 * tree-sra.c (candidates):
23896 Change type to hash_table. Update dependent calls and types.
23898 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
23900 (int_tree_map_hash): Likewise.
23902 * tree-ssa-dom.c (htab_t avail_exprs):
23903 Change type to hash_table. Update dependent calls and types.
23905 * tree-ssa-live.c (var_map_base_init::tree_to_index):
23906 Change type to hash_table. Update dependent calls and types.
23908 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
23909 Change type to hash_table. Update dependent calls and types.
23911 * tree-ssa-phiopt.c (seen_ssa_names):
23912 Change type to hash_table. Update dependent calls and types.
23914 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
23915 Change type to hash_table. Update dependent calls and types.
23917 * tree-ssa-uncprop.c (equiv):
23918 Change type to hash_table. Update dependent calls and types.
23920 2013-04-25 Jakub Jelinek <jakub@redhat.com>
23922 PR rtl-optimization/57003
23923 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
23924 call note_stores with kill_clobbered_value callback again after
23925 killing regs_invalidated_by_call.
23927 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23929 * config/aarch64/aarch64-simd.md
23930 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
23931 (aarch64_simd_bsl<mode>): Likewise.
23932 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
23934 2013-04-25 Marek Polacek <polacek@redhat.com>
23936 PR tree-optimization/57066
23937 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
23939 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23941 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
23943 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23945 * config/aarch64/aarch64-builtins.c
23946 (aarch64_fold_builtin): New.
23947 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
23948 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
23949 * config/aarch64/aarch64-simd-builtins.def (abs): New.
23950 * config/aarch64/arm_neon.h
23951 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
23953 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23954 Tejas Belagod <tejas.belagod@arm.com>
23956 * config/aarch64/aarch64-builtins.c
23957 (aarch64_gimple_fold_builtin): New.
23958 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
23959 * config/aarch64/aarch64-simd-builtins.def (addv): New.
23960 * config/aarch64/aarch64-simd.md (addpv4sf): New.
23961 (addvv4sf): Update.
23962 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
23964 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23966 * config/aarch64/aarch64.md
23967 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
23969 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23971 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
23972 (*ngcsi_uxtw): New pattern.
23974 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23975 Julian Brown <julian@codesourcery.com>
23977 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
23978 (TB_DREG): Add T_V4HF.
23979 (v4hf_UP): New macro.
23980 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
23981 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
23982 Handle initialisation of V4HF. Adjust initialisation of reinterpret
23984 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
23985 (arm_vector_mode_supported_p): Handle V4HF.
23986 (arm_mangle_map): Handle V4HFmode.
23987 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
23988 * config/arm/arm_neon_builtins.def: Add entries for
23989 vcvtv4hfv4sf, vcvtv4sfv4hf.
23990 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
23991 (neon_vcvtv4hfv4sf): Likewise.
23992 * config/arm/neon-gen.ml: Handle half-precision floating point
23994 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
23995 * config/arm/arm_neon.h: Regenerate.
23996 * config/arm/neon.ml (type elts): Add F16.
23997 (type vectype): Add T_float16x4, T_floatHF.
23998 (type vecmode): Add V4HF.
23999 (type features): Add Requires_FP_bit feature.
24000 (elt_width): Handle F16.
24001 (elt_class): Likewise.
24002 (elt_of_class_width): Likewise.
24003 (mode_of_elt): Refactor.
24004 (type_for_elt): Handle F16, fix error messages.
24005 (vectype_size): Handle T_float16x4.
24006 (vcvt_sh): New function.
24007 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
24008 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
24009 (string_of_mode): Handle V4HF.
24010 * doc/arm-neon-intrinsics.texi: Regenerate.
24012 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
24014 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
24015 format specifier in 'X' case.
24017 2013-04-25 Alan Modra <amodra@gmail.com>
24020 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
24021 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
24022 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
24023 Repeat for many other rotate/shift and mask patterns using subregs.
24024 Name lshiftrt insns.
24025 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
24026 on WORDS_BIG_ENDIAN.
24028 2013-04-25 Alan Modra <amodra@gmail.com>
24030 * config.gcc: Support little-endian powerpc-linux targets.
24031 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
24032 (LINK_OS_LINUX_SPEC): Define.
24033 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
24034 Preserve MASK_LITTLE_ENDIAN.
24035 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
24036 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
24037 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
24038 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
24039 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
24040 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
24041 Correct fp word order for little-endian. Don't shift toc entries
24042 smaller than a word for little-endian.
24043 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
24044 (bswapdi2 splits): Correct low-part subreg for little-endian.
24045 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
24046 low/high where such is correct only for be.
24047 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
24048 little-endian for -mcall-aixdesc.
24050 2013-04-25 Alan Modra <amodra@gmail.com>
24052 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
24053 replace_equiv_address_nv.
24055 2013-04-25 Alan Modra <amodra@gmail.com>
24057 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
24059 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24062 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24063 * rtl.h (struct rtx_def): ...
24065 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24067 PR rtl-optimizations/57046
24068 * lra-constraints (split_reg): Set up lra_risky_transformations_p
24069 for multi-reg splits.
24071 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
24073 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
24075 2013-04-24 Sterling Augustine <saugustine@google.com>
24077 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
24078 (comp_dir_string, debug_str_dwo_section): New.
24079 (DEBUG_STR_DWO_SECTION): Rename to ...
24080 (DEBUG_DWO_STR_SECTION): ... this.
24081 (DEBUG_NORM_STR_SECTION): Delete.
24082 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
24083 (DEBUG_STR_DWO_SECTION_FLAGS): New.
24084 (find_AT_string): Move most logic to ...
24085 (find_AT_string_in_table): ... here. New.
24086 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
24087 add_skeleton_AT_string. Delete logic.
24088 (output_skeleton_debug_sections): Remove call to
24089 add_top_level_skeleton_die_attrs.
24090 (add_comp_dir_attribute): Move logic to comp_dir_string.
24091 (dwarf2out_init): Initialize debug_str_dwo_section.
24092 (output_indirect_string): Call find_string_form.
24093 (output_indirect_strings): Rewrite.
24094 (prune_unused_types): Empty skeleton_debug_str_hash.
24095 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
24096 (dwarf2out_finish): Call output_indirect_strings.
24098 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
24100 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
24102 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
24104 * rtl.h (struct rtx_def): Add comment for field jump.
24105 (LRA_SUBREG_P): New macro.
24106 * recog.c (register_operand): Check LRA_SUBREG_P.
24107 * lra.c (lra): Add note at the end of RTL code. Align non-empty
24109 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
24110 (lra_final_code_change): Skip subreg change for operators.
24111 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
24112 if there are no operand changes.
24113 * lra-constraints.c (curr_insn_set): New.
24114 (match_reload): Set LRA_SUBREG_P.
24115 (emit_spill_move): Ditto.
24116 (check_and_process_move): Use curr_insn_set. Process only single
24117 set insns. Don't initialize sec_mem_p and change_p.
24118 (simplify_operand_subreg): Use LRA_SUBREG_P.
24119 (reg_in_class_p): New function.
24120 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
24121 of #ifdef. Add code to remove cycling.
24122 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
24123 non-null disp. Reload inner instead of disp when base and index
24124 are null. Try to put lo_sum into register.
24125 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
24126 (check_and_process_move): Move code for move cost check to
24127 simple_move_p. Remove equiv_substitution.
24128 (simple_move_p): New function.
24129 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
24130 curr_insn_set. Call check_and_process_move only for single set
24131 insns. Use the new function. Move call of check_and_process_move
24132 after operand equiv substitution and address process.
24134 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
24136 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
24137 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
24138 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
24140 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
24142 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
24144 2013-04-24 Marek Polacek <polacek@redhat.com>
24146 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
24147 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
24148 (select_loops_exit_conditions): Likewise.
24149 (number_of_iterations_for_all_loops): Likewise.
24150 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
24151 (scev_analysis): Likewise.
24153 2013-04-02 Catherine Moore <clm@codesourcery.com>
24154 Chao-ying Fu <fu@mips.com>
24156 * config/mips/micromips.md (jraddiusp): New pattern.
24157 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
24158 instruction if possible.
24160 2013-04-24 Alan Modra <amodra@gmail.com>
24162 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
24164 2013-04-24 Julian Brown <julian@codesourcery.com>
24165 Chung-Lin Tang <cltang@codesourcery.com>
24167 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
24168 dependency behavior in enumeration type DIE generation. Add TODO note
24169 to comments about future DW_FORM_sdata/udata re-work of related code.
24171 2013-04-23 Lawrence Crowl <crowl@google.com>
24173 * Makefile.in: Update as needed below.
24175 * hash-table.h (class hash_table):
24176 Correct many methods with parameter types compare_type to the correct
24177 value_type. (Correct code was unlikely to notice the change.)
24178 (hash_table::elements_with_deleted) New.
24179 (class hashtable::iterator): New.
24180 (hashtable::begin()): New.
24181 (hashtable::end()): New.
24182 (FOR_EACH_HASH_TABLE_ELEMENT): New.
24184 * statistics.c (statistics_hashes):
24185 Change type to hash_table. Update dependent calls and types.
24187 * tree-into-ssa.c (var_infos):
24188 Change type to hash_table. Update dependent calls and types.
24190 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
24191 Change type to hash_table. Update dependent calls and types.
24193 * tree-ssa-loop-im.c (struct mem_ref.refs):
24194 Change type to hash_table. Update dependent calls and types.
24196 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
24197 Change type to hash_table. Update dependent calls and types.
24199 * tree-ssa-sccvn.c (vn_tables_s::nary):
24200 Change type to hash_table. Update dependent calls and types.
24201 (vn_tables_s::phis): Likewise.
24202 (vn_tables_s::references): Likewise.
24204 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
24205 (vn_reference_eq): Update parameter and return types.
24207 * tree-ssa-structalias.c (pointer_equiv_class_table):
24208 Change type to hash_table. Update dependent calls and types.
24209 (location_equiv_class_table): Likewise.
24211 * tree-vect-data-refs.c: Consequential changes for making
24212 peeling a hash_table.
24214 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
24215 (destroy_loop_vec_info): Dependent hash_table update.
24217 * tree-vectorizer.h (peeling_htab):
24218 Change type to hash_table. Update dependent calls and types.
24220 2013-04-23 Shiva Chen <shiva0217@gmail.com>
24222 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
24223 to check the register content is equal or not.
24224 * lra-constraints.c (match_reload): Use lra_assign_reg_val
24225 to assign register content record.
24226 * lra-eliminations.c (update_reg_eliminate): Use
24227 lra_update_reg_val_offset to update register content offset.
24228 * lra-int.h (struct lra_reg): Add offset member.
24229 (lra_reg_val_equal_p): New static inline function.
24230 (lra_update_reg_val_offset): New static inline function.
24231 (lra_assign_reg_val): New static inline function.
24232 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
24233 to assign register content record.
24234 (initialize_lra_reg_info_element): Initial offset to zero.
24236 2013-04-23 Catherine Moore <clm@codesourcery.com>
24238 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
24239 operands. Record compression.
24241 2013-04-23 Xinliang David Li <davidxl@google.com>
24243 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
24245 2013-04-23 Richard Biener <rguenther@suse.de>
24247 PR middle-end/57036
24248 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
24249 parameter, only add abnormal goto edges from the copied body
24250 if the call could perform abnormal gotos.
24251 (copy_cfg_body): Adjust.
24253 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
24255 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
24257 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
24259 * coretypes.h (gimple_stmt_iterator): Add struct to make
24262 2013-04-23 Richard Biener <rguenther@suse.de>
24264 PR tree-optimization/57026
24265 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
24266 from SSA names occuring in abnormal PHI nodes.
24268 2013-04-22 Andi Kleen <ak@linux.intel.com>
24270 * lto/lto.c (print_lto_report_1): Fix LTO report names.
24272 2013-04-22 Andi Kleen <ak@linux.intel.com>
24274 * lto/lto.c (print_lto_report_1): Declare early.
24275 (read_cgraph_and_symbols): Call print_lto_report_1 early.
24277 2013-04-22 Andi Kleen <ak@linux.intel.com>
24279 * common.opt (-flto-report-wpa): Add.
24280 * doc/invoke.texi (-flto-report-wpa): Add.
24281 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
24284 2013-04-22 Xinliang David Li <davidxl@google.com>
24286 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
24287 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
24288 * Makefile.in: New dependency
24290 David Daney <ddaney.cavm@gmail.com>
24292 * configure.ac (gcc_cv_as_micromips_support): Use the
24293 --fatal-warnings option.
24294 * configure: Regenerate.
24296 2013-04-22 Marek Polacek <polacek@redhat.com>
24299 * tsan.c (instrument_expr): Don't instrument expression
24300 in case its size is zero.
24302 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
24306 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
24308 * config/alpha/alpha.c (TARGET_LRA_P): New define.
24310 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24312 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
24313 (gimple_stmt_iterator): New typedef.
24314 * gimple.h (gimple_stmt_iterator): Rename to...
24315 (gimple_stmt_iterator_d): ... This.
24316 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
24317 trees be valid for GIMPLE and GENERIC.
24318 (TARGET_GIMPLE_FOLD_BUILTIN): New.
24319 * gimple-fold.c (gimple_fold_call): Call target hook
24320 gimple_fold_builtin.
24321 * hooks.c (hook_bool_gsiptr_false): New.
24322 * hooks.h (hook_bool_gsiptr_false): New.
24323 * target.def (fold_stmt): New.
24324 * doc/tm.texi: Regenerate.
24326 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
24329 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
24330 a set sp if no stack realignment.
24332 2013-04-22 Nick Clifton <nickc@redhat.com>
24334 * config.gcc (tilegx-linux): Extend extra_objs rather than
24336 (tilepro-linux): Likewise.
24338 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24340 * config/aarch64/aarch64-builtins.c
24342 (CF0, CF1, CF2, CF3, CF4, CF10): New.
24343 (VAR<1-12>): Add MAP parameter.
24344 (BUILTIN_*): Likewise.
24345 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
24346 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
24347 (aarch64_ushl_n<mode>): Likewise.
24348 (aarch64_sshr_n<mode>): Likewise.
24349 (aarch64_ushr_n<mode>): Likewise.
24350 (aarch64_<maxmin><mode>): Likewise.
24351 (aarch64_sqrt<mode>): Likewise.
24352 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
24353 (vshr<q>_n_*): Likewise.
24355 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
24357 * config/aarch64/aarch64-builtins.c
24358 (aarch64_simd_builtin_type_mode): Handle SF types.
24360 (BUILTIN_GPF): Define.
24361 (aarch64_init_simd_builtins): Handle SF types.
24362 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
24363 (frecps): Likewise.
24364 (frecpx): Likewise.
24365 * config/aarch64/aarch64-simd.md
24366 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
24367 (aarch64_frecpe<mode>): New.
24368 (aarch64_frecps<mode>): Likewise.
24369 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
24370 (v8type): Add frecp<esx>.
24371 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
24372 (aarch64_frecps<mode>): Likewise.
24373 * config/aarch64/iterators.md (FRECP): New.
24374 (frecp_suffix): Likewise.
24375 * config/aarch64/arm_neon.h
24376 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
24378 2013-04-22 Christian Bruel <christian.bruel@st.com>
24381 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
24382 (REG_CLASS_NAMES): Idem.
24383 (REG_CLASS_CONTENTS): Idem.
24384 (REGCLASS_HAS_FP_REG): Idem.
24385 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
24386 (sh_conditional_register_usage): Idem.
24388 2013-04-21 Jeff Law <law@redhat.com>
24390 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
24391 (ssa_forward_propagate_and_combine): Use it.
24393 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
24395 * lra.c: Update the flow chart diagram.
24397 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
24399 PR rtl-optimization/56847
24400 * lra-constraints.c (process_alt_operands): Discourage alternative
24401 with non-matche doffsettable memory constraint fro memory with
24404 2013-04-19 Richard Biener <rguenther@suse.de>
24406 PR tree-optimization/56982
24407 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
24409 * gimplify.c (gimplify_call_expr): Notice special calls.
24410 (gimplify_modify_expr): Likewise.
24411 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
24412 abnormal control flow receivers.
24413 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
24414 in the same way as cfun->has_nonlocal_labels.
24415 (gimple_purge_dead_abnormal_call_edges): Likewise.
24416 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
24417 receivers start a basic-block.
24419 2013-04-19 Richard Biener <rguenther@suse.de>
24421 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
24423 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
24424 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
24425 (SLP_TREE_LOAD_PERMUTATION): Add.
24426 (vect_transform_slp_perm_load): Adjust prototype.
24427 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
24428 (vect_free_slp_instance): Likewise.
24429 (vect_create_new_slp_node): Likewise.
24430 (vect_supported_slp_permutation_p): Remove.
24431 (vect_slp_rearrange_stmts): Adjust.
24432 (vect_supported_load_permutation_p): Likewise. Inline
24433 vect_supported_slp_permutation_p here.
24434 (vect_analyze_slp_instance): Compute load permutations per
24435 slp node instead of per instance.
24436 (vect_get_slp_defs): Adjust.
24437 (vect_transform_slp_perm_load): Likewise.
24438 (vect_schedule_slp_instance): Remove redundant code.
24439 (vect_schedule_slp): Remove hack for PR56270, add it ...
24440 * tree-vect-stmts.c (vectorizable_load): ... here, do not
24441 CSE loads for SLP. Adjust.
24443 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
24445 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
24446 spelling in two comments.
24448 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
24451 * config/arm/arm.c (load_multiple_sequence): Require SP
24452 as base register for loads if SP is in the register list.
24454 2013-04-19 Martin Jambor <mjambor@suse.cz>
24456 PR tree-optimization/56718
24457 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
24458 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
24459 and made public. Adjusted all callers.
24460 (ipa_intraprocedural_devirtualization): New function.
24461 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
24462 (ipa_intraprocedural_devirtualization): Likewise.
24463 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
24465 2013-04-19 Richard Biener <rguenther@suse.de>
24467 PR tree-optimization/57000
24468 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
24470 2013-04-19 Terry Guo <terry.guo@arm.com>
24472 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
24474 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
24475 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
24476 (cortex_m4_fmacs): Use new reservations.
24477 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
24479 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
24481 PR rtl-optimization/56999
24482 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
24484 (lra_coalesce): Remove split_origin_bitmap and related code.
24485 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
24486 ranges if necessary.
24488 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
24490 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
24492 (ix86_expand_call): Remove clobbered_registers array and use
24493 x86_64_ms_sysv_extra_clobbered_registers instead.
24494 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
24496 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
24498 * config/i386/i386.md (*call_rex64_ms_sysv): Use
24499 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
24500 (*call_value_rex64_ms_sysv): Ditto.
24502 2013-04-18 Cary Coutant <ccoutant@google.com>
24504 * dwarf2out.c (output_pubnames): Check die_perennial_p of
24505 parent instead of die_mark.
24507 2013-04-18 Diego Novillo <dnovillo@google.com>
24509 * gimple.c (create_gimple_tmp): New.
24510 (get_expr_type): New.
24511 (build_assign): New.
24512 (build_type_cast): New.
24513 * gimple.h (enum ssa_mode): Define.
24514 (gimple_seq_set_location): New.
24515 * asan.c (build_check_stmt): Change some gimple_build_* calls
24516 to use build_assign and build_type_cast.
24518 2013-04-18 Richard Biener <rguenther@suse.de>
24520 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
24521 handle negative step. Remove redundant checks.
24522 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
24523 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
24524 for negative step and grouped loads fail to vectorize.
24526 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
24528 * emit-rtl.c (reset_insn_used_flags): New function.
24529 (reset_all_used_flags): Use it.
24530 (verify_insn_sharing): New function.
24531 (verify_rtl_sharing): Fix verification for SEQUENCEs.
24533 2013-04-18 Jakub Jelinek <jakub@redhat.com>
24535 PR tree-optimization/56984
24536 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
24537 and (x >> M) >= N don't register any assertion if N << M is the
24540 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
24542 * lower-subreg.c (resolve_simple_move): If called self-recursive,
24543 do not delete_insn insns that have not yet been emitted, only
24544 unlink them with remove_insn.
24545 * df-scan.c (df_insn_delete): Revert r197492.
24547 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
24549 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
24550 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
24552 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
24554 * config/arm/arm.md (movsicc_insn): Convert define_insn into
24555 define_insn_and_split.
24556 (and_scc,ior_scc,negscc): Likewise.
24557 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
24559 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
24561 * config/arm/arm.c (use_return_insn): Return 0 for targets that
24562 can benefit from using a sequence of LDRD instructions in epilogue
24563 instead of a single LDM instruction.
24565 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
24568 * doc/extend.texi: Fix typo.
24570 2013-04-17 Richard Biener <rguenther@suse.de>
24572 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
24573 (vect_build_slp_tree): ... here.
24574 (vect_build_slp_tree_1): Compute which stmts of the SLP group
24575 match. Remove special-casing of mismatched complex loads.
24576 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
24577 re-try the match with swapped commutative operands.
24578 (vect_supported_load_permutation_p): Remove special-casing of
24579 mismatched complex loads.
24580 (vect_analyze_slp_instance): Adjust.
24582 2013-04-17 Richard Biener <rguenther@suse.de>
24584 PR rtl-optimization/56921
24585 * cfgloop.h (struct loop): Add simple_loop_desc member.
24586 (struct niter_desc): Mark with GTY(()).
24587 (simple_loop_desc): Do not use aux field but simple_loop_desc.
24588 * loop-iv.c (get_simple_loop_desc): Likewise.
24589 (free_simple_loop_desc): Likewise.
24592 2013-04-16 Richard Biener <rguenther@suse.de>
24594 PR rtl-optimization/56921
24595 * loop-init.c (pass_rtl_move_loop_invariants): Add
24596 TODO_do_not_ggc_collect to todo_flags_finish.
24597 (pass_rtl_unswitch): Same.
24598 (pass_rtl_unroll_and_peel_loops): Same.
24599 (pass_rtl_doloop): Same.
24601 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
24603 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
24604 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
24605 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
24607 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
24608 * tree-streamer.c (record_common_node): Adjust reference in comment.
24610 2013-04-17 Terry Guo <terry.guo@arm.com>
24612 * config/arm/cortex-m4.md: Add a new bypass.
24614 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24616 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
24618 (*subs_<optab><mode>_multp2): New pattern.
24619 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
24620 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
24622 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24624 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
24625 (*subs_mul_imm_<mode>): New pattern.
24627 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
24630 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
24631 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
24632 (vsx_movti_32bit): Change j->wa to O->wa.
24634 2013-04-16 Richard Biener <rguenther@suse.de>
24636 PR rtl-optimization/56921
24637 * loop-init.c (pass_rtl_move_loop_invariants): Add
24638 TODO_do_not_ggc_collect to todo_flags_finish.
24639 (pass_rtl_unswitch): Same.
24640 (pass_rtl_unroll_and_peel_loops): Same.
24641 (pass_rtl_doloop): Same.
24643 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
24645 * config/arm/arm.c (emit_multi_reg_push): New declaration
24646 for an existing function.
24647 (arm_emit_strd_push): New function.
24648 (arm_expand_prologue): Used here.
24649 (arm_emit_ldrd_pop): New function.
24650 (arm_expand_epilogue): Used here.
24651 (arm_get_frame_offsets): Update condition.
24652 (arm_emit_multi_reg_pop): Add a special case for load of a single
24653 register with writeback.
24655 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
24657 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
24660 2013-04-16 Richard Biener <rguenther@suse.de>
24662 PR tree-optimization/56756
24663 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
24664 (first_mem_ref_loc): New.
24665 (execute_sm): Place the load temporarily before a previous
24666 access instead of in the latch edge to ensure its SSA dependencies
24667 are defined at points dominating the load.
24669 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
24671 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
24672 correct fix by moving header and footer insn to the footer of
24673 the merged basic block. Clear BB_END of the merged-away block.
24675 PR middle-end/43631
24676 * emit-rtl.c (make_note_raw): New function.
24677 (link_insn_into_chain): New static inline function.
24678 (add_insn): Use it.
24679 (add_insn_before, add_insn_after): Factor insn chain linking code...
24680 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
24681 using link_insn_into_chain.
24682 (note_outside_basic_block_p): New helper function for emit_note_after
24683 and emit_note_before.
24684 (emit_note_after): Use nobb variant of add_insn_after if the note
24685 should not be contained in a basic block.
24686 (emit_note_before): Use nobb variant of add_insn_before if the note
24687 should not be contained in a basic block.
24688 (emit_note_copy): Use make_note_raw.
24689 (emit_note): Likewise.
24690 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
24691 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
24692 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
24693 the moved barrier the tail of the basic block it follows.
24694 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
24696 2013-04-15 Jakub Jelinek <jakub@redhat.com>
24698 PR tree-optimization/56962
24699 * gimple-ssa-strength-reduction.c (record_increment): Only set
24700 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
24701 either rhs1 or rhs2 is equal to c->base_expr.
24703 2013-04-15 Richard Biener <rguenther@suse.de>
24705 PR tree-optimization/56933
24706 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
24708 (GROUP_READ_WRITE_DEPENDENCE): Remove.
24709 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
24710 * tree-vect-data-refs.c (vect_analyze_group_access): Move
24711 dependence check ...
24712 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
24714 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
24715 GROUP_READ_WRITE_DEPENDENCE.
24717 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24719 * emit-rtl.c (reset_all_used_flags): New function.
24720 (verify_rtl_sharing): Call reset_all_used_flags before and after
24721 performing the checks.
24723 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24725 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
24726 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
24727 * config/arm/constraints.md (De): New constraint.
24728 * config/arm/neon.md (anddi3_neon): Delete.
24729 (neon_vand<mode>): Expand to standard anddi3 pattern.
24730 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
24731 Move earlier in the file.
24732 (neon_inv_logic_op2): Likewise.
24733 (arm_anddi_operand_neon): New predicate.
24735 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24737 * configure.ac (gcc_cv_ld_as_needed): Set
24738 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
24739 Use -z ignore, -z record on *-*-solaris2*.
24740 (HAVE_LD_AS_NEEDED): Update comment.
24741 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
24742 * configure: Regenerate.
24743 * config.in: Regenerate.
24744 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
24745 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
24746 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
24747 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
24748 equivalents. Fix markup.
24749 * doc/tm.texi: Regenerate.
24751 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
24753 * config/i386/i386.opt: New option mstack-protector-guard=.
24754 * config/i386/i386-opts.h: Add enum stack_protector_guard.
24755 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
24756 TARGET_SSP_TLS_GUARD.
24757 * config/i386/i386.c (ix86_option_override_internal): Set
24758 ix86_stack_protector_guard.
24759 * config/i386/i386.md (stack_protect_set): Enable for
24760 TARGET_SSP_TLS_GUARD only.
24761 (stack_protect_set_<mode>): Ditto.
24762 (stack_protect_test): Ditto.
24763 (stack_protect_test_<mode>): Ditto.
24764 * doc/invoke.texi (i386 Option): Document.
24766 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
24769 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
24770 (S_MODES): Set H_MODE bit.
24771 (SF_MODES): Set only S_MODE and SF_MODE bits.
24772 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
24773 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
24774 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
24775 <MODE_FLOAT>: Likewise.
24777 2013-04-15 Joey Ye <joey.ye@arm.com>
24779 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
24781 2013-04-15 Joey Ye <joey.ye@arm.com>
24783 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
24785 (thumb_far_jump_used_p): Count instruction size and set
24788 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
24790 * reorg.c (fill_simple_delay_slots): Reindent block of code.
24791 * resource.c (mark_target_live_regs): Reformat conditional block.
24793 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
24795 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
24796 notes, they are emitted only just before final.
24797 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
24799 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
24801 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
24802 * cfgrtl.c (delete_insn): Call it here instead.
24803 * lra-spills.c (lra_final_code_change): Use delete_insn.
24804 * haifa-sched.c (sched_remove_insn): Likewise.
24805 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
24806 returning to the nop pool.
24807 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
24808 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
24810 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
24812 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
24813 * doc/tm.texi: Regenerated.
24815 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
24817 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
24820 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
24822 * df-core.c (df_find_def): Compare register numbers.
24823 (df_find_use): Likewise.
24825 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
24828 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
24829 lra_in_progress for return.
24831 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
24833 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
24834 define_insn into define_insn_and_split and emit movsicc patterns.
24836 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
24838 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
24840 2013-04-12 Richard Biener <rguenther@suse.de>
24842 * tree-pass.h (TODO_do_not_ggc_collect): New.
24843 * passes.c (execute_one_ipa_transform_pass): Honor
24844 TODO_do_not_ggc_collect.
24845 (execute_one_pass): Likewise.
24848 2013-04-10 Richard Biener <rguenther@suse.de>
24850 * passes.c (init_optimization_passes): Remove reload pass.
24851 * ira.c (do_reload): Merge into ...
24853 (rest_of_handle_reload): Remove.
24854 (pass_reload): Likewise.
24855 * config/i386/i386.c (ix86_option_override): Refer to ira instead
24856 of reload for vzeroupper pass placement.
24858 2013-04-12 Jakub Jelinek <jakub@redhat.com>
24860 PR tree-optimization/56918
24861 PR tree-optimization/56920
24862 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
24863 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
24864 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
24865 use wide_mul_with_sign method.
24867 2013-04-12 Richard Biener <rguenther@suse.de>
24869 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
24870 not be considered a gimple constant.
24872 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
24874 * fold-const.c (const_binop): Handle vector shifts by a scalar.
24875 (fold_binary_loc): Call const_binop also for mixed vector-scalar
24878 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
24879 Jakub Jelinek <jakub@redhat.com>
24881 * opts.c: Include diagnostic-color.h.
24882 (common_handle_option): Handle OPT_fdiagnostics_color_.
24883 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
24884 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
24885 (diagnostic-color.o): New.
24886 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
24887 (diagnostic_color_rule): New enum.
24888 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
24889 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
24890 the location string.
24891 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
24892 either NULL, or color kind.
24893 * diagnostic-color.c: New file.
24894 * diagnostic-color.h: New file.
24895 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
24897 * doc/invoke.texi (-fdiagnostics-color): Document.
24898 * pretty-print.h (pp_show_color): Define.
24899 (struct pretty_print_info): Add show_color field.
24900 * diagnostic.c: Include diagnostic-color.h.
24901 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
24902 macros. Colorize error:, warning: etc. strings and also the location
24904 (diagnostic_show_locus): Colorize the caret line.
24905 * pretty-print.c: Include diagnostic-color.h.
24906 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
24907 inside of %< %> quotes or quoted through q format modifier.
24909 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24911 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
24913 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24915 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
24916 code in CC_NZ mode.
24917 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
24920 2013-04-11 Marek Polacek <polacek@redhat.com>
24922 PR tree-optimization/48184
24923 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
24925 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
24927 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
24928 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
24929 (skip_simple_arithmetic): Tidy up.
24930 * tree.h (skip_simple_constant_arithmetic): Declare.
24932 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24934 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
24936 2013-04-11 Richard Biener <rguenther@suse.de>
24938 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24939 generate vector constants.
24941 2013-04-11 Richard Biener <rguenther@suse.de>
24943 PR tree-optimization/56878
24944 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
24945 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
24947 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
24948 Prefer to align the DR with the most invariant base address.
24950 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
24952 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
24955 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
24957 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
24958 floating-point vector comparisons against 0.
24960 2013-04-11 Jakub Jelinek <jakub@redhat.com>
24962 PR tree-optimization/56899
24963 * fold-const.c (extract_muldiv_1): Apply distributive law
24964 only if TYPE_OVERFLOW_WRAPS (ctype).
24966 2013-04-11 Bin Cheng <bin.cheng@arm.com>
24969 * ira-costs.c (scan_one_insn): Check whether the source rtx of
24970 loading has side effect.
24972 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
24974 * config/sparc/sparc.c: Include tree-pass.h.
24975 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
24976 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
24977 head of file. Change return type. Split off gate function.
24978 (sparc_gate_work_around_errata): New function.
24979 (pass_work_around_errata): New pass definition.
24980 (insert_pass_work_around_errata) New pass insert definition to
24981 insert pass_work_around_errata just after delayed-branch scheduling.
24982 (sparc_option_override): Insert the pass.
24983 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
24985 2013-04-10 David S. Miller <davem@davemloft.net>
24987 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
24988 or -mcpu=hypersparc.
24990 * target.def (cstore_mode): New hook.
24991 * target.h: Include insn-codes.h
24992 * targhooks.c: Likewise.
24993 (default_cstore_mode): New function.
24994 * targhooks.h: Declare it.
24995 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
24996 * doc/tm.texi: Rebuild.
24997 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
24998 target hook, rather than inspecting the insn_data.
24999 * config/sparc/sparc.c (sparc_cstore_mode): New function.
25000 (TARGET_CSTORE_MODE): Redefine.
25001 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
25003 * config/sparc/predicates.md (cstore_result_operand): New special
25005 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
25006 Use it for operand 0.
25007 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
25008 (*snesi_special): Likewise.
25009 (*snesi_zero): Likewise.
25010 (*seqsi_zero): Likewise.
25011 (*sltu_insn): Likewise.
25012 (*sgeu_insn): Likewise.
25013 (*seqdi_special): Make operand 0 and comparison operation be of
25015 (*snedi_special): Likewise.
25016 (*snedi_special_vis3): Likewise.
25017 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
25018 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
25019 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
25020 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
25021 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
25022 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
25023 (*sltu_extend_sp64): Likewise.
25024 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
25025 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
25026 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
25027 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
25028 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
25030 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
25032 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
25033 (aarch64_start_file): Use the new function.
25035 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
25036 Jason Merrill <jason@redhat.com>
25038 * common.opt: Add -gdwarf.
25039 * opts.c (common_handle_option): Handle it.
25040 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
25042 2013-04-10 Richard Biener <rguenther@suse.de>
25044 * passes.c (execute_todo): Do not call ggc_collect conditional here.
25045 (execute_one_ipa_transform_pass): But unconditionally here.
25046 (execute_one_pass): And here.
25047 (init_optimization_passes): Remove reload pass.
25048 * tree-pass.h (TODO_ggc_collect): Remove.
25049 (pass_reload): Likewise.
25050 * ira.c (do_reload): Merge into ...
25052 (rest_of_handle_reload): Remove.
25053 (pass_reload): Likewise.
25054 * config/i386/i386.c (ix86_option_override): Refer to ira instead
25055 of reload for vzeroupper pass placement.
25056 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
25057 and todo_flags_finish of all passes.
25059 2013-04-10 Richard Biener <rguenther@suse.de>
25061 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
25062 first_const_oprnd field, rename first_def_type to first_op_type.
25063 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
25064 (vect_get_and_check_slp_defs): Always use the type of the
25065 operand. Allow mixed vect_external_def, vect_constant_def types.
25066 (vect_get_constant_vectors): Handle mixed vect_external_def,
25067 vect_constant_def types.
25069 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
25071 PR tree-optimization/55524
25072 * tree-ssa-math-opts.c
25073 (convert_mult_to_fma): Don't use an fms construct
25074 when we don't have an fms operation, but fnma, and it looks
25075 likely that we'll be able to use the latter.
25077 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
25079 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
25081 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
25082 inline fail caused by overwritable functions.
25084 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
25086 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
25087 unnecessary bits in the constant power of two case.
25089 2013-04-10 Richard Biener <rguenther@suse.de>
25091 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
25092 broken code swapping operands.
25093 (vect_build_slp_tree): Do not compute load permutations here.
25094 (vect_analyze_slp_instance): Compute load permutations here,
25095 after building the SLP tree.
25097 2013-04-09 Christian Bruel <christian.bruel@st.com>
25099 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
25100 of next/prev_real_insn.
25102 2013-04-09 Jan Hubicka <jh@suse.cz>
25104 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
25105 Drop aliased parameter.
25106 (function_and_variable_visibility): Do not handle alias pairs.
25107 * cgraph.c (varpool_externally_visible_p): Update prototype.
25108 * varpool.c (varpool_add_new_variable): Update.
25110 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25112 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
25114 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
25116 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
25118 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
25120 2013-04-09 Marek Polacek <polacek@redhat.com>
25122 PR tree-optimization/48762
25123 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
25125 2013-04-09 Richard Biener <rguenther@suse.de>
25127 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
25129 (vect_build_slp_tree): Likewise.
25130 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
25131 calculating the cost of a SLP instance.
25132 (vect_analyze_slp_instance): Use it from here, after building
25135 2013-04-09 Jakub Jelinek <jakub@redhat.com>
25137 PR middle-end/56883
25138 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
25139 expand_omp_for_static_chunk): Use simple_p = true in
25140 force_gimple_operand_gsi calls when assigning to addressable decls.
25142 2013-04-09 Jeff Law <law@redhat.com>
25144 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
25145 when the boolean was created by converting a wider object which
25146 had a boolean range.
25148 2013-04-09 Richard Biener <rguenther@suse.de>
25150 * tree-vectorizer.h (slp_void_p): Remove.
25151 (slp_tree): Typedef before _slp_tree declaration.
25152 (struct _slp_tree): Use a vector of slp_tree as children.
25153 (vect_get_place_in_interleaving_chain): Remove.
25154 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
25156 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
25158 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
25159 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
25160 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
25161 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
25162 Use slp_node instead of slp_void_p and adjust.
25164 2013-04-09 Richard Biener <rguenther@suse.de>
25166 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
25167 work that is not necessary.
25169 2013-04-09 Jakub Jelinek <jakub@redhat.com>
25171 PR tree-optimization/56854
25172 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
25173 forward into clobber stmts if it would change MEM_REF lhs into
25176 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
25178 * tree.c (type_hash_lookup, type_hash_add): Make static.
25179 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
25181 2013-04-09 Richard Biener <rguenther@suse.de>
25183 * tree.h (unsave_expr_now): Remove.
25184 * tree-inline.c (mark_local_for_remap_r): Remove.
25185 (unsave_expr_1): Likewise.
25186 (unsave_r): Likewise.
25187 (unsave_expr_now): Likewise.
25188 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
25189 (propagate_tree_value): Likewise.
25191 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
25193 * doc/rtl.texi (sequence): Rewrite documentation to match the
25194 current use of SEQUENCE rtl objects.
25195 * rtl.def (SEQUENCE): Likewise.
25197 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
25198 Update documentation.
25199 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
25200 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
25202 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
25204 2013-04-08 Teresa Johnson <tejohnson@google.com>
25206 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
25207 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
25209 (estimate_edge_size_and_time): Add comment to suggest using rounding
25211 (estimate_node_size_and_time): Ditto.
25212 (remap_edge_change_prob): Use helper rounding divide methods.
25213 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
25214 (gimple_mod_pow2_value_transform): Ditto.
25215 (gimple_mod_subtract_transform): Ditto.
25216 (gimple_ic_transform): Ditto.
25217 (gimple_stringops_transform): Ditto.
25218 * stmt.c (conditional_probability): Ditto.
25219 (emit_case_dispatch_table): Ditto.
25220 * lto-cgraph.c (merge_profile_summaries): Ditto.
25221 * tree-optimize.c (execute_fixup_cfg): Ditto.
25222 * cfgcleanup.c (try_forward_edges): Ditto.
25223 * cfgloopmanip.c (scale_loop_profile): Ditto.
25225 (duplicate_loop_to_header_edge): Ditto.
25226 (lv_adjust_loop_entry_edge): Ditto.
25227 * tree-vect-loop.c (vect_transform_loop): Ditto.
25228 * profile.c (compute_branch_probabilities): Ditto.
25229 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
25230 * lto-streamer-in.c (input_cfg): Ditto.
25231 * gimple-streamer-in.c (input_bb): Ditto.
25232 * ipa-cp.c (update_profiling_info): Ditto.
25233 (update_specialized_profile): Ditto.
25234 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
25235 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
25237 * sched-rgn.c (compute_dom_prob_ps): Ditto.
25238 (compute_trg_info): Ditto.
25239 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
25240 (purge_dead_edges): Ditto.
25241 * loop-unswitch.c (unswitch_loop): Ditto.
25242 * cgraphclones.c (cgraph_clone_edge): Ditto.
25243 (cgraph_clone_node): Ditto.
25244 * tree-inline.c (copy_bb): Ditto.
25245 (copy_edges_for_bb): Ditto.
25246 (initialize_cfun): Ditto.
25247 (copy_cfg_body): Ditto.
25248 (expand_call_inline): Ditto.
25250 2013-04-08 Kai Tietz <ktietz@redhat.com>
25252 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
25253 TARGET_CYGWIN64 by TARGET_64BIT.
25255 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
25257 * config/epiphany/epiphany.md (GPR_1): New constant.
25258 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
25259 * config/epiphany/epiphany.c (gen_compare_reg):
25260 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
25261 is already in place.
25262 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
25263 Don't require being called during rtl expansion; If y operlaps r0,
25265 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
25266 (epiphany_expand_epilogue): Likewise.
25268 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
25269 Don't use CC_FPmode for ORDERED / UNORDERED.
25270 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
25272 * config/epiphany/constraints.md (CnL): New constraint.
25273 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
25274 * config/epiphany/predicates.md (add_operand): Allow 1024.
25276 * config/epiphany/epiphany.md (logical_op): New code iterator.
25277 (op_mnc): New code attribute.
25278 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
25279 (mov_f+1, mov_f+2): New peephole2 patterns.
25281 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
25282 (cstoresi4): Also allow re-use of zero result when doing a NE
25283 comparison to a non-zero operand.
25284 Use (clobber (scratch)) for first insn if the gpr output is not needed.
25286 * config/epiphany/epiphany.md (<insn_opname>v2si3):
25287 Use gen_addsi3_i / gen_subsi3_i.
25289 2013-04-08 Jakub Jelinek <jakub@redhat.com>
25293 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
25294 contain anything but clobbers, at most one __builtin_stack_restore,
25295 optionally debug stmts and final resx, and if it has at least one
25296 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
25297 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
25298 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
25299 which isn't defaut definition, remove them.
25300 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
25301 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
25302 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
25303 with MEM_REF LHS with SSA_NAME address.
25305 2013-04-08 Jeff Law <law@redhat.com>
25307 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
25309 2013-04-08 Richard Biener <rguenther@suse.de>
25311 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
25313 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
25314 determined vector type.
25315 (vect_analyze_data_refs): Likewise.
25316 (vect_get_new_vect_var): Adjust.
25317 (vect_create_destination_var): Preserve SSA name versions.
25318 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
25319 not dump anything here.
25321 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
25323 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
25324 Add member lr_slot_known.
25325 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
25327 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
25328 Remove code that sets lr_slot_offset according to what a previous
25329 version of epiphany_emit_save_restore used to do.
25330 (epiphany_emit_save_restore): When doing an lr save or restore,
25331 set/verify lr_slot_known and lr_slot_offset.
25333 2013-04-08 Xinyu Qi <xyqi@marvell.com>
25336 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
25339 2013-04-08 Richard Biener <rguenther@suse.de>
25341 * alias.c (find_base_term): Fix thinko in previous change.
25343 2013-04-08 Jakub Jelinek <jakub@redhat.com>
25345 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
25346 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
25347 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
25348 if possible to compute val.
25349 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
25350 For QImode integers don't require anything about precision. Use
25351 const_with_all_bytes_same to find out if the constant doesn't have
25352 repeated bytes in it.
25354 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25356 * config/s390/s390.c (s390_expand_insv): Only accept insertions
25359 2013-04-08 Marek Polacek <polacek@redhat.com>
25361 PR rtl-optimization/48182
25362 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
25365 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25368 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
25369 nuses, make sure we have a label.
25371 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25374 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
25375 (rs6000_emit_swdiv_low_precision): Remove.
25376 (rs6000_emit_swdiv): Rewrite to handle between one and four
25377 iterations of Newton-Raphson generally; modify required number of
25378 iterations for some cases.
25379 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
25381 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
25383 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
25384 set-but-unused variable.
25386 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
25387 basic blocks of released function bodies garbage-collectable.
25389 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
25390 (struct rtl_opt_pass): Add TODO_df_finish.
25392 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
25394 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25396 * config/arm/constraints.md (q): New constraint.
25397 * config/arm/ldrdstrd.md: New file.
25398 * config/arm/arm.md (ldrdstrd.md) New include.
25399 (arm_movdi): Use "q" instead of "r" constraint
25400 for double-word memory access.
25401 (movdf_soft_insn): Likewise.
25402 * config/arm/vfp.md (movdi_vfp): Likewise.
25403 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
25404 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
25405 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
25406 (mem_ok_for_ldrd_strd): Likewise.
25407 (output_move_double): Update assertion.
25409 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25411 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
25413 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25415 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
25416 define_insn_and_split.
25417 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
25419 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25421 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
25422 define_insn_and_split.
25423 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
25424 (shiftsi3_compare): New pattern.
25425 (rrx): New pattern.
25426 * config/arm/unspecs.md (UNSPEC_RRX): New.
25428 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25430 * config/arm/arm.md (negdi_extendsidi): New pattern.
25431 (negdi_zero_extendsidi): Likewise.
25433 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25435 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
25436 define_insn_and_split.
25437 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
25438 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
25440 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25442 * config/arm/arm.md (arm_subdi3): Convert define_insn into
25443 define_insn_and_split.
25444 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
25445 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
25447 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25449 * config/arm/arm.md (subsi3_carryin): New pattern.
25450 (subsi3_carryin_const): Likewise.
25451 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
25452 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
25454 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25456 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
25458 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
25460 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
25461 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
25463 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25465 * config/arm/arm.c (arm_expand_builtin): Change fcode
25466 type to unsigned int.
25468 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25470 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
25472 2013-04-04 Ian Lance Taylor <iant@google.com>
25474 * doc/standards.texi (Standards): The Go frontend supports the Go 1
25477 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
25479 PR middle-end/56729
25480 * df-scan.c (df_insn_delete): Disable failing assert.
25482 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25484 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
25485 New function prototype.
25486 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
25487 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
25488 (arm_builtin_vectorized_function): New function.
25490 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25492 * config/arm/arm_neon_builtins.def: New file.
25493 * config/arm/arm.c (neon_builtin_data): Move contents to
25494 arm_neon_builtins.def.
25495 (enum arm_builtins): Include neon builtin definitions.
25496 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
25497 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
25499 2013-04-04 Marek Polacek <polacek@redhat.com>
25501 PR tree-optimization/48186
25502 * predict.c (maybe_hot_frequency_p): Return false if
25503 HOT_BB_FREQUENCY_FRACTION is 0.
25504 (cgraph_maybe_hot_edge_p): Likewise.
25506 2013-04-04 Richard Biener <rguenther@suse.de>
25508 PR tree-optimization/56826
25509 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
25512 2013-04-04 Richard Biener <rguenther@suse.de>
25514 PR tree-optimization/56213
25515 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
25516 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
25518 2013-04-04 Richard Biener <rguenther@suse.de>
25520 PR tree-optimization/56837
25521 * tree-loop-distribution.c (classify_partition): For non-zero
25522 values require that the value has the same precision as its
25523 mode to be useful as memset value.
25525 2013-04-03 Nick Clifton <nickc@redhat.com>
25527 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
25528 (fmssf4): Use fmsf.s on E3V5 architectures.
25529 (fnmasf4): Use fnmaf.s on E3V5 architectures.
25530 (fnmssf4): Use fnmsf.s on E3V5 architectures.
25532 2013-04-03 Jeff Law <law@redhat.com>
25534 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
25535 (lra-eliminations.o): Likewise.
25537 2013-04-03 Teresa Johnson <tejohnson@google.com>
25539 * gcov-io.c (compute_working_sets): Moved most of body of old
25540 compute_working_sets here from profile.c.
25541 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
25542 (gcov_working_set_t): Moved typedef here from basic-block.h
25543 (compute_working_set): Declare.
25544 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
25545 (get_working_sets): Renamed from compute_working_set,
25546 replace most of body with call to new compute_working_sets.
25547 (get_exec_counts): Replace call to compute_working_sets
25548 to get_working_sets.
25549 * profile.h (get_working_sets): Renamed from compute_working_set.
25550 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
25551 to get_working_sets.
25552 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
25553 * gcov-dump.c (dump_working_sets): New function.
25555 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
25557 * hwint.c (sext_hwi, zext_hwi): New functions.
25558 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
25559 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
25560 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
25561 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
25562 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
25563 (sext_hwi, zext_hwi): New functions.
25565 2013-04-03 Jeff Law <law@redhat.com>
25567 PR tree-optimization/56799
25568 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
25569 back test for widening conversion erroneously dropped in prior change.
25571 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25574 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
25575 instead of next_real_insn.
25577 2013-04-03 Marek Polacek <polacek@redhat.com>
25580 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
25582 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25585 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
25587 (thumb1_output_casesi): Likewise.
25588 (thumb2_output_casesi): Likewise.
25590 2013-04-03 Richard Biener <rguenther@suse.de>
25592 PR tree-optimization/56817
25593 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
25595 (tree_unroll_loops_completely_1): ... new function to manually
25596 walk the loop tree, properly defering outer loops of unrolled
25597 loops to later iterations.
25599 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
25601 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
25602 (vectorizable_load): Likewise.
25603 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
25604 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
25606 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
25608 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
25611 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
25613 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
25615 2013-04-03 Bin Cheng <bin.cheng@arm.com>
25617 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
25619 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
25621 PR tree-optimization/56790
25622 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
25625 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
25627 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
25629 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
25630 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
25633 2013-04-03 Jakub Jelinek <jakub@redhat.com>
25636 * tree.h (force_folding_builtin_constant_p): New decl.
25637 * builtins.c (force_folding_builtin_constant_p): New variable.
25638 (fold_builtin_constant_p): Fold immediately also if
25639 force_folding_builtin_constant_p.
25641 2013-04-03 Richard Biener <rguenther@suse.de>
25643 PR tree-optimization/56812
25644 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
25645 DRs of the same interleaving chain are independent.
25647 2013-04-02 Jason Merrill <jason@redhat.com>
25649 * gdbinit.in (pbb): Use debug fn.
25651 2013-04-02 Lawrence Crowl <crowl@google.com>
25653 * sese.h (struct ivtype_map_elt_s): Remove unused.
25654 (extern debug_ivtype_map): Remove unused.
25655 (extern eq_ivtype_map_elts): Remove unused.
25656 * sese.c (debug_ivtype_map): Removed unused.
25657 (debug_ivtype_map_1): Removed unused.
25658 (debug_ivtype_elt): Remove unused.
25659 (eq_ivtype_map_elts): Remove unused.
25661 2013-04-02 Kai Tietz <ktietz@redhat.com>
25664 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
25665 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
25666 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
25668 (legitimize_pe_coff_symbol): Likewise.
25669 (is_imported_p): New helper-function.
25670 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
25671 for Windows x64 targets.
25672 (ix86_expand_prologue): Optimize for pe-coff targets.
25673 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
25674 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
25675 medium/large code-model.
25676 (legitimize_pic_address): Likewise.
25677 (legitimize_tls_address): Likewise.
25678 (ix86_expand_call): Likewise.
25679 (x86_output_mi_thunk): Likewise.
25680 (get_dllimport_decl): Add new beimport argument.
25681 (construct_plt_address): Don't assert for x64 pe-coff targets.
25682 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
25684 (SYMBOL_FLAG_STUBVAR): New macro.
25685 (SYMBOL_REF_STUBVAR_P): Likewise.
25686 * config/i386/winnt.c (stub_list): New structure.
25687 (stub_head): New local variable.
25688 (i386_pe_record_stub): New function.
25689 (i386_pe_file_end): Emit refptr-stubs.
25691 2013-04-02 Jakub Jelinek <jakub@redhat.com>
25693 PR rtl-optimization/56745
25694 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
25695 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
25698 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
25699 and both of them are MEM_REFs, just compare first argument for
25700 equality and attempt to deal even with differing offsets.
25703 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
25704 of gimple_clobber_p to be MEM_REF.
25705 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
25706 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
25707 after gimplification.
25708 * asan.c (get_mem_ref_of_assignment): Don't instrument
25709 gimple_clobber_p stmts.
25710 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
25711 gimple_clobber_p stmt if they have MEM_REF lhs and
25712 are dead because of another gimple_clobber_p stmt.
25713 * tree-ssa-live.c (clear_unused_block_pointer): Treat
25714 gimple_clobber_p stmts like debug stmts.
25715 (remove_unused_locals): Remove clobbers with MEM_REF lhs
25716 that refer to unused VAR_DECLs or uninitialized values.
25717 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
25718 gimple_clobber_p stmts if they refer to removed parameters.
25719 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
25722 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
25724 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
25725 using SWI48 mode attribute.
25727 2013-04-02 Wei Mi <wmi@google.com>
25729 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
25730 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
25731 *<rotate_insn><mode>3_mask in i386.md.
25733 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
25735 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
25737 2013-04-02 Richard Biener <rguenther@suse.de>
25739 PR tree-optimization/56778
25740 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
25741 Runtime alias tests are not supported for gather loads.
25742 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
25743 stmts referenced from SSA operands before updating SSA form.
25745 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
25746 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25748 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
25749 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
25750 * config/arm/cortex-a53.md: New file.
25751 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
25752 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
25753 * config/arm/arm.c (arm_issue_rate): Likewise.
25754 * config/arm/arm-tune.md: Regenerate
25755 * config/arm/arm-tables.opt: Regenerate.
25756 * config/arm/arm-cores.def: Add cortex-a53.
25758 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
25760 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
25763 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
25765 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
25766 scalar load/store operations using B/H registers.
25767 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
25769 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
25771 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
25773 * config/aarch64/aarch64.c
25774 (aarch64_simd_scalar_immediate_valid_for_move): New.
25775 * config/aarch64/aarch64-protos.h
25776 (aarch64_simd_scalar_immediate_valid_for_move): New.
25777 * config/aarch64/constraints.md (Dh, Dq): New.
25778 * config/aarch64/iterators.md (hq): New.
25780 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
25782 * reorg.c (get_branch_condition): Deal with conditional returns.
25783 (fill_simple_delay_slots): Remove dead code dealing with jumps.
25785 2013-04-01 Wei Mi <wmi@google.com>
25787 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
25788 Truncate operand 2 using %b asm operand modifier.
25789 (*<shift_insn><mode>3_mask): Ditto.
25790 (*<rotate_insn><mode>3_mask): Ditto.
25792 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
25794 PR middle-end/56798
25795 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
25797 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
25799 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
25801 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
25803 2013-03-30 Lawrence Crowl <crowl@google.com>
25805 * dse.c (clear_alias_sets): Remove never set.
25806 (disqualified_clear_alias_sets): Remove never set.
25807 (clear_alias_mode_pool): Remove never set.
25808 (dse_step0): Remove condition that is never true.
25809 (canon_address): Remove condition that is never true.
25810 (dse_step7): Remove condition that is never true.
25811 (rest_of_handle_dse): Remove condition that is never true.
25812 (rest_of_handle_dse::did_global): Remove never read from above.
25813 (dse_step2_spill): Remove never called from above.
25814 (dse_step5_spill): Remove never called from above.
25816 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
25818 * doc/md.texi (Standard Names) <casesi>: Update documentation for
25819 JUMP_TABLE_DATA changes.
25820 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
25821 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
25822 (Insns) <jump_table_data>: New entry.
25823 * doc/tm.texi: Regenerate.
25825 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
25827 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
25828 for table jump at the end of a basic block using tablejump_p.
25829 * targhooks.c (default_invalid_within_doloop): Likewise.
25830 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
25831 target hook implementation that is identical to the default hook.
25832 (rs6000_invalid_within_doloop): Remove.
25834 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
25835 unused variable from tablejump_p call.
25837 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
25838 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
25839 (INSN_DELETED_P): Likewise.
25840 (emit_jump_table_data): New prototype.
25841 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
25842 after 4th as unused.
25843 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
25844 * sched-vis.c (print_insn): Likewise.
25845 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
25846 insn for compatibility with back ends that use next_active_insn to
25847 identify jump table data.
25848 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
25849 (remove_insn): Likewise.
25850 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
25852 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
25853 (emit_jump_table_data): New function.
25855 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
25856 basic block, a JUMP_TABLE_DATA never is.
25857 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
25858 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
25859 off from code handling real insns.
25860 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
25861 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
25862 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
25863 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
25864 is not a NONDEBUG_INSN_P.
25865 * ira-costs.c (scan_one_insn): Likewise.
25866 * jump.c (mark_all_labels): Likewise.
25867 (mark_jump_label_1): Likewise.
25868 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
25869 * lra.c (get_insn_freq): Expect all insns reaching here to be in
25871 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
25872 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
25873 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
25874 JUMP_TABLE_DATA_P insns.
25875 (calculate_elim_costs_all_insns): Likewise.
25876 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
25877 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
25878 (delete_output_reload): Code style fixups.
25879 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
25880 insn flags on this non-insn.
25881 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
25882 as scheduling barriers, for pre-change compatibility.
25883 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
25884 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
25886 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
25887 redundant JUMP_TABLE_DATA_P test.
25888 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
25889 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
25890 (frv_for_each_packet): Likewise.
25891 * config/i386/i386.c (min_insn_size): Likewise.
25892 (ix86_avoid_jump_mispredicts): Likewise.
25893 * config/m32r/m32r.c (m32r_is_insn): Likewise.
25894 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
25895 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
25896 (mips16_insn_length): Robustify.
25897 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
25898 (mips16_split_long_branches): Likewise.
25899 * config/pa/pa.c (pa_combine_instructions): Likewise.
25900 * config/rs6000/rs6000.c (get_next_active_insn): Treat
25901 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
25902 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
25903 as contributing to pool range lengths.
25904 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
25905 Remove redundant JUMP_TABLE_DATA_P test.
25906 (sh_loop_align): Likewise.
25907 (split_branches): Likewise.
25908 (sh_insn_length_adjustment): Likewise.
25909 * config/spu/spu.c (get_branch_target): Likewise.
25911 2013-03-29 Jan Hubicka <jh@suse.cz>
25913 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
25914 gcov streaming; stream hot bb threshold to ltrans.
25915 * predict.c (get_hot_bb_threshold): Break out from ....
25916 (maybe_hot_count_p): ... here.
25917 (set_hot_bb_threshold): New function.
25918 * lto-section-in.c (lto_section_name): Add profile.
25919 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
25920 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
25921 and data-streamer.h
25922 (histogram_entry): New structure.
25923 (histogram, histogram_pool): New global vars.
25924 (histogram_hash): New structure.
25925 (histogram_hash::hash): New method.
25926 (histogram_hash::equal): Likewise.
25927 (account_time_size): New function.
25928 (cmp_counts): New function.
25929 (dump_histogram): New function.
25930 (ipa_profile_generate_summary): New function.
25931 (ipa_profile_write_summary): New function.
25932 (ipa_profile_read_summary): New function.
25933 (ipa_profile): Decide on threshold.
25934 (pass_ipa_profile): Add ipa_profile_write_summary and
25935 ipa_profile_read_summary.
25936 * Makefile.in (ipa.o): Update dependencies.
25937 * lto-streamer.h (LTO_section_ipa_profile): New section.
25939 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
25941 * tree.h (VAR_P): New.
25943 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
25946 * doc/invoke.texi ([-fwhole-program]): Fix typo.
25948 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
25950 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
25951 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
25952 (control_flow_insn_p): Likewise.
25953 * cfgrtl.c (duplicate_insn_chain): Likewise.
25954 * final.c (get_attr_length_1): Likewise.
25955 (shorten_branches): Likewise.
25956 (final_scan_insn): Likewise.
25957 * function.c (instantiate_virtual_regs): Likewise.
25958 * gcse.c (insert_insn_end_basic_block): Likewise.
25959 * ira-costs.c (scan_one_insn): Likewise.
25960 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
25961 * lra.c (check_rtl): Likewise.
25962 * reload1.c (elimination_costs_in_insn): Likewise.
25963 * reorg.c (follow_jumps): Likewise.
25965 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
25966 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
25967 (thumb_far_jump_used_p): Likewise.
25968 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
25969 (workaround_speculation): Likewise.
25970 (add_sched_insns_for_speculation): Likewise.
25971 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
25972 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
25973 (frv_for_each_packet): Likewise.
25974 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
25975 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
25976 (final_emit_insn_group_barriers): Likewise.
25977 * config/m32r/m32r.c (m32r_is_insn): Likewise.
25978 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
25979 (mips16_insn_length): Likewise.
25980 * config/pa/pa.c (pa_reorg): Likewise.
25981 (pa_combine_instructions): Likewise.
25982 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
25983 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
25984 (sh_reorg): Likewise.
25985 (split_branches): Likewise.
25986 * config/spu/spu.c (get_branch_target): Likewise.
25988 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
25991 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
25993 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
25994 Fix declaration name.
25996 2013-03-28 Lawrence Crowl <crowl@google.com>
25998 * graphds.h (struct graph.indicies): Remove unused.
25999 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
26000 (SCOP_ORIGINAL_PDDRS): Remove unused.
26001 * sese.h (extern insert_loop_close_phis): Removed unused.
26002 (extern insert_guard_phis): Removed unused.
26003 (extern ivtype_map_elt_info): Removed unused.
26004 (new_ivtype_map_elt): Removed unused.
26005 * sese.c (ivtype_map_elt_info): Removed unused.
26007 2013-03-28 Lawrence Crowl <crowl@google.com>
26009 * Makefile.in: Add several missing include dependences.
26011 (test-dump.o): New. This object is not added to any executable,
26012 but is present for ad-hoc testing.
26014 (debug (const bitmap_head_def &)): New.
26015 (debug (const bitmap_head_def *)): New.
26017 (extern debug (const bitmap_head_def &)): New.
26018 (extern debug (const bitmap_head_def *)): New.
26020 (debug (edge_def &)): New.
26021 (debug (edge_def *)): New.
26023 (debug (basic_block_def &)): New.
26024 (debug (basic_block_def *)): New.
26026 (dump_node (const_tree, int, FILE *)): Correct source file.
26028 (debug (die_struct &)): New.
26029 (debug (die_struct *)): New.
26031 (extern debug (die_struct &)): New.
26032 (extern debug (die_struct *)): New.
26033 * gimple-pretty-print.c
26034 (debug (gimple_statement_d &)): New.
26035 (debug (gimple_statement_d *)): New.
26036 * gimple-pretty-print.h
26037 (extern debug (gimple_statement_d &)): New.
26038 (extern debug (gimple_statement_d *)): New.
26040 (debug (ira_allocno_copy &)): New.
26041 (debug (ira_allocno_copy *)): New.
26042 (debug (ira_allocno &)): New.
26043 (debug (ira_allocno *)): New.
26045 (extern debug (ira_allocno_copy &)): New.
26046 (extern debug (ira_allocno_copy *)): New.
26047 (extern debug (ira_allocno &)): New.
26048 (extern debug (ira_allocno *)): New.
26050 (debug (live_range &)): New.
26051 (debug (live_range *)): New.
26053 (debug (lra_live_range &)): New.
26054 (debug (lra_live_range *)): New.
26056 (debug (lra_live_range &)): New.
26057 (debug (lra_live_range *)): New.
26059 (debug (omega_pb_d &)): New.
26060 (debug (omega_pb_d *)): New.
26062 (extern debug (omega_pb_d &)): New.
26063 (extern debug (omega_pb_d *)): New.
26065 (debug (const rtx_def &)): New.
26066 (debug (const rtx_def *)): New.
26068 (debug_tree (tree): Move within file.
26069 (debug_raw (const tree_node &)): New.
26070 (debug_raw (const tree_node *)): New.
26071 (dump_tree_via_hooks (const tree_node *, int)): New.
26072 (debug (const tree_node &)): New.
26073 (debug (const tree_node *)): New.
26074 (debug_verbose (const tree_node &)): New.
26075 (debug_verbose (const tree_node *)): New.
26076 (debug_head (const tree_node &)): New.
26077 (debug_head (const tree_node *)): New.
26078 (debug_body (const tree_node &)): New.
26079 (debug_body (const tree_node *)): New.
26080 (debug_vec_tree (tree): Move and reimplement in terms of dump.
26081 (debug (vec<tree, va_gc> &)): New.
26082 (debug (vec<tree, va_gc> *)): New.
26084 (extern debug (const rtx_def &)): New.
26085 (extern debug (const rtx_def *)): New.
26087 (debug_raw (simple_bitmap_def &)): New.
26088 (debug_raw (simple_bitmap_def *)): New.
26089 (debug (simple_bitmap_def &)): New.
26090 (debug (simple_bitmap_def *)): New.
26092 (extern debug (simple_bitmap_def &)): New.
26093 (extern debug (simple_bitmap_def *)): New.
26094 (extern debug_raw (simple_bitmap_def &)): New.
26095 (extern debug_raw (simple_bitmap_def *)): New.
26097 (debug (vinsn_def &)): New.
26098 (debug (vinsn_def *)): New.
26099 (debug_verbose (vinsn_def &)): New.
26100 (debug_verbose (vinsn_def *)): New.
26101 (debug (expr_def &)): New.
26102 (debug (expr_def *)): New.
26103 (debug_verbose (expr_def &)): New.
26104 (debug_verbose (expr_def *)): New.
26105 (debug (vec<rtx> &)): New.
26106 (debug (vec<rtx> *)): New.
26108 (extern debug (vinsn_def &)): New.
26109 (extern debug (vinsn_def *)): New.
26110 (extern debug_verbose (vinsn_def &)): New.
26111 (extern debug_verbose (vinsn_def *)): New.
26112 (extern debug (expr_def &)): New.
26113 (extern debug (expr_def *)): New.
26114 (extern debug_verbose (expr_def &)): New.
26115 (extern debug_verbose (expr_def *)): New.
26116 (extern debug (vec<rtx> &)): New.
26117 (extern debug (vec<rtx> *)): New.
26119 (_list_iter_cond_expr): Make inline instead of static.
26121 (debug (sreal &)): New.
26122 (debug (sreal *)): New.
26124 (extern debug (sreal &)): New.
26125 (extern debug (sreal *)): New.
26127 (extern debug_raw (const tree_node &)): New.
26128 (extern debug_raw (const tree_node *)): New.
26129 (extern debug (const tree_node &)): New.
26130 (extern debug (const tree_node *)): New.
26131 (extern debug_verbose (const tree_node &)): New.
26132 (extern debug_verbose (const tree_node *)): New.
26133 (extern debug_head (const tree_node &)): New.
26134 (extern debug_head (const tree_node *)): New.
26135 (extern debug_body (const tree_node &)): New.
26136 (extern debug_body (const tree_node *)): New.
26137 (extern debug (vec<tree, va_gc> &)): New.
26138 (extern debug (vec<tree, va_gc> *)): New.
26140 (debug (struct loop &)): New.
26141 (debug (struct loop *)): New.
26142 (debug_verbose (struct loop &)): New.
26143 (debug_verbose (struct loop *)): New.
26144 * tree-dump.c: Add header dependence.
26146 (extern debug (struct loop &)): New.
26147 (extern debug (struct loop *)): New.
26148 (extern debug_verbose (struct loop &)): New.
26149 (extern debug_verbose (struct loop *)): New.
26151 (debug (data_reference &)): New.
26152 (debug (data_reference *)): New.
26153 (debug (vec<data_reference_p> &)): New.
26154 (debug (vec<data_reference_p> *)): New.
26155 (debug (vec<ddr_p> &)): New.
26156 (debug (vec<ddr_p> *)): New.
26158 (extern debug (data_reference &)): New.
26159 (extern debug (data_reference *)): New.
26160 (extern debug (vec<data_reference_p> &)): New.
26161 (extern debug (vec<data_reference_p> *)): New.
26162 (extern debug (vec<ddr_p> &)): New.
26163 (extern debug (vec<ddr_p> *)): New.
26165 (debug (pt_solution &)): New.
26166 (debug (pt_solution *)): New.
26168 (extern debug (pt_solution &)): New.
26169 (extern debug (pt_solution *)): New.
26171 (debug (_var_map &)): New.
26172 (debug (_var_map *)): New.
26173 (debug (tree_live_info_d &)): New.
26174 (debug (tree_live_info_d *)): New.
26176 (extern debug (_var_map &)): New.
26177 (extern debug (_var_map *)): New.
26178 (extern debug (tree_live_info_d &)): New.
26179 (extern debug (tree_live_info_d *)): New.
26181 2013-03-28 Jan Hubicka <jh@suse.cz>
26183 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
26185 2013-03-28 Ian Bolton <ian.bolton@arm.com>
26187 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
26188 record only when desired or required.
26190 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
26192 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
26193 *vec_extractv2di_1_rex64. Use x64 isa attribute.
26195 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26197 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
26198 (*andsi3_compare0_uxtw): New pattern.
26199 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
26200 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
26202 2013-03-28 Jan Hubicka <jh@suse.cz>
26204 * data-streamer-in.c (streamer_read_gcov_count): New function.
26205 * gimple-streamer-out.c: Include value-prof.h.
26206 (output_gimple_stmt): Output histogram.
26207 (output_bb): Use streamer_write_gcov_count.
26208 * value-prof.c: Include data-streamer.h
26209 (dump_histogram_value): Add HIST_TYPE_MAX.
26210 (stream_out_histogram_value): New function.
26211 (stream_in_histogram_value): New function.
26212 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
26213 (stream_out_histogram_value, stream_in_histogram_value): Declare.
26214 * data-streamer-out.c (streamer_write_gcov_count): New function.
26215 (streamer_write_gcov_count_stream): New function.
26216 * lto-cgraph.c (lto_output_edge): Update counter streaming.
26217 (lto_output_node): Likewise.
26218 (input_node, input_edge): Likewise.
26219 * lto-streamer-out.c (output_cfg): Update streaming.
26220 * lto-streamer-in.c (input_cfg): Likewise.
26221 * data-streamer.h (streamer_write_gcov_count,
26222 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
26223 * gimple-streamer-in.c: Include value-prof.h
26224 (input_gimple_stmt): Input histograms.
26225 (input_bb): Update profile streaming.
26227 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
26229 * genmodes.c (emit_max_int): New function.
26230 (emit_insn_modes_h): Added call to emit_max_function.
26231 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
26233 * machmode.def: Fixed comment.
26235 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
26237 * combine.c (try_combine): Removed useless assert.
26238 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
26240 2013-03-28 Marek Polacek <polacek@redhat.com>
26241 Richard Biener <rguenther@suse.de>
26243 PR tree-optimization/56695
26244 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
26245 build signed result of a vector comparison.
26246 * tree-cfg.c (verify_gimple_comparison): Check that a result
26247 of a vector comparison has signed type.
26249 2013-03-28 Richard Biener <rguenther@suse.de>
26251 PR tree-optimization/37021
26252 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
26253 do not restrict gaps between groups.
26254 * tree-vect-stmts.c (vectorizable_load): Properly account for
26255 a gap between groups.
26257 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
26259 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
26260 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
26263 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
26265 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
26266 * doc/extend.texi (Named Address Spaces): Ditto.
26267 (Variable Attributes): Ditto.
26269 2013-03-27 Kai Tietz <ktietz@redhat.com>
26271 * config.build: Add support for cygwin x64 target.
26272 * config.gcc: Likewise.
26273 * config.host: Likewise.
26274 * configure.ac: Likewise
26275 * configure: Regenerated.
26277 2013-03-27 Kai Tietz <ktietz@redhat.com>
26279 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
26280 * config/i386/t-cygwin-w64: New file.
26281 * config/i386/cygwin-w64.h: New file.
26282 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
26283 and add support for x64-cygwin target.
26284 (CPP_SPEC): Likewise.
26285 (CXX_WRAP_SPEC_LIST): Undefine before define.
26286 (LIBGCJ_SONAME): Use 15 as version.
26288 2013-03-27 Richard Biener <rguenther@suse.de>
26290 PR tree-optimization/56716
26291 * tree-ssa-structalias.c (perform_var_substitution): Adjust
26292 dumping for ref nodes.
26294 2013-03-27 Martin Jambor <mjambor@suse.cz>
26296 PR tree-optimization/55334
26297 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
26298 restricted pointers to arrays.
26300 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
26302 * Makefile.in (.SUFFIXES): Add .cc.
26303 (.c.o): Apply same recipe for implicit rule .cc.o.
26305 2013-03-27 Richard Biener <rguenther@suse.de>
26307 PR tree-optimization/37021
26308 * tree-vect-data-refs.c (vect_check_strided_load): Allow
26309 REALPART/IMAGPART_EXPRs around the supported refs.
26310 * tree-ssa-structalias.c (find_func_aliases): Assume that
26311 floating-point values are not used to transfer pointers.
26313 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
26315 * target.def (TARGET_HAS_IFUNC_P): New target hook.
26316 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
26317 * doc/tm.texi: Regenerate.
26318 * targhooks.h (default_has_ifunc_p): New.
26319 * targhooks.c (default_has_ifunc_p): Ditto.
26320 * config/linux-protos.h: New file.
26321 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
26322 hook for linux which disables support of indirect functions in android.
26323 * config/linux-android.c: New file.
26324 * config/t-linux-android.c: Ditto.
26325 * config.gcc: Added new object file linux-android.o.
26326 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
26327 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
26328 * varasm.c (do_assemble_alias): Likewise.
26329 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
26330 doesn't support indirect functions.
26331 * configure: Regenerate.
26333 2013-03-27 Bin Cheng <bin.cheng@arm.com>
26336 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
26337 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
26340 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26342 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
26344 2013-03-27 Terry Guo <terry.guo@arm.com>
26346 * config/arm/arm-cores.def: Added core cortex-r7.
26347 * config/arm/arm-tune.md: Regenerated.
26348 * config/arm/arm-tables.opt: Regenerated.
26349 * doc/invoke.texi: Added entry for core cortex-r7.
26351 2013-03-27 Walter Lee <walt@tilera.com>
26353 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
26354 double-decrement of next_scratch_regno.
26356 2013-03-27 Walter Lee <walt@tilera.com>
26358 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
26360 (insn_v1mulus): Ditto.
26361 (insn_v2muls): Ditto.
26363 2013-03-27 Walter Lee <walt@tilera.com>
26365 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
26366 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
26368 2013-03-27 Walter Lee <walt@tilera.com>
26370 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
26371 (*sibcall_value): Ditto.
26373 2013-03-27 Walter Lee <walt@tilera.com>
26375 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
26376 (insn_mnz_v8qi): ... this ...
26377 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
26379 (insn_v<n>mnz): Replaced by ...
26380 (insn_v1mnz): ... this ...
26381 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
26383 (insn_mz_<mode>): Replaced by ...
26384 (insn_mz_v8qi): ... this ...
26385 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
26387 (insn_v<n>mz): Replaced by ...
26388 (insn_v1mz): ... this ...
26389 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
26392 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
26394 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
26396 2013-03-26 Roland McGrath <mcgrathr@google.com>
26398 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
26399 than fprintf with a non-constant, non-format string.
26401 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
26403 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
26404 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
26405 operand 0 predicate.
26406 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
26407 attribute. Use general_x64nomem_operand as operand 1 predicate.
26408 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
26409 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
26410 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
26411 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
26412 (mov<mode>_insv_1): Remove expander. Merge insn with
26413 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
26414 Use general_x64nomem_operand as operand 1 predicate.
26415 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
26416 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
26417 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
26418 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
26419 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
26420 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
26421 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
26422 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
26423 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
26424 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
26425 (general_x64nomem_operand): Ditto.
26427 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
26429 * config/rtems.opt: Add -pthread option.
26431 2013-03-26 Richard Biener <rguenther@suse.de>
26433 * alias.c (find_base_term): Avoid redundant and not used recursion.
26434 (base_alias_check): Get the initial base term from the caller.
26435 (true_dependence_1): Compute and pass base terms to base_alias_check.
26436 (write_dependence_p): Likewise.
26437 (may_alias_p): Likewise.
26439 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
26441 * config/aarch64/aarch64.c (aarch64_classify_address): Support
26442 PC-relative load in SI modes and above only.
26444 2013-03-26 Xinyu Qi <xyqi@marvell.com>
26446 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
26447 * config/arm/iwmmxt.md (WCGR0): Update.
26448 (WCGR1, WCGR2, WCGR3): Likewise.
26450 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
26452 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
26453 Use x64 and nox64 isa attributes.
26455 2013-03-26 Richard Biener <rguenther@suse.de>
26457 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
26458 alignment computations and rely on get_object_alignment_1
26459 for the !TYPE_P case.
26460 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
26462 2013-03-26 Walter Lee <walt@tilera.com>
26464 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
26465 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
26467 2013-03-25 Jeff Law <law@redhat.com>
26469 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
26470 check for INTEGRAL_TYPE_P that was missing due to checking in
26471 wrong version of prior patch.
26473 2013-03-25 Walter Lee <walt@tilera.com>
26475 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
26476 TILEGX_INSN_SHUFFLEBYTES1.
26477 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
26479 (tilegx_builtins): Ditto.
26480 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
26482 2013-03-25 Walter Lee <walt@tilera.com>
26484 * config/tilegx/tilegx.md (floatsisf2): New pattern.
26485 (floatunssisf2): New pattern.
26486 (floatsidf2): New pattern.
26487 (floatunssidf2): New pattern.
26489 2013-03-25 Walter Lee <walt@tilera.com>
26491 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
26492 tests for constraint J, K, N, P.
26494 2013-03-25 Walter Lee <walt@tilera.com>
26496 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
26497 Use indirect/pcrel encoding.
26498 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
26501 2013-03-25 Steve Ellcey <sellcey@mips.com>
26503 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
26504 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
26505 * config/mips/mips.c (mips_option_override): Set IMADD default.
26506 * config/mips/mips.h (PTF_AVOID_IMADD): New.
26507 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
26508 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
26509 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
26510 * doc/invoke.texi (-mimadd/-mno-imadd): New.
26512 2013-03-25 Jeff Law <law@redhat.com>
26514 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
26515 slightly to avoid creating and folding useless trees. Simplify
26516 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
26518 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
26520 * config/i386/i386.md (*zero_extendsidi2): Merge with
26521 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
26522 * config/i386/predicates.md (x86_64_zext_operand): Rename from
26523 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
26524 targets. Clarify comment.
26526 2013-03-25 Martin Jambor <mjambor@suse.cz>
26528 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
26529 pass-through jump functions differently.
26530 (ipa_read_jump_function): Likewise. Also use setter functions to set
26533 2013-03-25 Martin Jambor <mjambor@suse.cz>
26535 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
26536 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
26538 (ipa_get_indirect_edge_target): New function.
26539 (devirtualization_time_bonus): New parameter known_aggs, pass it to
26540 ipa_get_indirect_edge_target. Update all callers.
26541 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
26542 ipa_get_indirect_edge_target_1 instead of calling
26543 ipa_get_indirect_edge_target.
26544 (create_specialized_node): Pass aggvlas to
26545 ipcp_discover_new_direct_edges.
26547 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26549 * config/arm/arm.md (f_sels, f_seld): New types.
26550 (*cmov<mode>): New pattern.
26551 * config/arm/predicates.md (arm_vsel_comparison_operator): New
26554 2013-03-25 Kai Tietz <ktietz@redhat.com>
26556 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
26557 POSIX-printf for mingw-hosted builds.
26559 2013-03-25 Richard Biener <rguenther@suse.de>
26561 PR middle-end/56694
26562 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
26563 must-not-throw stmt location.
26565 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26567 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
26568 Emit load-acquire versions when acq is true.
26569 (arm_emit_store_exclusive): Add rel parameter.
26570 Emit store-release versions when rel is true.
26571 (arm_split_compare_and_swap): Use acquire-release instructions
26573 of barriers when appropriate.
26574 (arm_split_atomic_op): Likewise.
26575 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
26576 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
26577 (VUNSPEC_SLX): Likewise.
26578 (VUNSPEC_LDA): Likewise.
26579 (VUNSPEC_STL): Likewise.
26580 * config/arm/sync.md (atomic_load<mode>): New pattern.
26581 (atomic_store<mode>): Likewise.
26582 (arm_load_acquire_exclusive<mode>): Likewise.
26583 (arm_load_acquire_exclusivesi): Likewise.
26584 (arm_load_acquire_exclusivedi): Likewise.
26585 (arm_store_release_exclusive<mode>): Likewise.
26587 2013-03-25 Catherine Moore <clm@codesourcery.com>
26589 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
26590 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
26591 * config/mip/predicates.md (lwsp_swsp_operand,
26592 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
26593 sb16_operand, db4_operand, db7_operand, ib3_operand,
26594 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
26595 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
26596 andi16_operand): New predicates.
26597 * config/mips/mips.md (compression): New attribute.
26598 (enabled): New attribute.
26599 (length): Consider compression in computing length.
26600 (shift_compression): New code attribute.
26601 (*add<mode>3): New operands. Record compression.
26602 (sub<mode>3): Likewise.
26603 (one_cmpl<mode>2): Likewise.
26604 (*and<mode>3): Likewise.
26605 (*ior<mode>3): Likewise.
26606 (unnamed pattern for xor): Likewise.
26607 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
26608 (*<optab><mode>3): Likewise.
26609 (*mov<mode>_internal: Likewise.
26610 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
26611 (mips_unsigned_immediate_p): New.
26612 (umips_lwsp_swsp_address_p): New.
26613 (m16_based_address_p): New.
26614 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
26615 (mips_unsigned_immediate_p): New prototype.
26616 (lwsp_swsp_address_p): New prototype.
26617 (m16_based_address_p): New prototype.
26618 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
26619 (mips_signed_immediate_p): New function.
26620 (m16_based_address_p): New function.
26621 (lwsp_swsp_address_p): New function.
26622 (mips_print_operand_punctuation): Recognize short delay slot insns
26623 for microMIPS.add<mode>3"
26625 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26628 * config/arm/iterators.md (v_cmp_result): New mode attribute.
26629 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
26631 2013-03-25 Richard Biener <rguenther@suse.de>
26633 PR tree-optimization/56689
26634 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
26637 2013-03-25 Richard Biener <rguenther@suse.de>
26639 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
26641 (memory_references): Likewise.
26642 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
26643 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
26644 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
26645 (gather_mem_refs_in_loops): Fold into ...
26646 (analyze_memory_references): ... this. Move initialization
26647 to tree_ssa_lim_initialize.
26648 (fill_always_executed_in): Rename to ...
26649 (fill_always_executed_in_1): ... this.
26650 (fill_always_executed_in): Move contains_call computation to
26651 this new function from ...
26652 (tree_ssa_lim_initialize): ... here.
26653 (tree_ssa_lim): Call fill_always_executed_in.
26655 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
26657 * postreload.c (reload_combine): Fix code detecting returns.
26659 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
26661 * function.c (emit_use_return_register_into_block): On cc0 targets,
26662 do not emit the sequence between cc0 setter and user.
26664 2013-03-25 Kai Tietz <ktietz@redhat.com>
26666 * config/i386/predicates.md (local_symbolic_operand): Interpret
26667 dll-imported symbols as none-local.
26669 2013-03-25 Richard Biener <rguenther@suse.de>
26671 * tree-ssa-loop-im.c (struct depend): Remove.
26672 (struct lim_aux_data): Make depends a vec of gimples.
26673 (free_lim_aux_data): Adjust.
26674 (add_dependency): Likewise.
26675 (set_level): Likewise.
26677 2013-03-25 Richard Biener <rguenther@suse.de>
26679 PR middle-end/56434
26680 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
26681 the pointer returned by calls with ECF_MALLOC set.
26683 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
26685 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
26687 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
26689 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
26690 using MMXMODE mode iterator.
26691 (*move<mode>_internal): Merge with *movv2sf_internal and
26692 *movv2sf_internal_rex64 using MMXMODE mode iterator.
26694 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
26696 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
26697 (record_last_mem_set_info): Likewise.
26699 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
26700 of XNEWVEC followed by memset.
26701 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
26703 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
26705 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
26706 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
26707 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
26708 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
26709 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
26710 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
26711 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
26712 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
26713 BARRIER_P instead of GET_CODE.
26715 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
26717 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
26718 inaccuracy in the probing code.
26720 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
26721 (ctrapdi4): Likewise.
26723 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
26725 * calls.c (expand_call): Add missing guard to code handling return
26726 of non-BLKmode structures in MSB.
26727 * function.c (expand_function_end): Likewise.
26729 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
26731 * combine.c (try_combine): Adjust comment. Do not add the set of
26732 insn #0 if the destination indirectly is set or dies in insn #2.
26733 Tidy up code to distribute a new note.
26735 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
26737 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
26738 also for alternatives 16 and 17.
26740 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
26742 * config/i386/sse.md (*mov<mode>_internal): Merge with
26743 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
26744 Emit insn template depending on type attribute. Use
26745 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
26746 movd instead of movq mnemonic for interunit moves. Rewrite mode
26747 attribute calculation. Remove unit attribute calculation.
26748 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
26749 Set prefix_data16 attribute for DImode ssemov types.
26750 Use Ym instead of y for SSE-MMX conversion alternatives.
26751 Reorder operand constraints.
26753 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
26755 * df.h (df_insn_delete): Adjust prototype.
26756 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
26757 and let it decide whether mark the basic block dirty.
26758 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
26759 * df-scan.c (df_insn_info_delete): New helper function, split
26760 off from df_insn_delete.
26761 (df_scan_free_bb_info): Use it.
26762 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
26764 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
26765 that the insn is actually an insn and it has a non-NULL basic block.
26766 Do not mark basic block dirty if only deleting a DEBUG_INSN.
26768 2013-03-22 Richard Biener <rguenther@suse.de>
26770 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
26772 (mem_ref_alloc): Do not allocate them.
26773 (refs_independent_p): Do not query or maintain a cache.
26775 2013-03-22 Richard Biener <rguenther@suse.de>
26777 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
26778 (gather_mem_refs_in_loops): Do not compute it.
26779 (analyze_memory_references): Do not allocate it.
26780 (tree_ssa_lim_finalize): Do not free it.
26781 (for_all_locs_in_loop): Do not query all_refs_in_loop.
26783 2013-03-22 Richard Biener <rguenther@suse.de>
26785 * is-a.h (as_a): Use gcc_checking_assert.
26787 2013-03-22 Ian Bolton <ian.bolton@arm.com>
26789 * config/aarch64/aarch64.c (aarch64_print_operand): New
26790 format specifier for printing a constant in hex.
26791 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
26792 format specifier for printing second operand.
26794 2013-03-22 Richard Biener <rguenther@suse.de>
26796 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
26798 (gather_mem_refs_in_loops): Perform store accumulation here.
26799 (create_vop_ref_mapping_loop): Remove.
26800 (create_vop_ref_mapping): Likewise.
26801 (analyze_memory_references): Initialize refs_stored_in_loop.
26802 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
26803 (record_indep_loop): Remove.
26804 (record_dep_loop): New function.
26805 (ref_indep_loop_p_1): Adjust to only walk over references
26806 in the loop, not its subloops.
26807 (ref_indep_loop_p): Rename to ...
26808 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
26809 maintaining a more fine-grained cache.
26810 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
26811 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
26813 2013-03-22 Richard Biener <rguenther@suse.de>
26815 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
26816 (struct mem_ref): Make accesses_in_loop a vec of a vec of
26817 aggregate mem_ref_loc.
26818 (free_mem_ref_locs): Inline into ...
26819 (memref_free): ... this and adjust.
26820 (mem_ref_alloc): Adjust.
26821 (mem_ref_locs_alloc): Remove.
26822 (record_mem_ref_loc): Adjust.
26823 (get_all_locs_in_loop): Rewrite into ...
26824 (for_all_locs_in_loop): ... this iterator.
26825 (rewrite_mem_ref_loc): New functor.
26826 (rewrite_mem_refs): Use for_all_locs_in_loop.
26827 (sm_set_flag_if_changed): New functor.
26828 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
26829 (ref_always_accessed): New functor.
26830 (ref_always_accessed_p): Use for_all_locs_in_loop.
26832 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
26834 * tree-pass.h (PROP_gimple_lvec): New.
26835 * passes.c (dump_properties): Handle PROP_gimple_lvec.
26836 (init_optimization_passes): Move pass_lower_vector.
26837 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
26839 (pass_lower_vector): Provide PROP_gimple_lvec.
26840 (pass_lower_vector_ssa): Likewise.
26841 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
26843 2013-03-21 Mark Wielaard <mjw@redhat.com>
26845 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
26847 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
26849 * config/i386/i386.md (*movdi_internal): Disparage slightly
26850 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
26851 conversion alternatives.
26853 2013-03-21 Jakub Jelinek <jakub@redhat.com>
26855 PR middle-end/48087
26856 * diagnostic.def (DK_WERROR): New kind.
26857 * diagnostic.h (werrorcount): Define.
26858 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
26859 promoted to DK_ERROR, increment DK_WERROR counter instead of
26861 * toplev.c (toplev_main): Call print_ignored_options even if
26862 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
26863 even if just werrorcount is non-zero.
26866 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
26868 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
26869 (string_cst_pool_decl): New function.
26870 (optimize_one_addr_into_implicit_ptr): New function.
26871 (resolve_addr_in_expr): Optimize DWARF location expression
26872 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
26873 which doesn't live in memory, but has DW_AT_location or
26874 DW_AT_const_value, or refers to a string literal, into
26875 DW_OP_GNU_implicit_pointer.
26876 (optimize_location_into_implicit_ptr): New function.
26877 (resolve_addr): If removing DW_AT_location of a variable because
26878 it was DW_OP_addr of address of the variable, but the variable doesn't
26879 live in memory, try to emit const value attribute for the initializer.
26881 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
26883 * tree.h (VECTOR_TYPE_P): New macro.
26884 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
26885 TYPE_MODE): Use it.
26886 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
26887 VEC_COND_EXPR cannot be lvalues.
26888 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
26890 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
26892 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
26893 Restrict the transformation to equal modes.
26895 2013-03-21 Richard Biener <rguenther@suse.de>
26897 PR tree-optimization/39326
26898 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
26899 (MEM_ANALYZABLE): Adjust.
26900 (record_mem_ref_loc): Move bitmap ops ...
26901 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
26902 unanalyzable refs, do not record locations for it.
26903 (analyze_memory_references): Allocate ref zero as shared
26905 (refs_independent_p): Do not test for unanalyzed mems here.
26906 (ref_indep_loop_p_1): Special-case disambiguation against
26907 the unanalyzed ref.
26908 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
26910 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
26912 * config/arm/arm-protos.h (tune_params): Add
26913 prefer_neon_for_64bits field.
26914 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
26915 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
26916 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
26917 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
26918 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
26919 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
26920 (arm_option_override): Handle -mneon-for-64bits new option.
26921 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
26922 (prefer_neon_for_64bits): Declare new variable.
26923 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
26924 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
26925 (arch_enabled): Handle new arch types. Remove support for onlya8
26927 (one_cmpldi2): Use new arch names.
26928 (zero_extend<mode>di2, extend<mode>di2): Ditto.
26929 * config/arm/arm.opt (mneon-for-64bits): Add option.
26930 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
26931 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
26932 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
26934 * doc/invoke.texi (-mneon-for-64bits): Document.
26936 2013-03-21 Richard Biener <rguenther@suse.de>
26938 PR tree-optimization/39326
26939 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
26940 (sort_bbs_in_loop_postorder_cmp): New function.
26941 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
26943 2013-03-21 Richard Biener <rguenther@suse.de>
26945 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
26946 (vect_insert_into_interleaving_chain): Likewise.
26947 (vect_drs_dependent_in_basic_block): Inline ...
26948 (vect_slp_analyze_data_ref_dependence): ... here. New function,
26950 (vect_analyze_data_ref_dependence): ... here. Simplify.
26951 (vect_check_interleaving): Simplify.
26952 (vect_analyze_data_ref_dependences): Likewise. Split out ...
26953 (vect_slp_analyze_data_ref_dependences): ... this new function.
26954 (dr_group_sort_cmp): New function.
26955 (vect_analyze_data_ref_accesses): Compute data-reference groups
26956 here instead of in vect_analyze_data_ref_dependence. Use
26957 a more efficient algorithm.
26958 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
26959 vect_slp_analyze_data_ref_dependences. Call
26960 vect_analyze_data_ref_accesses earlier.
26961 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
26962 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
26963 (vect_slp_analyze_data_ref_dependences): New prototype.
26965 2013-03-21 Richard Biener <rguenther@suse.de>
26967 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
26968 ref is stored in the loop.
26969 (find_refs_for_sm): Walk only over all stores.
26970 (store_motion_loop): Allocate from lim_bitmap_obstack.
26971 (store_motion): Likewise.
26973 2013-03-21 Richard Biener <rguenther@suse.de>
26975 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
26976 Update virtual SSA form.
26978 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
26980 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
26981 * configure: Regenerate.
26982 * config.in: Regenerate.
26983 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
26984 if !HAVE_LD_EH_FRAME_CIEV3.
26986 2013-03-21 Richard Biener <rguenther@suse.de>
26988 * tree-cfg.c (verify_expr_no_block): New function.
26989 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
26990 nor DECL_VALUE_EXPR have locations with associated blocks.
26991 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
26992 (clear_unused_block_pointer): Remove code dealing with
26993 blocks in DECL_DEBUG_EXPR locations.
26995 2013-03-21 Richard Biener <rguenther@suse.de>
26997 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
26998 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
26999 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
27000 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
27001 instead of DECL_DEBUG_EXPR_IS_FROM.
27002 * gimplify.c (gimplify_modify_expr): Likewise.
27003 * tree-cfg.c (verify_expr_location_1): Likewise.
27004 * tree-complex.c (create_one_component_var): Likewise.
27005 * tree-sra.c (create_access_replacement): Likewise.
27006 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
27007 (clear_unused_block_pointer): Likewise.
27008 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
27009 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
27010 * var-tracking.c (var_debug_decl): Likewise.
27011 (track_expr_p): Likewise.
27012 * tree-inline.c (add_local_variables): Likewise. Set
27013 DECL_HAS_DEBUG_EXPR_P after copying it.
27014 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
27015 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
27017 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
27020 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
27021 * configure: Regenerate.
27022 * config.in: Regenerate.
27023 * config/i386/i386.md (*movdf_internal): Use
27024 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
27025 movd instead of movq mnemonic for interunit moves.
27026 (*movdi_internal): Ditto.
27028 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27030 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
27031 (abd<mode>_3): New pattern.
27032 (aba<mode>_3): New pattern.
27033 (fabd<mode>_3): New pattern.
27035 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27037 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
27038 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
27039 occurrence of REGISTER_PREFIX as its empty string.
27041 2013-03-20 Jeff Law <law@redhat.com>
27043 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
27044 addititional equivalences for equality comparisons between an SSA_NAME
27045 and a constant where the SSA_NAME was set from a widening conversion.
27047 2013-03-20 Walter Lee <walt@tilera.com>
27049 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
27051 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27053 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
27054 depending on type attribute.
27055 (*movti_internal): Ditto.
27056 (*movtf_internal): Ditto.
27057 (*movxf_internal): Ditto.
27058 (*movdf_internal): Ditto.
27059 (*movsf_internal): Ditto.
27061 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27063 * config/i386/i386.md (*movti_internal): Set prefix attribute to
27064 maybe_vex for sselog1 and ssemov types.
27065 (*movdi_internal): Reorder operand constraints.
27066 (*movsi_internal): Ditto. Set prefix attribute to
27067 maybe_vex for sselog1 and ssemov types.
27068 (*movtf_internal): Set prefix attribute to maybe_vex
27069 for sselog1 and ssemov types.
27070 (*movdf_internal): Ditto. Set prefix_data16 attribute for
27071 DImode ssemov types. Reorder operand constraints.
27072 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
27073 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
27074 attribute for SImode ssemov types. Reorder operand constraints.
27076 2013-03-20 Martin Jambor <mjambor@suse.cz>
27078 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
27079 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
27081 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
27083 * config/rs6000/predicates.md (indexed_address, update_address_mem
27084 update_indexed_address_mem): New predicates.
27085 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
27086 attribute for load/store instructions.
27087 * config/rs6000/dfp.md (movsd_store): Likewise.
27088 (movsd_load): Likewise.
27089 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
27090 (unnamed HI->DI extend define_insn): Likewise.
27091 (unnamed SI->DI extend define_insn): Likewise.
27092 (unnamed QI->SI extend define_insn): Likewise.
27093 (unnamed QI->HI extend define_insn): Likewise.
27094 (unnamed HI->SI extend define_insn): Likewise.
27095 (unnamed HI->SI extend define_insn): Likewise.
27096 (extendsfdf2_fpr): Likewise.
27097 (movsi_internal1): Likewise.
27098 (movsi_internal1_single): Likewise.
27099 (movhi_internal): Likewise.
27100 (movqi_internal): Likewise.
27101 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
27102 attribute for load/store instructions.
27103 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
27105 (mov<mode>_softfloat): Likewise.
27106 (mov<mode>_hardfloat32): Likewise.
27107 (mov<mode>_hardfloat64): Likewise.
27108 (mov<mode>_softfloat64): Likewise.
27109 (movdi_internal32): Likewise.
27110 (movdi_internal64): Likewise.
27111 (probe_stack_<mode>): Likewise.
27113 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
27115 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
27116 floating point, and decimal floating point to reload iterator.
27118 * config/rs6000/constraints.md (wl constraint): New constraints to
27119 return FLOAT_REGS if certain options are used to reduce the number
27120 of separate patterns that exist in the file.
27121 (wx constraint): Likewise.
27122 (wz constraint): Likewise.
27124 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
27125 -mdebug=reg, print wg, wl, wx, and wz constraints.
27126 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
27127 Initialize the reload functions for 64-bit binary/decimal floating
27129 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
27130 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
27131 create the buffer on the stack to overcome not having a 32-bit
27133 (rs6000_emit_move): Likewise.
27134 (rs6000_secondary_memory_needed_rtx): Likewise.
27135 (rs6000_alloc_sdmode_stack_slot): Likewise.
27136 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
27137 via xxlxor, just like DFmode 0.0.
27139 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
27140 define as 1 if we are running on a power7 or newer.
27141 (enum r6000_reg_class_enum): Add new constraints.
27143 * config/rs6000/dfp.md (movsd): Delete, combine with binary
27144 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
27145 with other moves by using conditional constraits (wg). Use LFIWZX
27146 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
27147 (movsd splitter): Likewise.
27148 (movsd_hardfloat): Likewise.
27149 (movsd_softfloat): Likewise.
27151 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
27152 binary and decimal floating point moves.
27153 (fmove_ok): New attributes to combine binary and decimal floating
27154 point moves, and to combine power6x (mfpgpr) moves along normal
27156 (real_value_to_target): Likewise.
27157 (f32_lr): Likewise.
27158 (f32_lm): Likewise.
27159 (f32_li): Likewise.
27160 (f32_sr): Likewise.
27161 (f32_sm): Likewise.
27162 (f32_si): Likewise.
27163 (movsf): Combine binary and decimal floating point moves. Combine
27164 power6x (mfpgpr) moves with other moves by using conditional
27165 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
27166 (mov<mode> for SFmode/SDmode); Likewise.
27167 (SFmode/SDmode splitters): Likewise.
27168 (movsf_hardfloat): Likewise.
27169 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
27170 (movsf_softfloat): Likewise.
27171 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
27173 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
27174 wx and wz constraints.
27176 * config/rs6000/constraints.md (wg constraint): New constraint to
27177 return FLOAT_REGS if -mmfpgpr (power6x) was used.
27179 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
27182 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
27183 -mdebug=reg, print wg, wl, wx, and wz constraints.
27184 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
27185 Initialize the reload functions for 64-bit binary/decimal floating
27187 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
27188 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
27189 create the buffer on the stack to overcome not having a 32-bit
27191 (rs6000_emit_move): Likewise.
27192 (rs6000_secondary_memory_needed_rtx): Likewise.
27193 (rs6000_alloc_sdmode_stack_slot): Likewise.
27194 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
27195 via xxlxor, just like DFmode 0.0.
27197 * config/rs6000/dfp.md (movdd): Delete, combine with binary
27198 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
27199 with other moves by using conditional constraits (wg). Use LFIWZX
27200 and STFIWX for loading SDmode on power7.
27201 (movdd splitters): Likewise.
27202 (movdd_hardfloat32): Likewise.
27203 (movdd_softfloat32): Likewise.
27204 (movdd_hardfloat64_mfpgpr): Likewise.
27205 (movdd_hardfloat64): Likewise.
27206 (movdd_softfloat64): Likewise.
27208 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
27209 64-bit binary and decimal floating point moves.
27210 (FMOVE64X): Likewise.
27211 (movdf): Combine 64-bit binary and decimal floating point moves.
27212 Combine power6x (mfpgpr) moves with other moves by using
27213 conditional constraits (wg).
27214 (mov<mode> for DFmode/DDmode): Likewise.
27215 (DFmode/DDmode splitters): Likewise.
27216 (movdf_hardfloat32): Likewise.
27217 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
27218 (movdf_softfloat32): Likewise.
27219 (movdf_hardfloat64_mfpgpr): Likewise.
27220 (movdf_hardfloat64): Likewise.
27221 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
27222 (movdf_softfloat64): Likewise.
27223 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
27224 (reload_<mode>_load): Move to later in the file so they aren't in
27225 the middle of the floating point move insns.
27226 (reload_<mode>_store): Likewise.
27228 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
27231 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
27232 constraint if -mdebug=reg.
27233 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
27234 Enable using dd reload support if needed.
27236 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
27237 binary and decimal floating point moves in rs6000.md.
27238 (movtd_internal): Likewise.
27240 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
27241 decimal floating point moves.
27243 (movtf_internal): Likewise.
27244 (mov<mode>_internal, TDmode/TFmode): Likewise.
27245 (movtf_softfloat): Likewise.
27246 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
27248 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
27249 movdi_internal64, using wg constraint for move direct operations.
27250 (movdi_internal64): Likewise.
27252 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
27253 MODES_TIEABLE_P for selected modes. Print the numerical value of
27254 the various virtual registers. Use GPR/FPR first/last values,
27255 instead of hard coding the register numbers. Print which modes
27256 have reload functions registered.
27257 (rs6000_option_override_internal): If -mdebug=reg, trace the options
27258 settings before/after setting cpu, target and subtarget settings.
27259 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
27260 and for secondary reload failures in rs6000_secondary_reload_inner.
27261 (rs6000_secondary_reload_fail): Likewise.
27262 (rs6000_secondary_reload_inner): Likewise.
27264 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
27265 macros for first/last GPR and FPR registers.
27266 (LAST_GPR_REGNO): Likewise.
27267 (FIRST_FPR_REGNO): Likewise.
27268 (LAST_FPR_REGNO): Likewise.
27270 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
27271 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
27272 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
27273 (vcond<mode><mode>): Likewise.
27274 (vcondu<mode><mode>): Likewise.
27275 (vector_gtu<mode>): Likewise.
27276 (vector_gte<mode>): Likewise.
27277 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
27278 to prevent the compiler from converting DImode operations to TImode.
27279 (ior<mode>3): Likewise.
27280 (and<mode>3): Likewise.
27281 (one_cmpl<mode>2): Likewise.
27282 (nor<mode>3): Likewise.
27283 (andc<mode>3): Likewise.
27285 * config/rs6000/constraints.md (wt constraint): New constraint
27286 that returns VSX_REGS if TImode is allowed in VSX registers.
27288 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
27289 constant under VSX.
27291 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
27292 similar to TImode, but it is restricted to being in the GPRs.
27294 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
27295 TImode to occupy a single VSX register.
27297 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
27298 -mvsx-timode for power7/power8.
27299 (power7 cpu): Likewise.
27300 (power8 cpu): Likewise.
27302 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
27303 sure that TFmode/TDmode take up two registers if they are ever
27304 allowed in the upper VSX registers.
27305 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
27307 (rs6000_init_hard_regno_mode_ok): Likewise.
27308 (rs6000_debug_reg_global): Add debugging for PTImode and wt
27309 constraint. Print if LRA is turned on.
27310 (rs6000_option_override_internal): Give an error if -mvsx-timode
27311 and VSX is not enabled.
27312 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
27313 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
27314 to reg+offset addressing. Use PTImode when checking offset
27315 addresses for validity.
27316 (reg_offset_addressing_ok_p): Likewise.
27317 (rs6000_legitimate_offset_address_p): Likewise.
27318 (rs6000_legitimize_address): Likewise.
27319 (rs6000_legitimize_reload_address): Likewise.
27320 (rs6000_legitimate_address_p): Likewise.
27321 (rs6000_eliminate_indexed_memrefs): Likewise.
27322 (rs6000_emit_move): Likewise.
27323 (rs6000_secondary_reload): Likewise.
27324 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
27325 reloads to fpr registers to continue to use reg+offset addressing,
27326 but 64-bit reloads to altivec registers need reg+reg addressing.
27327 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
27328 it. Treat LO_SUM like a PLUS operation.
27329 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
27330 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
27331 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
27332 registers to share a register with a smaller sized type, since VSX
27333 puts scalars in the upper 64-bits.
27334 (print_operand): Add support for PTImode.
27335 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
27336 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
27337 registers, but don't have arithmetic support.
27338 (rs6000_memory_move_cost): Add test for VSX.
27339 (rs6000_opt_masks): Add -mvsx-timode.
27341 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
27344 (VSr): Use wt constraint for TImode.
27345 (VSv): Drop TImode support.
27346 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
27347 (vsx_movti_64bit): Likewise.
27348 (vsx_movti_32bit): Likewise.
27349 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
27350 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
27351 one '?' on the appropriate output constraint. Do not allow TImode
27352 logical operations on 32-bit systems.
27353 (vsx_ior<mode>3): Likewise.
27354 (vsx_xor<mode>3): Likewise.
27355 (vsx_one_cmpl<mode>2): Likewise.
27356 (vsx_nor<mode>3): Likewise.
27357 (vsx_andc<mode>3): Likewise.
27358 (vsx_concat_<mode>): Likewise.
27359 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
27361 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
27362 OPTION_MASK_VSX_TIMODE.
27363 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
27364 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
27366 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
27367 (TI2 iterator): New iterator for TImode, PTImode.
27368 (wd mode attribute): Add values for vector types.
27369 (movti_string): Replace TI move operations with operations for TImode
27370 and PTImode. Add support for TImode being allowed in VSX registers.
27371 (mov<mode>_string, TImode/PTImode): Likewise.
27372 (movti_ppc64): Likewise.
27373 (mov<mode>_ppc64, TImode/PTImode): Likewise.
27374 (TI mode splitters): Likewise.
27376 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
27379 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
27381 PR tree-optimization/56355
27382 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
27383 Also handle integers with undefined overflow.
27385 2013-03-20 Catherine Moore <clm@codesourcery.com>
27386 Maciej W. Rozycki <macro@codesourcery.com>
27387 Tom de Vries <tom@codesourcery.com>
27388 Nathan Sidwell <nathan@codesourcery.com>
27389 Iain Sandoe <iain@codesourcery.com>
27390 Nathan Froyd <froydnj@codesourcery.com>
27391 Chao-ying Fu <fu@mips.com>
27393 * doc/extend.texi: (micromips, nomicromips, nocompression):
27394 Document new function attributes.
27395 * doc/invoke.texi (minterlink-compressed, mmicromips,
27396 m14k, m14ke, m14kec): Document new options.
27397 (minterlink-mips16): Update documentation.
27398 * doc/md.texi (ZC, ZD): Document new constraints.
27399 * configure.ac (gcc_cv_as_micromips): Check if linker
27400 supports the .set micromips directive.
27401 * configure: Regenerate.
27402 * config.in: Regenerate.
27403 * config/mips/mips-tables.opt: Regenerate.
27404 * config/mips/micromips.md: New file.
27405 * constraints.md (ZC, ZD): New constraints.
27406 * config/mips/predicates.md (movep_src_register): New predicate.
27407 (movep_src_operand): New predicate.
27408 (non_volatile_mem_operand): New predicate.
27409 * config/mips/mips.md (multimem): New type.
27410 (length): Differentiate between 17-bit and 18-bit branch offsets.
27411 (MOVEP1, MOVEP2): New mode iterator.
27412 (mov_<load>l): Use ZC constraint.
27413 (mov_<load>r): Likewise.
27414 (mov_<store>l): Likewise.
27415 (mov_<store>r): Likewise.
27416 (*branch_equality<mode>_inverted): Add microMIPS support.
27417 (*branch_equality<mode>): Likewise.
27418 (*jump_absolute): Likewise.
27419 (indirect_jump_<mode>): Likewise.
27420 (tablejump_<mode>): Likewise.
27421 (<optab>_internal): Likewise.
27422 (sibcall_internal): Likewise.
27423 (sibcall_value_internal): Likewise.
27424 (prefetch): Use constraint ZD.
27425 * config/mips/mips.opt (minterlink-compressed): New option.
27426 (minterlink-mips16): Now an alias for minterlink-compressed.
27427 (mmicromips): New option.
27428 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27429 (compare_and_swap_12): Likewise.
27430 (sync_add<mode>): Likewise.
27431 (sync_<optab>_12): Likewise.
27432 (sync_old_<optab>_12): Likewise.
27433 (sync_new_<optab>_12): Likewise.
27434 (sync_nand_12): Likewise.
27435 (sync_old_nand_12): Likewise.
27436 (sync_new_nand_12): Likewise.
27437 (sync_sub<mode>): Likewise.
27438 (sync_old_add<mode>): Likewise.
27439 (sync_old_sub<mode>): Likewise.
27440 (sync_new_add<mode>): Likewise.
27441 (sync_new_sub<mode>): Likewise.
27442 (sync_<optab><mode>): Likewise.
27443 (sync_old_<optab><mode>): Likewise.
27444 (sync_new_<optab><mode>): Likewise.
27445 (sync_nand<mode>): Likewise.
27446 (sync_old_nand<mode>): Likewise.
27447 (sync_new_nand<mode>): Likewise.
27448 (sync_lock_test_and_set<mode>): Likewise.
27449 (test_and_set_12): Likewise.
27450 (atomic_compare_and_swap<mode>): Likewise.
27451 (atomic_exchange<mode>_llsc): Likewise.
27452 (atomic_fetch_add<mode>_llsc): Likewise.
27453 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27454 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27455 (umips_save_restore_pattern_p): Likewise.
27456 (umips_load_store_pair_p): Likewise.
27457 (umips_output_load_store_pair): Likewise.
27458 (umips_movep_target_p): Likewise.
27459 (umips_12bit_offset_address_p): Likewise.
27460 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27461 (mips_base_mips16): Rename this...
27462 (mips_base_compression_flags): ...to this. Update all uses.
27463 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27464 (mips_mips16_decl_p): Delete.
27465 (mips_nomips16_decl_p): Delete.
27466 (mips_get_compress_on_flags): New function.
27467 (mips_get_compress_off_flags): New function.
27468 (mips_get_compress_mode): New function.
27469 (mips_get_compress_on_name): New function.
27470 (mips_get_compress_off_name): New function.
27471 (mips_insert_attributes): Support multiple compression types.
27472 (mips_merge_decl_attributes): Likewise.
27473 (umips_12bit_offset_address_p): New function.
27474 (mips_start_function_definition): Emit .set micromips directive.
27475 (mips_call_may_need_jalx_p): New function.
27476 (mips_function_ok_for_sibcall): Add microMIPS support.
27477 (mips_print_operand_punctuation): Support short delay slots and
27479 (umips_swm_mask, umips_swm_encoding): New.
27480 (umips_build_save_restore): New function.
27481 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27482 (was_mips16_p): Remove.
27483 (old_compression_mode): New.
27484 (mips_set_compression_mode): New function.
27485 (mips_set_current_function): Add microMIPS support.
27486 (mips_option_override): Likewise.
27487 (umips_save_restore_pattern_p): New function.
27488 (umips_output_save_restore): New function.
27489 (umips_load_store_pair_p_1): New function.
27490 (umips_load_store_pair_p): New function.
27491 (umips_output_load_store_pair_1): New function.
27492 (umips_output_load_store_pair): New function.
27493 (umips_movep_target_p) New function.
27494 (mips_prepare_pch_save): Add microMIPS support.
27495 * config/mips/mips.h (TARGET_COMPRESSION): New.
27496 (TARGET_CPU_CPP_BUILTINS): Update macro
27497 to use new compression flags and to support microMIPS.
27498 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27499 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27500 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27501 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27502 (ASM_SPEC): Support mmicromips and mno-micromips.
27503 (M16STORE_REG_P): New macro.
27504 (MIPS_CALL): Support TARGET_MICROMIPS.
27505 (MICROMIPS_J): New macro.
27506 (mips_base_mips16): Rename this...
27507 (mips_base_compression_flags): ...to this.
27508 (UMIPS_12BIT_OFFSET_P): New macro.
27509 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27510 (MULTILIB_DIRNAMES): Likewise.
27511 2013-03-20 Richard Biener <rguenther@suse.de>
27513 PR tree-optimization/56661
27514 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
27515 the result does not have to be distinct.
27517 2013-03-20 Richard Biener <rguenther@suse.de>
27519 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
27522 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
27523 Steven Bosscher <steven@gcc.gnu.org>
27525 PR rtl-optimization/56605
27526 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
27528 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
27531 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
27532 that require movd instead of movq.
27534 2013-03-20 Richard Biener <rguenther@suse.de>
27536 * tree-ssa-structalias.c (struct variable_info): Add pointer
27537 to the first field of an aggregate with sub-vars. Make
27538 this and the pointer to the next subfield its ID.
27539 (vi_next): New function.
27540 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
27541 storedanything_id, integer_id): Increment by one.
27542 (new_var_info, get_call_vi, lookup_call_clobber_vi,
27543 get_call_clobber_vi): Adjust.
27544 (solution_set_expand): Simplify and speedup.
27545 (solution_set_add): Inline into ...
27546 (set_union_with_increment): ... this. Adjust accordingly.
27547 (do_sd_constraint): Likewise.
27548 (do_ds_constraint): Likewise.
27549 (do_complex_constraint): Simplify.
27550 (build_pred_graph): Adjust.
27551 (solve_graph): Likewise. Simplify and speedup.
27552 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
27553 get_constraint_for_component_ref, get_constraint_for_1,
27554 first_vi_for_offset, first_or_preceding_vi_for_offset,
27555 create_function_info_for, create_variable_info_for_1,
27556 create_variable_info_for, intra_create_variable_infos): Adjust.
27557 (init_base_vars): Push NULL for ID zero.
27558 (compute_points_to_sets): Adjust.
27560 2013-03-20 Richard Biener <rguenther@suse.de>
27562 * cfgloop.c (verify_loop_structure): Streamline and avoid
27563 ICEing on corrupt loop tree.
27564 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
27567 2013-03-20 Richard Biener <rguenther@suse.de>
27569 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
27570 check whether an SSA update is needed.
27572 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
27574 * config/mips/constraints.md (T): Rename to...
27578 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
27579 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
27581 2013-03-19 Ian Bolton <ian.bolton@arm.com>
27583 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
27584 (*subsi3_carryin_uxtw): Likewise.
27586 2013-03-19 Ian Bolton <ian.bolton@arm.com>
27588 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
27589 (*rorsi3_insn_uxtw): Likewise.
27591 2013-03-19 Ian Bolton <ian.bolton@arm.com>
27593 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
27594 (*extrsi5_insn_uxtw): Likewise.
27596 2013-03-19 Richard Biener <rguenther@suse.de>
27598 PR tree-optimization/56273
27599 * passes.c (init_optimization_passes): Move second VRP after DOM.
27601 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
27603 * config/i386/i386.md (*movti_internal): Merge from
27604 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
27605 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
27606 nox64 isa attributes.
27608 2013-03-18 Richard Biener <rguenther@suse.de>
27610 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
27612 (merge_node_constraints): Likewise.
27613 (build_succ_graph): Likewise.
27614 (valid_graph_edge): Inline into single caller.
27615 (unify_nodes): Likewise. Use bitmap_set_bit return value
27617 (scc_visit): Fix formatting and variable use.
27618 (do_sd_constraint): Use gcc_checking_assert.
27619 (do_ds_constraint): Likewise.
27620 (do_complex_constraint): Likewise.
27621 (condense_visit): Likewise. Cleanup.
27622 (dump_pred_graph): New function.
27623 (perform_var_substitution): Dump the pred-graph before
27624 variable substitution.
27625 (find_equivalent_node): Use gcc_checking_assert.
27626 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
27628 2013-03-18 Richard Biener <rguenther@suse.de>
27630 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
27631 Remove cond_expr_stmt_list argument and do not gimplify the
27633 (vect_loop_versioning): Adjust.
27634 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
27635 Cleanup to use less temporaries.
27636 (vect_create_data_ref_ptr): Cleanup.
27638 2013-03-18 Jakub Jelinek <jakub@redhat.com>
27640 PR tree-optimization/56635
27641 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
27642 require types_compatible_p types.
27644 2013-03-18 Nick Clifton <nickc@redhat.com>
27646 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
27647 spurious backslash.
27649 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
27650 Add missing line to comment describing function.
27652 2013-03-18 Richard Biener <rguenther@suse.de>
27654 PR tree-optimization/56210
27655 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
27656 Handle string / character search functions.
27657 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
27659 2013-03-18 Richard Biener <rguenther@suse.de>
27661 PR middle-end/56483
27662 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
27663 and implement properly.
27664 * gimple.h (gimple_cond_single_var_p): Remove.
27666 2013-03-18 Richard Biener <rguenther@suse.de>
27668 * tree-data-ref.h (find_data_references_in_loop): Declare.
27669 * tree-data-ref.c (get_references_in_stmt): Use a stack
27670 vector pre-allocated in the callers.
27671 (find_data_references_in_stmt): Adjust.
27672 (graphite_find_data_references_in_stmt): Likewise.
27673 (create_rdg_vertices): Likewise.
27674 (find_data_references_in_loop): Export.
27675 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
27676 Compute dependences here...
27677 (vect_analyze_data_refs): ...not here. When we encounter
27678 a non-vectorizable data reference in basic-block vectorization
27679 truncate the data reference vector. Do not bother to
27680 fixup data-dependence information for gather loads.
27681 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
27682 of data references, as reported.
27684 2013-03-18 Richard Biener <rguenther@suse.de>
27686 PR tree-optimization/3713
27687 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
27688 has_constants and expr.
27689 (stmt_has_constants): Properly valueize SSA names when deciding
27690 whether the stmt has constants.
27692 2013-03-18 Richard Biener <rguenther@suse.de>
27694 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
27695 whole function when there is nothing to do.
27696 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
27697 * tree-vectorizer.c (vectorize_loops): Update virtual and
27698 loop-closed SSA once.
27699 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
27701 2013-03-18 Richard Biener <rguenther@suse.de>
27703 PR middle-end/56113
27704 * domwalk.c (bb_postorder): New global static.
27705 (cmp_bb_postorder): New function.
27706 (walk_dominator_tree): Replace scheme imposing an order for
27707 visiting dominator sons by one sorting them at the time they
27708 are pushed on the stack.
27710 2013-03-18 Richard Biener <rguenther@suse.de>
27712 PR tree-optimization/39326
27713 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
27714 (struct mem_ref): Replace mem member with ao_ref typed member.
27715 (MEM_ANALYZABLE): Adjust.
27716 (memref_eq): Likewise.
27717 (mem_ref_alloc): Likewise.
27718 (gather_mem_refs_stmt): Likewise.
27719 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
27720 (execute_sm_if_changed_flag_set): Adjust.
27721 (execute_sm): Likewise.
27722 (ref_always_accessed_p): Likewise.
27723 (refs_independent_p): Likewise.
27724 (can_sm_ref_p): Likewise.
27726 2013-03-18 Jakub Jelinek <jakub@redhat.com>
27729 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
27730 return 1 even for !unsignedp.
27732 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
27734 * config/i386/i386.md (isa): Add x64 and nox64.
27735 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
27736 (*pushtf): Enable *roF alternative for x64 isa only.
27737 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
27738 mode attribute of integer alternatives to DImode for TARGET_64BIT.
27739 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
27740 (*movtf_internal): Merge from *movtf_internal_rex64 and
27741 *movtf_internal_sse. Use x64 and nox64 isa attributes.
27742 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
27743 nox64 isa attributes.
27744 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
27745 nox64 isa attributes.
27746 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
27748 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
27750 * config/alpha/alpha.c (TARGET_LRA_P): New define.
27752 2013-03-17 Jakub Jelinek <jakub@redhat.com>
27755 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
27756 class names. Remove trailing comma after "ALL_REGS".
27758 2013-03-16 Jan Hubicka <jh@suse.cz>
27760 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
27761 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
27762 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
27763 of cgraph_get_create_node.
27764 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
27766 2013-03-16 Jason Merrill <jason@redhat.com>
27769 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
27770 with DW_AT_default_value.
27772 2013-03-16 Jakub Jelinek <jakub@redhat.com>
27774 * BASE-VER: Set to 4.9.0.
27776 2013-03-14 Andi Kleen <ak@linux.intel.com>
27779 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
27780 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
27781 Document _x* TSX intrinsics.
27783 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
27784 David Holsgrove <david.holsgrove@xilinx.com>
27786 * configure.ac: Add MicroBlaze TLS support detection.
27787 * configure: Regenerate.
27788 * config/microblaze/microblaze-protos.h
27789 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
27790 symbol_mentioned_p, label_mentioned_p): Add prototypes.
27791 * config/microblaze/microblaze.c (microblaze_address_type): Add
27792 ADDRESS_TLS and tls_reloc address types.
27793 (microblaze_address_info): Add tls_reloc.
27794 (TARGET_HAVE_TLS): Define.
27795 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
27796 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
27797 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
27798 load_tls_operand, microblaze_call_tls_get_addr,
27799 microblaze_legitimize_tls_address): New functions.
27800 (microblaze_classify_unspec): Handle UNSPEC_TLS.
27801 (get_base_reg): Use microblaze_tls_symbol_p.
27802 (microblaze_classify_address): Handle TLS.
27803 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
27804 label_mentioned_p and microblaze_tls_referenced_p.
27805 (microblaze_legitimize_address): Handle TLS.
27806 (microblaze_address_insns): Handle ADDRESS_TLS.
27807 (pic_address_needs_scratch): Handle TLS.
27808 (print_operand_address): Handle TLS.
27809 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
27810 (microblaze_expand_move): Handle TLS.
27811 (microblaze_legitimate_constant_p): Check
27812 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
27813 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
27814 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
27815 (PIC_OFFSET_TABLE_REGNUM): Set.
27816 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
27817 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
27818 (addsi3, movsi_internal2, movdf_internal): Update constraints
27819 * config/microblaze/predicates.md (arith_plus_operand): Define
27820 (move_operand): Redefine as move_src_operand,
27821 check microblaze_tls_referenced_p.
27823 2013-03-14 Ian Bolton <ian.bolton@arm.com>
27825 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
27826 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
27828 2013-03-14 Ian Bolton <ian.bolton@arm.com>
27830 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
27833 2013-03-14 Jakub Jelinek <jakub@redhat.com>
27835 PR tree-optimization/53265
27836 * common.opt (Waggressive-loop-optimizations): New option.
27837 * tree-ssa-loop-niter.c: Include tree-pass.h.
27838 (do_warn_aggressive_loop_optimizations): New function.
27839 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
27840 if number_of_latch_executions returned constant.
27841 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
27842 early. If number_of_latch_executions returned constant, set
27843 nb_iterations_upper_bound back to it.
27844 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
27846 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
27847 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
27849 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
27850 (MULTILIB_OSDIRNAMES): Set.
27851 * genmultilib: If defaultosdirname doesn't start with :: , set
27852 defaultosdirname2 instead, clear it and emit two . multilib_raw
27853 entries instead of just one.
27855 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
27857 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
27858 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
27859 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
27860 (SUBTARGET_OVERRIDE_OPTIONS): New.
27862 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
27865 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
27866 (musermode): Convert to Var(TARGET_USERMODE).
27867 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
27868 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
27869 * config/sh/sh.c (sh_option_override): Use
27870 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
27871 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
27873 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
27875 (udivsi3_i4_single, divsi3_i4_single): Use
27876 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
27878 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
27880 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
27883 2013-03-13 Richard Biener <rguenther@suse.de>
27885 PR tree-optimization/56608
27886 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
27887 calls when vectorizing basic-blocks.
27889 2013-03-13 Jakub Jelinek <jakub@redhat.com>
27892 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
27895 2013-03-12 Jakub Jelinek <jakub@redhat.com>
27897 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
27899 2013-03-11 Jan Hubicka <jh@suse.cz>
27902 * lto-streamer-out.c (output_symbol_p): Skip references from
27903 constructors of external variables.
27905 2013-03-11 Jan Hubicka <jh@suse.cz>
27907 PR middle-end/56571
27908 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
27910 * emit-rtl.c (verify_rtx_sharing): Likewise.
27911 (copy_insn_1): Likewise.
27912 * rtl.c (copy_rtx): Likewise.
27914 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
27917 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
27918 output_operand_lossage message.
27920 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
27923 * arm.c (shift_op): Validate RTL pattern on the fly.
27924 (arm_print_operand, case 'S'): Don't use shift_operator to validate
27927 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27930 * config/pa/pa.md (call_value): Check for calls to powf and direct to
27931 new call patterns that clobber %fr12.
27932 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
27933 split and postreload patterns.
27934 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
27935 registers %fr12 and %fr12R as call used.
27937 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
27939 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
27940 (canon_address, record_store, replace_read, check_mem_read_rtx,
27941 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
27942 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
27943 rest_of_handle_dse): Likewise.
27945 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
27947 PR middle-end/56524
27948 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
27950 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
27951 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
27952 (save_optabs_if_changed): Replace with...
27953 (init_tree_optimization_optabs): ...this.
27954 * optabs.c (save_optabs_if_changed): Rename to...
27955 (init_tree_optimization_optabs): ...this. Take the optimization node
27956 as argument. Do nothing if the base optabs are already correct.
27957 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
27958 to recompute optabs.
27959 * function.h (function): Remove optabs field.
27960 * function.c (invoke_set_current_function_hook): Call
27961 init_tree_optimization_optabs. Use the result to initialize
27964 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
27966 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
27967 if GTMA_HAS_NO_INSTRUMENTATION.
27968 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
27969 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
27970 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
27971 * gimple-pretty-print.c (dump_gimple_transaction): Handle
27972 GTMA_HAS_NO_INSTRUMENTATION.
27974 2013-03-08 Jakub Jelinek <jakub@redhat.com>
27976 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
27979 2013-03-08 Marek Polacek <polacek@redhat.com>
27980 Jakub Jelinek <jakub@redhat.com>
27982 PR tree-optimization/56478
27983 * predict.c (is_comparison_with_loop_invariant_p): Change the
27984 type of loop_step to tree.
27985 (predict_loops): Adjust.
27986 (predict_iv_comparison): Perform the computations on double_ints.
27988 2013-03-08 Richard Biener <rguenther@suse.de>
27990 PR tree-optimization/56570
27991 * tree-cfg.c (verify_expr_location_1): Verify locations for
27993 * tree-sra.c (create_access_replacement): Strip locations
27994 from DECL_DEBUG_EXPRs.
27996 2013-03-08 Richard Biener <rguenther@suse.de>
27998 * tree-inline.c (expand_call_inline): Do not associate
27999 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
28000 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
28002 2013-03-08 Richard Biener <rguenther@suse.de>
28004 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
28005 or block changes with -Og. Fix for location / block encoding
28006 changes and PHI arguments with locations.
28008 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
28010 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
28012 (struct output_info): Likewise.
28013 (register_overhead): Remove bad gcc_assert.
28014 (bitmap_find_bit): If there is only a single bitmap element, do not
28015 count a miss as a search.
28016 (print_statistics): Update for counter type changes.
28017 (dump_bitmap_statistics): Likewise. Print headers such that they
28018 are properly lined up with the printed counters.
28020 2013-03-07 Jakub Jelinek <jakub@redhat.com>
28022 PR tree-optimization/56559
28023 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
28024 check that it has only a single use.
28026 2013-03-07 Richard Biener <rguenther@suse.de>
28028 * doc/invoke.texi (fwhole-program): Discourage use in combination
28031 2013-03-06 Jakub Jelinek <jakub@redhat.com>
28033 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
28035 PR tree-optimization/56539
28036 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
28037 instead of GSI_CONTINUE_LINKING as last argument to
28038 force_gimple_operand_gsi. Adjust function comment.
28040 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
28043 PR middle-end/56548
28044 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
28045 promoted mode, convert the result back to the original mode.
28047 2013-03-06 Richard Biener <rguenther@suse.de>
28049 PR middle-end/56294
28050 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
28051 (insert_updated_phi_nodes_compare_uids): New function.
28052 (update_ssa): Sort symbols_to_rename after UID before
28053 traversing it to insert PHI nodes.
28055 2013-03-06 Richard Biener <rguenther@suse.de>
28057 PR middle-end/50494
28058 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
28059 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
28062 2013-02-13 Richard Biener <rguenther@suse.de>
28065 * varasm.c (output_constant_def_1): Get the decl representing
28066 the constant as argument.
28067 (output_constant_def): Wrap output_constant_def_1.
28068 (make_decl_rtl): Use output_constant_def_1 with the decl
28069 representing the constant.
28070 (build_constant_desc): Optionally re-use a decl already
28071 representing the constant.
28072 (tree_output_constant_def): Adjust.
28074 2013-03-06 Joey Ye <joey.ye@arm.com>
28077 * gcc.c (convert_white_space): New function.
28078 (main): Handles white space in function name.
28080 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
28083 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
28084 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
28085 to SH_DIV_CALL_TABLE for TARGET_SH2.
28086 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
28088 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
28089 call-table options.
28091 2013-03-05 Sterling Augustine <saugustine@google.com>
28092 Cary Coutant <ccoutant@google.com>
28095 * dwarf2out.c (resolve_addr): Don't call
28096 remove_loc_list_addr_table_entries a second time for the same
28099 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28102 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
28103 (avoid_complex_debug_insns): New function.
28104 (expand_debug_locations): Call it.
28106 PR rtl-optimization/56484
28107 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
28108 lifetimes of hard registers on small register class machines.
28110 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
28112 * config/microblaze/microblaze-protos.h: Rename
28113 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
28114 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
28116 (microblaze_fast_interrupt_function_p): New function.
28117 (microblaze_is_interrupt_handler): Rename to
28118 microblaze_is_interrupt_variant and add fast_interrupt check.
28119 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
28120 (save_restore_insns): Likewise.
28121 (compute_frame_size): Likewise.
28122 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
28123 (microblaze_globalize_label): Likewise.
28124 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
28125 * config/microblaze/microblaze.md: Use wrapper
28126 microblaze_is_interrupt_variant.
28128 2013-03-05 Kai Tietz <ktietz@redhat.com>
28130 * sdbout.c (sdbout_one_type): Switch to current function's section
28131 supporting cold/hot.
28133 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
28135 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
28138 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28140 PR middle-end/56461
28141 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
28142 if VALGRIND_GET_VBITS is defined, temporarily make object
28143 memory all defined, and restore previous valgrind addressability
28144 and definability afterwards. Free this_object at the end.
28146 PR middle-end/56461
28147 * lra.c (lra): Call lra_clear_live_ranges if live_p,
28148 right before calling lra_create_live_ranges, also call it
28149 when clearing live_p. Only call lra_clear_live_ranges
28150 at the end if live_p.
28152 PR middle-end/56461
28153 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
28155 2013-03-05 Richard Biener <rguenther@suse.de>
28157 PR tree-optimization/56521
28158 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
28161 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
28164 * except.h (remove_unreachable_eh_regions): New prototype.
28165 * except.c (remove_eh_handler_splicer): New function, split out
28166 of remove_eh_handler.
28167 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
28168 warning about running it on many EH regions one at a time.
28169 (remove_unreachable_eh_regions_worker): New function, walk the
28170 EH tree in depth-first order and remove non-marked regions.
28171 (remove_unreachable_eh_regions): New function.
28172 * tree-eh.c (mark_reachable_handlers): New function, split out
28173 from remove_unreachable_handlers.
28174 (remove_unreachable_handlers): Use mark_reachable_handlers and
28175 remove_unreachable_eh_regions.
28176 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
28177 and remove_unreachable_eh_regions.
28179 2013-03-05 Richard Biener <rguenther@suse.de>
28181 PR middle-end/56525
28182 * loop-init.c (fix_loop_structure): Remove loops in two stages,
28183 not freeing them until the end.
28185 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28187 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
28189 2013-03-05 Richard Biener <rguenther@suse.de>
28191 PR tree-optimization/56270
28192 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
28193 of loads after scheduling an SLP instance.
28195 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28197 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
28199 (check_gcc_parallelize): Run guality.exp as a separate job from
28200 vect.exp with unsorted.exp and $(dg_target_exps) separately from
28201 struct-layout-1.exp with stackalign.exp.
28203 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
28205 PR middle-end/56461
28206 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
28207 load_index sbitmap even if some bit in it isn't set.
28209 PR middle-end/56461
28210 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
28211 (discover_iteration_bound_by_body_walk): Change queues to
28212 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
28213 spelling in comment. Call safe_push on queues[bound_index] directly.
28214 Release queues[queue_index] in every iteration unconditionally.
28215 Release bounds vector.
28217 PR middle-end/56461
28218 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
28219 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
28220 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
28223 2013-03-05 Richard Biener <rguenther@suse.de>
28226 * tree-inline.c (remap_blocks_to_null): New function.
28227 (expand_call_inline): When expanding a call stmt without
28228 an associated block inline remap all callee blocks to NULL.
28230 2013-03-05 Jakub Jelinek <jakub@redhat.com>
28232 PR rtl-optimization/56494
28233 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
28234 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
28235 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
28237 PR middle-end/56461
28238 * sel-sched-ir.c (free_sched_pools): Release
28239 succs_info_pool.stack[succs_info_pool.max_top] vectors too
28240 if succs_info_pool.max_top isn't -1.
28243 * opts.c (opts_obstack, opts_concat): Moved to...
28244 * opts-common.c (opts_obstack, opts_concat): ... here.
28246 2013-03-04 Jakub Jelinek <jakub@redhat.com>
28248 PR middle-end/56461
28249 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
28251 2013-03-04 Martin Jambor <mjambor@suse.cz>
28253 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
28254 all appropriate places.
28256 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
28258 PR tree-optimization/56424
28259 * ipa-split.c (split_function): Do not set the RSO flag if result is
28260 not by reference and its type is a register type.
28262 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
28264 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
28265 (microblaze_legitimate_pic_operand): Likewise
28266 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
28267 new function microblaze_legitimate_pic_operand
28268 * config/microblaze/microblaze-protos.h
28269 (microblaze_legitimate_pic_operand): Declare.
28271 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28273 * config/microblaze/predicates.md (call_insn_simple_operand):
28274 New predicate for supported rtx code types.
28275 * config/microblaze/microblaze.md (call_internal1): Use
28276 call_insn_simple_operand predicate.
28278 2013-03-04 Jakub Jelinek <jakub@redhat.com>
28280 PR middle-end/56461
28281 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
28282 partitions.ordered_remove.
28284 PR middle-end/56461
28285 * tree-vect-stmts.c (vectorizable_conversion): Don't call
28286 vec_oprnds0.create (1) for modifier == NONE.
28288 PR middle-end/56461
28289 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
28290 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
28291 vec_oprnds1 right before pushing anything to it for
28294 PR middle-end/56461
28295 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
28296 set nbbs to 0 instead of having separate code path.
28297 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
28298 instead of false as last argument if returning NULL.
28300 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
28302 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
28303 the attribute is now called "target" instead of "option".
28304 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
28305 * doc/tm.texi.in (Target Attributes): Likewise document the correct
28306 attribute/pragma name for TARGET_OPTION_VALID_P and
28307 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
28308 * doc/tm.texi: Regenerated.
28310 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
28312 * config/microblaze/microblaze.c:
28313 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
28314 * config/microblaze/microblaze.h: Add -mxl-reorder to
28316 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
28317 instructions emitted if TARGET_REORDER.
28318 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
28319 or 0 for -m/-mno case, but initialises as 2 to detect default use case
28322 2013-03-01 Xinliang David Li <davidxl@google.com>
28324 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
28327 2013-03-01 Jakub Jelinek <jakub@redhat.com>
28329 PR middle-end/56461
28330 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
28331 vector even when returning true. Fix up function comment formatting.
28333 PR middle-end/56461
28334 * ira-build.c (ira_loop_nodes_count): New variable.
28335 (create_loop_tree_nodes): Initialize it.
28336 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
28338 PR middle-end/56461
28339 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
28340 method on dr_chain and result_chain.
28341 * tree-vect-stmts.c (vectorizable_store): Only call
28342 result_chain.create if j == 0.
28344 PR middle-end/56461
28345 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
28346 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
28347 before overwriting it.
28349 2013-03-01 Tobias Burnus <burnus@net-b.de>
28351 * doc/extended.texi (C Extensions): Change order in @menu
28353 (Other MIPS Built-in Functions): Move last MIPS entry before
28354 "picoChip Built-in Functions".
28355 (SH Built-in Functions): Move after RX Built-in Functions.
28356 * doc/gcc.texi (Introduction): Change order in @menu
28358 * doc/md.texi (Constraints): Ditto.
28359 * gty.texi (Type Information): Ditto.
28360 (User-provided marking routines for template types): Make
28362 * doc/invoke.texi (AArch64 Options): Move before
28363 "Adapteva Epiphany Options".
28365 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
28366 Jakub Jelinek <jakub@redhat.com>
28369 * asan.c (gate_asan): Lookup no_sanitize_address instead of
28370 no_address_safety_analysis attribute.
28371 * doc/extend.texi (no_address_safety_attribute): Rename to
28372 no_sanitize_address attribute, mention no_address_safety_analysis
28373 attribute as deprecated alias.
28375 2013-02-28 Jakub Jelinek <jakub@redhat.com>
28377 PR middle-end/56461
28378 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
28379 type to vec<vec<tree> > *.
28380 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
28381 to be vec<tree> instead of vec<tree> *, set vec_defs
28382 to vNULL and call vec_defs.create (number_of_vects), adjust other
28384 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
28385 vectorizable_condition): Adjust vect_get_slp_defs callers.
28387 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28389 * config/aarch64/aarch64.c
28390 (aarch64_float_const_representable): Remove unused variable.
28392 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28394 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
28396 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28398 * config/aarch64/aarch64-builtins.c
28399 (aarch64_init_simd_builtins): Make static.
28401 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
28403 * config/aarch64/aarch64.c
28404 (aarch64_simd_make_constant): Make static.
28406 2013-02-28 Martin Jambor <mjambor@suse.cz>
28408 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
28409 with no initialization to the RHS of debug statements.
28411 2013-02-28 Martin Jambor <mjambor@suse.cz>
28413 PR tree-optimization/56294
28414 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
28416 (get_access_replacement): Do not call create_access_replacement.
28417 Assert a replacement exists.
28418 (get_repl_default_def_ssa_name): Create the replacement declaration
28421 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28423 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
28424 final_end_function.
28426 2013-02-28 Marek Polacek <polacek@redhat.com>
28428 PR rtl-optimization/56466
28429 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
28430 if we're changing a loop.
28431 (peel_loops_completely): Likewise.
28433 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
28436 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
28438 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28441 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
28442 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
28443 INTX_FTYPE_FX, FX_FTYPE_INTX.
28444 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
28446 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28448 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
28449 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
28450 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
28451 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
28452 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
28453 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
28454 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
28455 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
28456 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
28457 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
28458 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
28459 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
28460 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
28461 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
28462 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
28463 (avrxmega6): Increase max flash segments from 5 to 6.
28464 * config/avr/t-multilib: Regenerate.
28465 * config/avr/avr-tables.opt: Regenerate.
28466 * doc/avr-mmcu.texi: Regenerate.
28468 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
28470 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
28471 (avr_device_to_arch): Rename to avr_device_to_ld.
28472 (avr_device_to_as): New prototype.
28473 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
28474 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
28475 * config/avr/driver-avr.c (avr_device_to_as): New.
28476 (avr_device_to_arch): Rename to avr_device_to_ld.
28478 2013-02-27 Jakub Jelinek <jakub@redhat.com>
28480 PR middle-end/56461
28481 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
28482 method on dr_chain and result_chain.
28484 PR middle-end/56461
28485 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
28486 pointer_set_destroy on not_executed_last_iteration.
28488 PR middle-end/56461
28489 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
28491 PR middle-end/56461
28492 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
28493 FOR_EACH_DEFINED_FUNCTION when freeing state.
28495 PR middle-end/56461
28496 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
28498 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
28501 PR middle-end/56461
28502 * ipa-cp.c (decide_whether_version_node): Call vec_free on
28503 known_aggs[i].items and release known_aggs vector.
28505 PR middle-end/56461
28506 * ipa-reference.c (propagate): Free node_info even for alias nodes.
28508 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28510 * config/microblaze/microblaze.c (microblaze_emit_compare):
28511 Use xor for EQ/NE comparisions.
28512 * config/microblaze/microblaze.md (cstoresf4): Add constraints
28513 (cbranchsf4): Adjust operator to comparison_operator.
28515 2013-02-27 Jakub Jelinek <jakub@redhat.com>
28517 PR middle-end/56461
28518 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
28520 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
28521 vec_safe_push, always update *slot.
28522 (redirect_edge_var_map_clear): Use vec_free.
28523 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
28524 (free_var_map_entry): Use vec_free.
28525 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
28526 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
28528 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
28530 PR middle-end/45472
28531 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
28532 when the may_trap_p bit of the exprs being merged differs.
28533 Reorder tests for speculativeness in the logical and operator.
28535 2013-02-27 Jakub Jelinek <jakub@redhat.com>
28537 * incpath.c (add_standard_paths): Use reconcat instead of concat
28538 where appropriate and avoid leaking memory.
28540 * opts.h: Include obstack.h.
28541 (opts_concat): New prototype.
28542 (opts_obstack): New declaration.
28543 * opts.c (opts_concat): New function.
28544 (opts_obstack): New variable.
28545 (init_options_struct): Call gcc_init_obstack on opts_obstack.
28546 (finish_options): Use opts_concat instead of concat
28547 and XOBNEWVEC instead of XNEWVEC.
28548 * opts-common.c (generate_canonical_option, decode_cmdline_option,
28549 generate_option): Likewise.
28550 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
28551 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
28554 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
28555 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
28557 2013-02-26 Jakub Jelinek <jakub@redhat.com>
28559 PR middle-end/56461
28560 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
28562 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
28564 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
28565 (arm_block_move_unaligned_straight): Likewise.
28566 (arm_adjust_block_mem): Likewise.
28568 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
28571 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
28572 temp, cond and label.
28573 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
28576 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
28577 * config/c6x/c6x.h (dbx_register_map): Update declaration.
28580 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
28581 of prologue/epilogue functions.
28584 * config/tilegx/tilegx.c (tilegx_expand_prologue):
28585 Remove unused variable cfa_offset.
28586 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
28589 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
28590 type promotion to unsigned.
28593 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
28594 for HOST_WIDE_INT of 32 bit / same size as int.
28595 (arm_block_move_unaligned_straight): Likewise.
28596 (arm_adjust_block_mem): Likewise.
28599 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
28602 2013-02-26 Marek Polacek <polacek@redhat.com>
28604 PR tree-optimization/56426
28605 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
28607 2013-02-26 Richard Biener <rguenther@suse.de>
28610 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
28611 unused variable loops.
28613 2013-02-26 Jakub Jelinek <jakub@redhat.com>
28615 PR tree-optimization/56448
28616 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
28617 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
28618 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
28619 later operands of the references, or even first operand for
28620 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
28622 PR tree-optimization/56443
28623 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
28624 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
28625 to type_for_mode langhook.
28627 2013-02-25 Matt Turner <mattst88@gmail.com>
28629 * doc/invoke.texi: Document r4700.
28631 2013-02-25 Richard Biener <rguenther@suse.de>
28633 PR tree-optimization/56175
28634 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
28636 (simplify_bitwise_binary): ... here. Also guard the conversion
28637 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
28639 2013-02-25 Catherine Moore <clm@codesourcery.com>
28642 2013-02-24 Catherine Moore <clm@codesourcery.com>
28643 Maciej W. Rozycki <macro@codesourcery.com>
28644 Tom de Vries <tom@codesourcery.com>
28645 Nathan Sidwell <nathan@codesourcery.com>
28646 Iain Sandoe <iain@codesourcery.com>
28647 Nathan Froyd <froydnj@codesourcery.com>
28648 Chao-ying Fu <fu@mips.com>
28650 * doc/extend.texi: (micromips, nomicromips, nocompression):
28651 Document new function attributes.
28652 * doc/invoke.texi (minterlink-compressed, mmicromips,
28653 m14k, m14ke, m14kec): Document new options.
28654 (minterlink-mips16): Update documentation.
28655 * doc/md.texi (ZC, ZD): Document new constraints.
28656 * configure.ac (gcc_cv_as_micromips): Check if linker
28657 supports the .set micromips directive.
28658 * configure: Regenerate.
28659 * config.in: Regenerate.
28660 * config/mips/mips-tables.opt: Regenerate.
28661 * config/mips/micromips.md: New file.
28662 * constraints.md (ZC, AD): New constraints.
28663 * config/mips/predicates.md (movep_src_register): New predicate.
28664 (movep_src_operand): New predicate.
28665 (non_volatile_mem_operand): New predicate.
28666 * config/mips/mips.md (multimem): New type.
28667 (length): Differentiate between 17-bit and 18-bit branch offsets.
28668 (MOVEP1, MOVEP2): New mode iterator.
28669 (mov_<load>l): Use ZC constraint.
28670 (mov_<load>r): Likewise.
28671 (mov_<store>l): Likewise.
28672 (mov_<store>r): Likewise.
28673 (*branch_equality<mode>_inverted): Add microMIPS support.
28674 (*branch_equality<mode>): Likewise.
28675 (*jump_absolute): Likewise.
28676 (indirect_jump_<mode>): Likewise.
28677 (tablejump_<mode>): Likewise.
28678 (<optab>_internal): Likewise.
28679 (sibcall_internal): Likewise.
28680 (sibcall_value_internal): Likewise.
28681 (prefetch): Use constraint ZD.
28682 * config/mips/mips.opt (minterlink-compressed): New option.
28683 (minterlink-mips16): Now an alias for minterlink-compressed.
28684 (mmicromips): New option.
28685 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
28686 (compare_and_swap_12): Likewise.
28687 (sync_add<mode>): Likewise.
28688 (sync_<optab>_12): Likewise.
28689 (sync_old_<optab>_12): Likewise.
28690 (sync_new_<optab>_12): Likewise.
28691 (sync_nand_12): Likewise.
28692 (sync_old_nand_12): Likewise.
28693 (sync_new_nand_12): Likewise.
28694 (sync_sub<mode>): Likewise.
28695 (sync_old_add<mode>): Likewise.
28696 (sync_old_sub<mode>): Likewise.
28697 (sync_new_add<mode>): Likewise.
28698 (sync_new_sub<mode>): Likewise.
28699 (sync_<optab><mode>): Likewise.
28700 (sync_old_<optab><mode>): Likewise.
28701 (sync_new_<optab><mode>): Likewise.
28702 (sync_nand<mode>): Likewise.
28703 (sync_old_nand<mode>): Likewise.
28704 (sync_new_nand<mode>): Likewise.
28705 (sync_lock_test_and_set<mode>): Likewise.
28706 (test_and_set_12): Likewise.
28707 (atomic_compare_and_swap<mode>): Likewise.
28708 (atomic_exchange<mode>_llsc): Likewise.
28709 (atomic_fetch_add<mode>_llsc): Likewise.
28710 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
28711 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
28712 (umips_save_restore_pattern_p): Likewise.
28713 (umips_load_store_pair_p): Likewise.
28714 (umips_output_load_store_pair): Likewise.
28715 (umips_movep_target_p): Likewise.
28716 (umips_12bit_offset_address_p): Likewise.
28717 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
28718 (mips_base_mips16): Rename this...
28719 (mips_base_compression_flags): ...to this. Update all uses.
28720 (mips_attribute_table): Add micromips, nomicromips and nocompression.
28721 (mips_mips16_decl_p): Delete.
28722 (mips_nomips16_decl_p): Delete.
28723 (mips_get_compress_on_flags): New function.
28724 (mips_get_compress_off_flags): New function.
28725 (mips_get_compress_mode): New function.
28726 (mips_get_compress_on_name): New function.
28727 (mips_get_compress_off_name): New function.
28728 (mips_insert_attributes): Support multiple compression types.
28729 (mips_merge_decl_attributes): Likewise.
28730 (umips_12bit_offset_address_p): New function.
28731 (mips_start_function_definition): Emit .set micromips directive.
28732 (mips_call_may_need_jalx_p): New function.
28733 (mips_function_ok_for_sibcall): Add microMIPS support.
28734 (mips_print_operand_punctuation): Support short delay slots and
28736 (umips_swm_mask, umips_swm_encoding): New.
28737 (umips_build_save_restore): New function.
28738 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
28739 (was_mips16_p): Remove.
28740 (old_compression_mode): New.
28741 (mips_set_compression_mode): New function.
28742 (mips_set_current_function): Add microMIPS support.
28743 (mips_option_override): Likewise.
28744 (umips_save_restore_pattern_p): New function.
28745 (umips_output_save_restore): New function.
28746 (umips_load_store_pair_p_1): New function.
28747 (umips_load_store_pair_p): New function.
28748 (umips_output_load_store_pair_1): New function.
28749 (umips_output_load_store_pair): New function.
28750 (umips_movep_target_p) New function.
28751 (mips_prepare_pch_save): Add microMIPS support.
28752 * config/mips/mips.h (TARGET_COMPRESSION): New.
28753 (TARGET_CPU_CPP_BUILTINS): Update macro
28754 to use new compression flags and to support microMIPS.
28755 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
28756 (MIPS_ARCH_FLOAT_SPEC): Likewise.
28757 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
28758 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
28759 (ASM_SPEC): Support mmicromips and mno-micromips.
28760 (M16STORE_REG_P): New macro.
28761 (MIPS_CALL): Support TARGET_MICROMIPS.
28762 (MICROMIPS_J): New macro.
28763 (mips_base_mips16): Rename this...
28764 (mips_base_compression_flags): ...to this.
28765 (UMIPS_12BIT_OFFSET_P): New macro.
28766 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
28767 (MULTILIB_DIRNAMES): Likewise.
28769 2013-02-25 Tom de Vries <tom@codesourcery.com>
28771 PR rtl-optimization/56131
28772 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
28773 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
28774 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
28776 2013-02-25 Tobias Burnus <burnus@net-b.de>
28778 * doc/invoke.texi (-fsanitize=): Move from optimization
28779 to debugging options.
28781 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
28783 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
28785 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
28786 Alexander Monakov <amonakov@ispras.ru>
28788 PR middle-end/56077
28789 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
28790 flush pending lists also on non-jumps. Adjust comment.
28792 2013-02-24 Catherine Moore <clm@codesourcery.com>
28793 Maciej W. Rozycki <macro@codesourcery.com>
28794 Tom de Vries <tom@codesourcery.com>
28795 Nathan Sidwell <nathan@codesourcery.com>
28796 Iain Sandoe <iain@codesourcery.com>
28797 Nathan Froyd <froydnj@codesourcery.com>
28798 Chao-ying Fu <fu@mips.com>
28800 * doc/extend.texi: (micromips, nomicromips, nocompression):
28801 Document new function attributes.
28802 * doc/invoke.texi (minterlink-compressed, mmicromips,
28803 m14k, m14ke, m14kec): Document new options.
28804 (minterlink-mips16): Update documentation.
28805 * doc/md.texi (ZC, ZD): Document new constraints.
28806 * configure.ac (gcc_cv_as_micromips): Check if linker
28807 supports the .set micromips directive.
28808 * configure: Regenerate.
28809 * config.in: Regenerate.
28810 * config/mips/mips-tables.opt: Regenerate.
28811 * config/mips/micromips.md: New file.
28812 * constraints.md (ZC, AD): New constraints.
28813 * config/mips/predicates.md (movep_src_register): New predicate.
28814 (movep_src_operand): New predicate.
28815 (non_volatile_mem_operand): New predicate.
28816 * config/mips/mips.md (multimem): New type.
28817 (length): Differentiate between 17-bit and 18-bit branch offsets.
28818 (MOVEP1, MOVEP2): New mode iterator.
28819 (mov_<load>l): Use ZC constraint.
28820 (mov_<load>r): Likewise.
28821 (mov_<store>l): Likewise.
28822 (mov_<store>r): Likewise.
28823 (*branch_equality<mode>_inverted): Add microMIPS support.
28824 (*branch_equality<mode>): Likewise.
28825 (*jump_absolute): Likewise.
28826 (indirect_jump_<mode>): Likewise.
28827 (tablejump_<mode>): Likewise.
28828 (<optab>_internal): Likewise.
28829 (sibcall_internal): Likewise.
28830 (sibcall_value_internal): Likewise.
28831 (prefetch): Use constraint ZD.
28832 * config/mips/mips.opt (minterlink-compressed): New option.
28833 (minterlink-mips16): Now an alias for minterlink-compressed.
28834 (mmicromips): New option.
28835 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
28836 (compare_and_swap_12): Likewise.
28837 (sync_add<mode>): Likewise.
28838 (sync_<optab>_12): Likewise.
28839 (sync_old_<optab>_12): Likewise.
28840 (sync_new_<optab>_12): Likewise.
28841 (sync_nand_12): Likewise.
28842 (sync_old_nand_12): Likewise.
28843 (sync_new_nand_12): Likewise.
28844 (sync_sub<mode>): Likewise.
28845 (sync_old_add<mode>): Likewise.
28846 (sync_old_sub<mode>): Likewise.
28847 (sync_new_add<mode>): Likewise.
28848 (sync_new_sub<mode>): Likewise.
28849 (sync_<optab><mode>): Likewise.
28850 (sync_old_<optab><mode>): Likewise.
28851 (sync_new_<optab><mode>): Likewise.
28852 (sync_nand<mode>): Likewise.
28853 (sync_old_nand<mode>): Likewise.
28854 (sync_new_nand<mode>): Likewise.
28855 (sync_lock_test_and_set<mode>): Likewise.
28856 (test_and_set_12): Likewise.
28857 (atomic_compare_and_swap<mode>): Likewise.
28858 (atomic_exchange<mode>_llsc): Likewise.
28859 (atomic_fetch_add<mode>_llsc): Likewise.
28860 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
28861 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
28862 (umips_save_restore_pattern_p): Likewise.
28863 (umips_load_store_pair_p): Likewise.
28864 (umips_output_load_store_pair): Likewise.
28865 (umips_movep_target_p): Likewise.
28866 (umips_12bit_offset_address_p): Likewise.
28867 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
28868 (mips_base_mips16): Rename this...
28869 (mips_base_compression_flags): ...to this. Update all uses.
28870 (mips_attribute_table): Add micromips, nomicromips and nocompression.
28871 (mips_mips16_decl_p): Delete.
28872 (mips_nomips16_decl_p): Delete.
28873 (mips_get_compress_on_flags): New function.
28874 (mips_get_compress_off_flags): New function.
28875 (mips_get_compress_mode): New function.
28876 (mips_get_compress_on_name): New function.
28877 (mips_get_compress_off_name): New function.
28878 (mips_insert_attributes): Support multiple compression types.
28879 (mips_merge_decl_attributes): Likewise.
28880 (umips_12bit_offset_address_p): New function.
28881 (mips_start_function_definition): Emit .set micromips directive.
28882 (mips_call_may_need_jalx_p): New function.
28883 (mips_function_ok_for_sibcall): Add microMIPS support.
28884 (mips_print_operand_punctuation): Support short delay slots and
28886 (umips_swm_mask, umips_swm_encoding): New.
28887 (umips_build_save_restore): New function.
28888 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
28889 (was_mips16_p): Remove.
28890 (old_compression_mode): New.
28891 (mips_set_compression_mode): New function.
28892 (mips_set_current_function): Add microMIPS support.
28893 (mips_option_override): Likewise.
28894 (umips_save_restore_pattern_p): New function.
28895 (umips_output_save_restore): New function.
28896 (umips_load_store_pair_p_1): New function.
28897 (umips_load_store_pair_p): New function.
28898 (umips_output_load_store_pair_1): New function.
28899 (umips_output_load_store_pair): New function.
28900 (umips_movep_target_p) New function.
28901 (mips_prepare_pch_save): Add microMIPS support.
28902 * config/mips/mips.h (TARGET_COMPRESSION): New.
28903 (TARGET_CPU_CPP_BUILTINS): Update macro
28904 to use new compression flags and to support microMIPS.
28905 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
28906 (MIPS_ARCH_FLOAT_SPEC): Likewise.
28907 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
28908 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
28909 (ASM_SPEC): Support mmicromips and mno-micromips.
28910 (M16STORE_REG_P): New macro.
28911 (MIPS_CALL): Support TARGET_MICROMIPS.
28912 (MICROMIPS_J): New macro.
28913 (mips_base_mips16): Rename this...
28914 (mips_base_compression_flags): ...to this.
28915 (UMIPS_12BIT_OFFSET_P): New macro.
28916 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
28917 (MULTILIB_DIRNAMES): Likewise.
28919 2013-02-24 Jakub Jelinek <jakub@redhat.com>
28922 * target-globals.c (save_target_globals): For init_reg_sets and
28923 target_reinit remporarily set this_fn_optabs to this_target_optabs.
28925 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
28927 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
28928 * config/aarch64/t-aarch64
28929 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
28931 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
28933 PR inline-asm/56148
28934 * lra-constraints.c (process_alt_operands): Reload operand
28935 conflicting with earlier clobber only if no more other conflicting
28938 2013-02-22 Jakub Jelinek <jakub@redhat.com>
28941 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
28942 if not linking a shared library.
28944 2013-02-22 Seth LaForge <sethml@google.com>
28946 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
28948 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
28950 * config/arm/arm.md (split for extendsidi): Update condition.
28951 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
28952 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
28953 (qhs_zextenddi_cstr): Likewise.
28955 2013-02-21 Jakub Jelinek <jakub@redhat.com>
28957 PR middle-end/56420
28958 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
28959 avoid signed wrapping.
28960 (expand_mult): Handle properly multiplication by
28961 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
28962 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
28963 in the compiler if coeff is HOST_WIDE_INT_MIN.
28964 (expand_divmod): Don't make ext_op1 static, change it's type to
28965 uhwi. Avoid undefined behavior in -INTVAL (op1).
28967 PR rtl-optimization/50339
28968 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
28970 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
28971 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
28972 into splitting_ashiftrt field.
28973 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
28975 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
28978 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
28980 PR middle-end/56108
28981 * trans-mem.c (execute_tm_mark): Do not expand transactions that
28982 are sure to go irrevocable.
28984 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
28986 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
28987 scalars are valid operands.
28989 2013-02-21 Martin Jambor <mjambor@suse.cz>
28991 PR tree-optimization/56310
28992 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
28993 only matching indices and non-negative final offsets.
28994 (intersect_aggregates_with_edge): Pass src_idx to
28995 agg_replacements_to_vector. Pass src_idx insstead of index to
28996 intersect_with_agg_replacements.
28998 2013-02-21 Martin Jambor <mjambor@suse.cz>
29000 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
29001 instead of hard-wired defaults.
29003 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
29005 * doc/invoke.texi (MIPS Options): Update documentation of the
29006 floating-point multiply-accumulate instruction restrictions.
29008 2013-02-21 Kostya Serebryany <kcc@google.com>
29010 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
29011 asan_shadow_offset on x86_64 linux.
29013 2013-02-21 Richard Biener <rguenther@suse.de>
29015 PR tree-optimization/56415
29017 2013-02-11 Richard Biener <rguenther@suse.de>
29019 PR tree-optimization/56273
29020 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
29023 2013-02-21 Jakub Jelinek <jakub@redhat.com>
29026 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
29029 PR inline-asm/56405
29030 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
29031 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
29033 2013-02-20 Jan Hubicka <jh@suse.cz>
29035 PR tree-optimization/56265
29036 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
29037 when target is referenced for first time.
29039 2013-02-20 Richard Biener <rguenther@suse.de>
29041 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
29042 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
29043 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
29044 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
29045 not return anything.
29046 (rename_ssa_copies): Do not remove unused locals.
29047 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
29048 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
29049 * passes.c (execute_function_todo): Do not schedule unused locals
29050 removal if cleanup_tree_cfg did something.
29051 * tree-ssa-live.c (remove_unused_locals): Dump statistics
29052 about the number of removed locals.
29054 2013-02-20 Richard Biener <rguenther@suse.de>
29056 PR tree-optimization/56398
29057 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
29059 2013-02-20 Martin Jambor <mjambor@suse.cz>
29061 PR tree-optimization/55334
29062 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
29063 restricted pointers to arrays.
29065 2013-02-20 Richard Biener <rguenther@suse.de>
29066 Jakub Jelinek <jakub@redhat.com>
29068 PR tree-optimization/56396
29069 * tree-ssa-ccp.c (n_const_val): New static variable.
29070 (get_value): Return NULL for SSA names we don't have a lattice
29072 (ccp_initialize): Initialize n_const_val.
29073 * tree-ssa-copy.c (n_copy_of): New static variable.
29074 (init_copy_prop): Initialize n_copy_of.
29075 (get_value): Return NULL_TREE for SSA names we don't have a
29078 2013-02-20 Martin Jambor <mjambor@suse.cz>
29080 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
29082 2013-02-20 Richard Biener <rguenther@suse.de>
29084 * genpreds.c (write_lookup_constraint): Do not compare first
29085 letter of the constraint again.
29087 2013-02-20 Richard Biener <rguenther@suse.de>
29089 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
29091 (get_use_iv_cost): Terminate hashtable walk when coming across
29094 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
29096 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
29097 reassociation for avx2 targets.
29099 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29101 * config/microblaze/microblaze.c: microblaze_has_clz = 0
29102 Add version check for v8.10.a to enable microblaze_has_clz
29103 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
29104 version and TARGET_PATTERN_COMPARE check
29105 * config/microblaze/microblaze.md: New clzsi2 instruction
29107 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29109 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
29110 function before branching.
29112 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
29114 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
29116 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
29118 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
29120 PR middle-end/55889
29121 * sel-sched.c: Include ira.h.
29122 (implicit_clobber_conflict_p): New function.
29123 (moveup_expr): Use it.
29124 * Makefile.in (sel-sched.o): Depend on ira.h.
29126 2013-02-19 Richard Biener <rguenther@suse.de>
29128 PR tree-optimization/56384
29129 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
29130 (vn_hash_type): Split out from ...
29131 (vn_hash_constant_with_type): ... here.
29132 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
29133 (vn_phi_eq): Compare types from vn_phi_s structure.
29134 (vn_phi_lookup): Populate vn_phi_s type.
29135 (vn_phi_insert): Likewise.
29137 2013-02-19 Jakub Jelinek <jakub@redhat.com>
29139 PR tree-optimization/56350
29140 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
29141 if haven't found reduction or nested cycle operand, rather than
29142 asserting we must find it.
29144 PR tree-optimization/56381
29145 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
29148 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
29149 Jakub Jelinek <jakub@redhat.com>
29152 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
29153 (swap_optab_enable): Same.
29154 (init_all_optabs): Use argument instead of global.
29155 * tree.h (struct tree_optimization_option): New field target_optabs.
29156 * expr.h (init_all_optabs): Add argument to prototype.
29157 (TREE_OPTIMIZATION_OPTABS): New.
29158 (save_optabs_if_changed): Protoize.
29159 * optabs.h: Declare this_fn_optabs.
29160 * optabs.c (save_optabs_if_changed): New.
29161 Declare this_fn_optabs.
29162 (init_optabs): Add argument to init_all_optabs() call.
29163 * function.c (invoke_set_current_function_hook): Handle per
29165 * function.h (struct function): New field optabs.
29166 * config/mips/mips.c (mips_set_mips16_mode): Handle when
29167 optimization_current_node has changed.
29168 * target-globals.h (save_target_globals_default_opts): Protoize.
29169 * target-globals.c (save_target_globals_default_opts): New.
29171 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29174 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
29175 registers %fr12 and %fr12R as call used.
29178 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
29179 and HImode, require all displacements to be an integer multiple of
29181 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
29182 only allow QImode and HImode when reload is in progress and strict is
29183 true. Likewise for symbolic addresses. Use base14_operand to check
29184 displacements in REG+BASE addresses.
29186 2013-02-18 Richard Biener <rguenther@suse.de>
29188 PR tree-optimization/56366
29189 * tree-vect-loop.c (get_initial_def_for_induction): Properly
29190 handle sign-conversion of outer-loop initial induction value.
29192 2013-02-18 Richard Biener <rguenther@suse.de>
29194 PR middle-end/56349
29195 * cfghooks.c (merge_blocks): If we merge a latch into another
29196 block adjust references to it.
29197 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
29198 (verify_loop_structure): Verify that a recorded latch is in fact
29201 2013-02-18 Richard Biener <rguenther@suse.de>
29203 PR tree-optimization/56321
29204 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
29205 order SSA name release and virtual operand unlinking.
29207 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29209 * config/microblaze/microblaze.md (save_stack_block): Define.
29210 (restore_stack_block): Likewise.
29212 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29214 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
29215 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
29216 * config/microblaze/microblaze.c (microblaze_option_override):
29217 Bail out early for PIC modes when target does not support PIC.
29219 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29221 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
29222 Replace with a microblaze version.
29223 (microblaze_trampoline_init): Adapt for microblaze.
29224 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
29227 2013-02-16 Jakub Jelinek <jakub@redhat.com>
29228 Dodji Seketeli <dodji@redhat.com>
29231 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
29232 (instrument_mem_region_access): Do not forget to always put
29233 instrumentation of the of 'base' and 'base + len' in a "if (len !=
29234 0) statement, even for cases where either 'base' or 'base + len'
29235 are not instrumented -- because they have been previously
29236 instrumented. Simplify the logic by putting all the statements
29237 instrument 'base + len' inside a sequence, and then insert that
29238 sequence right before the current insertion point. Then, to
29239 instrument 'base + len', just get an iterator on that statement.
29240 And do not forget to update the pointer to iterator the function
29241 received as argument.
29243 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
29245 PR rtl-optimization/56348
29246 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
29248 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
29250 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
29251 (clean_graph_dump_file): Pass base to start_graph_dump.
29253 2013-02-14 Richard Henderson <rth@redhat.com>
29256 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
29258 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
29260 * collect2-aix.h: Define F_LOADONLY.
29262 2013-02-14 Richard Biener <rguenther@suse.de>
29265 * varasm.c (output_constant_def_1): Get the decl representing
29266 the constant as argument.
29267 (output_constant_def): Wrap output_constant_def_1.
29268 (make_decl_rtl): Use output_constant_def_1 with the decl
29269 representing the constant.
29270 (build_constant_desc): Optionally re-use a decl already
29271 representing the constant.
29272 (tree_output_constant_def): Adjust.
29274 2013-02-14 Dodji Seketeli <dodji@redhat.com>
29277 * asan.c (instrument_builtin_call): Really put the length of the
29278 second source argument into src1_len.
29280 2013-02-13 Jakub Jelinek <jakub@redhat.com>
29282 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
29283 argument. If it is false, don't create edge from then_bb to
29285 (insert_if_then_before_iter): Pass true to it.
29286 (build_check_stmt): Pass false to it.
29287 (transform_statements): Flush hash table only on extended basic
29288 block boundaries, rather than at the beginning of every bb.
29289 Don't flush hash table on nonfreeing_call_p calls.
29290 * tree-flow.h (nonfreeing_call_p): New prototype.
29291 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
29293 2013-02-13 David S. Miller <davem@davemloft.net>
29295 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
29297 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
29300 * ira.c (max_regno_before_ira): Move from ...
29302 (fix_reg_equiv_init): Use max_regno_before_ira instead of
29305 2013-02-13 Jakub Jelinek <jakub@redhat.com>
29307 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
29309 2013-02-13 Richard Biener <rguenther@suse.de>
29312 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
29313 globals in MEM_REFs.
29315 2013-02-13 Richard Biener <rguenther@suse.de>
29317 * loop-init.c (loop_optimizer_init): Clear loop state when
29318 re-initializing preserved loops.
29319 * loop-unswitch.c (unswitch_single_loop): Return whether
29320 we unswitched the loop. Do not verify loop state here.
29321 (unswitch_loops): When we unswitched a loop discover new loops.
29323 2013-02-13 Kostya Serebryany <kcc@google.com>
29325 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
29327 * sanitizer.def: Rename __asan_init to __asan_init_v1.
29329 2013-02-12 Dodji Seketeli <dodji@redhat.com>
29331 Avoid instrumenting duplicated memory access in the same basic block
29332 * Makefile.in (asan.o): Add new dependency on hash-table.h
29333 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
29334 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
29335 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
29336 (free_mem_ref_resources, has_mem_ref_been_instrumented)
29337 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
29338 (get_mem_ref_of_assignment): New functions.
29339 (get_mem_refs_of_builtin_call): Extract from
29340 instrument_builtin_call and tweak a little bit to make it fit with
29342 (instrument_builtin_call): Use the new
29343 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
29344 of is_gimple_builtin_call.
29345 (instrument_derefs, instrument_mem_region_access): Insert the
29346 instrumented memory reference into the hash table.
29347 (maybe_instrument_assignment): Renamed instrument_assignment into
29348 this, and change it to advance the iterator when instrumentation
29349 actually happened and return true in that case. This makes it
29350 homogeneous with maybe_instrument_assignment, and thus give a
29351 chance to callers to be more 'regular'.
29352 (transform_statements): Clear the memory reference hash table
29353 whenever we enter a new BB, when we cross a function call, or when
29354 we are done transforming statements. Use
29355 maybe_instrument_assignment instead of instrumentation. No more
29356 need to special case maybe_instrument_assignment and advance the
29357 iterator after calling it; it's now handled just like
29358 maybe_instrument_call. Update comment.
29360 2013-02-13 Richard Biener <rguenther@suse.de>
29362 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
29363 Fix loop discovery code.
29365 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
29367 PR inline-asm/56148
29368 * lra-constraints.c (process_alt_operands): Match early clobber
29369 operand with itself. Check conflicts with earlyclobber only if
29370 the operand is not reloaded. Prefer to reload conflicting operand
29371 if earlyclobber and matching operands are the same.
29373 2013-02-12 Richard Biener <rguenther@suse.de>
29376 * lto-streamer-out.c (write_symbol): Do not output symbols
29377 for hard register variables.
29379 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
29382 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
29383 (umulsidi3_insn, mulsidi3_insn): New insns.
29385 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
29387 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
29388 (struct tune_params): Add vec_costs field.
29389 * config/arm/arm.c (arm_builtin_vectorization_cost)
29390 (arm_add_stmt_cost): New functions.
29391 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
29392 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
29393 (arm_default_vec_cost): New struct of type cpu_vec_costs.
29394 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
29395 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
29396 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
29397 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
29399 2013-02-12 Richard Biener <rguenther@suse.de>
29402 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
29403 decls again if possible.
29405 2013-02-12 Richard Biener <rguenther@suse.de>
29407 PR middle-end/56288
29408 * tree-ssa.c (verify_ssa_name): Fix check, move
29409 SSA_NAME_IN_FREE_LIST check up.
29411 2013-02-12 Jakub Jelinek <jakub@redhat.com>
29412 Steven Bosscher <steven@gcc.gnu.org>
29414 PR rtl-optimization/56151
29415 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
29416 equal to op0 or op1, and last_insn pattern is CODE operation
29417 with MEM dest and one of the operands matches that MEM.
29419 2013-02-11 Sriraman Tallam <tmsriram@google.com>
29421 * doc/extend.texi: Document Function Multiversioning and "default"
29422 parameter string to target attribute.
29423 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
29424 target attribute parameter is "default".
29425 (ix86_compare_version_priority): Remove checks for target attribute.
29426 (ix86_mangle_function_version_assembler_name): Change error to sorry.
29427 Remove check for target attribute equal to NULL. Add assert.
29428 (ix86_generate_version_dispatcher_body): Change error to sorry.
29430 2013-02-11 Iain Sandoe <iain@codesourcery.com>
29431 Jack Howarth <howarth@bromo.med.uc.edu>
29432 Patrick Marlier <patrick.marlier@gmail.com>
29435 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
29436 define ENDFILE_SPEC as TM_DESTRUCTOR.
29437 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
29439 2013-02-11 Alexander Potapenko <glider@google.com>
29440 Jack Howarth <howarth@bromo.med.uc.edu>
29441 Jakub Jelinek <jakub@redhat.com>
29444 * config/darwin.c (cdtor_record): Rename ctor_record.
29445 (sort_cdtor_records): Rename sort_ctor_records.
29446 (finalize_dtors): New routine to sort destructors by
29447 priority before use in assemble_integer.
29448 (machopic_asm_out_destructor): Use finalize_dtors if needed.
29450 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
29452 PR rtl-optimization/56275
29453 * simplify-rtx.c (avoid_constant_pool_reference): Check that
29454 offset is non-negative and less than cmode size before
29455 calling simplify_subreg.
29457 2013-02-11 Richard Biener <rguenther@suse.de>
29459 PR tree-optimization/56264
29460 * cfgloop.h (fix_loop_structure): Adjust prototype.
29461 * loop-init.c (fix_loop_structure): Return the number of
29462 newly discovered loops.
29463 * tree-cfgcleanup.c (repair_loop_structures): When new loops
29464 are discovered, do a full loop-closed SSA rewrite.
29466 2013-02-11 Richard Biener <rguenther@suse.de>
29468 PR tree-optimization/56273
29469 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
29471 (check_array_ref): Fix missing newline in dumps.
29472 (search_for_addr_array): Likewise.
29474 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
29476 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
29478 2013-02-09 Jakub Jelinek <jakub@redhat.com>
29481 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
29483 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
29485 PR rtl-optimization/56246
29486 * lra-constraints.c (simplify_operand_subreg): Try to reuse
29488 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
29489 constraints are satisfied.
29491 2013-02-08 Jeff Law <law@redhat.com>
29494 * emit-rtl.c (reg_is_parm_p): New function.
29495 * regs.h (reg_is_parm_p): New prototype.
29496 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
29497 callee-clobbered registers.
29499 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
29502 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
29503 If there is no implicit builtin declaration, just return NULL.
29505 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
29507 * config/i386/sse.md (FMAMODEM): New mode iterator.
29508 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
29509 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
29511 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
29513 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
29514 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
29515 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
29517 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29519 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
29520 (microblaze*-*-elf): Likewise.
29521 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
29523 * config/microblaze/microblaze-c.c: Add builtin defines for
29524 _LITTLE_ENDIAN and _BIG_ENDIAN.
29525 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
29526 add to TARGET_DEFAULT flags.
29527 Expand ASM_SPEC and LINK_SPEC.
29528 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
29529 * config/microblaze/microblaze.md: Update extendsidi2 and
29530 movdi_internal instructions to use low-order / high-order reg
29532 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
29533 options and inversemask / mask of LITTLE_ENDIAN.
29534 * config/microblaze/t-microblaze: Expand multilib options to
29535 include mlittle-endian (le) and update exceptions patterns.
29537 2013-02-08 Jakub Jelinek <jakub@redhat.com>
29539 PR rtl-optimization/56195
29540 * lra-constraints.c (get_reload_reg): Don't reuse regs
29541 if they have smaller mode than requested, if they have
29542 wider mode than requested, try to return a SUBREG.
29544 PR tree-optimization/56250
29545 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
29546 if type is unsigned and code isn't MULT_EXPR.
29548 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
29550 PR tree-optimization/56064
29551 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
29552 bits according to mode.
29553 * fixed-value.h (fixed_from_double_int)
29554 (const_fixed_from_double_int): Adjust comments.
29556 2013-02-08 Richard Biener <rguenther@suse.de>
29559 * lto-streamer.h (struct data_in): Remove current_file, current_line
29560 and current_col members.
29561 * lto-streamer-out.c (lto_output_location): Stream changed bits
29562 en-block for efficiency.
29563 * lto-streamer-in.c (clear_line_info): Remove.
29564 (lto_input_location): Cache current file, line and column
29565 globally via local statics. Read changed bits en-block.
29566 (input_function): Do not call clear_line_info.
29567 (lto_read_body): Likewise.
29568 (lto_input_toplevel_asms): Likewise.
29570 2013-02-08 Michael Matz <matz@suse.de>
29572 PR tree-optimization/52448
29573 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
29574 (nt_call_phase): New static.
29575 (add_or_mark_expr): Only mark accesses with newer phase than any
29577 (nonfreeing_call_p): New.
29578 (nt_init_block): Update nt_call_phase, mark blocks as visited.
29579 (nt_fini_block): Keep blocks marked as visited.
29580 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
29582 2013-02-08 Richard Biener <rguenther@suse.de>
29584 * ira.c (ira): Free broken dominator information.
29586 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
29588 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
29590 2013-02-08 Marek Polacek <polacek@redhat.com>
29592 * cfgloop.c (verify_loop_structure): Add more checking of headers.
29594 2013-02-08 Richard Biener <rguenther@suse.de>
29596 PR middle-end/56181
29597 * cfgloop.h (flow_loops_find): Adjust.
29598 (bb_loop_header_p): Declare.
29599 * cfgloop.c (bb_loop_header_p): New function split out from ...
29600 (flow_loops_find): ... here. Adjust function signature,
29601 support incremental loop structure update.
29602 (verify_loop_structure): Cleanup. Verify a loop is a loop.
29603 * cfgloopmanip.c (fix_loop_structure): Move ...
29604 * loop-init.c (fix_loop_structure): ... here.
29605 (apply_loop_flags): Split out from ...
29606 (loop_optimizer_init): ... here.
29607 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
29608 in incremental mode, only remove dead loops here.
29610 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
29613 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
29614 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
29615 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
29616 (*round<mode>3.libgcc): New insns for fixed-modes.
29617 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
29618 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
29619 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
29620 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
29621 implementations. Define to __builtin_avr_absFX,
29622 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
29623 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
29624 __builtin_avr_countlsFX, respectively.
29625 * config/avr/avr-c.c (target.h): Include it.
29626 (enum avr_builtin_id): New enum.
29627 (avr_resolve_overloaded_builtin): New static function.
29628 (avr_register_target_pragmas): Use it to set
29629 targetm.resolve_overloaded_builtin.
29630 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
29631 tree nodes used by DEF_BUILTIN.
29632 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
29633 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
29634 <AVR_BUILTIN_xxBITS>: Same.
29636 2013-02-08 Richard Biener <rguenther@suse.de>
29638 * cfgloop.c (verify_loop_structure): Properly handle
29639 a loop exiting to another loop header.
29640 * ira-int.h (ira_loops): Remove.
29641 * ira.c (ira_loops): Remove.
29642 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
29643 (do_reload): Use loop_optimizer_finalize.
29644 * ira-build.c (create_loop_tree_nodes): Use get_loops and
29645 number_of_loops to access the loop tree.
29646 (more_one_region_p): Likewise.
29647 (finish_loop_tree_nodes): Likewise.
29648 (rebuild_regno_allocno_maps): Likewise.
29649 (mark_loops_for_removal): Likewise.
29650 (mark_all_loops_for_removal): Likewise.
29651 (remove_unnecessary_regions): Likewise.
29652 (ira_build): Likewise.
29653 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
29655 2013-02-08 Richard Biener <rguenther@suse.de>
29657 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
29658 * ipa-pure-const.c (analyze_function): Avoid calling
29659 mark_irreducible_loops twice.
29660 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
29662 2013-02-07 David S. Miller <davem@davemloft.net>
29664 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
29666 * var-tracking.c (vt_add_function_parameter): Test the presence of
29667 HAVE_window_save properly and do not remap argument registers when
29668 we have a leaf function.
29670 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
29673 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
29675 * config/i386/i386.c (ix86_print_operand): Ditto.
29677 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
29679 * lra-constraints.c (process_alt_operands): Fix recently added comment.
29681 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
29683 PR rtl-optimization/56225
29684 * lra-constraints.c (process_alt_operands): Check that reload hard
29685 reg can hold value for strict_low_part.
29687 2013-02-07 Jakub Jelinek <jakub@redhat.com>
29690 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
29691 dwarf2out_end_function.
29692 (in_first_function_p, maybe_at_text_label_p,
29693 first_loclabel_num_not_at_text_label): New variables.
29694 (dwarf2out_var_location): In the first function find out
29695 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
29696 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
29699 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
29701 PR rtl-optimization/56178
29702 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
29703 SUBREG of a register. Tidy up related block of code.
29704 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
29705 note if the source is a register or a SUBREG of a register.
29707 2013-02-07 Jakub Jelinek <jakub@redhat.com>
29710 * config/rs6000/rs6000.md (ptrm): New mode attr.
29711 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
29712 call_value_indirect_aix<pttrsize>,
29713 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
29716 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
29718 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
29719 if -bnortl. Convert to strcmp and strncmp.
29721 2013-02-07 Alan Modra <amodra@gmail.com>
29724 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
29725 addresses won't wrap when offsetting.
29726 (rs6000_secondary_reload): Provide secondary reloads needed for
29727 wrapping LO_SUM addresses.
29729 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
29731 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
29732 MACH, just __MACH__.
29734 2013-02-06 Richard Biener <rguenther@suse.de>
29736 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
29737 instead of calling fix_loop_structure.
29739 2013-02-06 Jakub Jelinek <jakub@redhat.com>
29741 PR middle-end/56217
29742 * omp-low.c (use_pointer_for_field): Return false if
29743 lower_send_shared_vars doesn't generate any copy-out code.
29745 2013-02-06 Tom de Vries <tom@codesourcery.com>
29747 PR rtl-optimization/56131
29748 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
29749 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
29750 of the label is NULL. Add comment.
29752 2013-02-05 Jakub Jelinek <jakub@redhat.com>
29754 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
29757 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
29758 (STATIC_LIBTSAN_LIBS): Likewise.
29759 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
29760 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
29761 is defined, don't add anything else beyond that.
29762 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
29763 (LINK_COMMAND_SPEC): Use them.
29765 PR tree-optimization/56205
29766 * tree-stdarg.c (check_all_va_list_escapes): Return true if
29767 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
29768 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
29770 2013-02-05 Richard Biener <rguenther@suse.de>
29772 PR tree-optimization/53342
29773 PR tree-optimization/53185
29774 * tree-vectorizer.h (vect_check_strided_load): Remove.
29775 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
29776 not disallow peeling for vectorized strided loads.
29777 (vect_check_strided_load): Make static and simplify.
29778 (vect_analyze_data_refs): Adjust.
29779 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
29780 correctly when vectorizing strided loads.
29782 2013-02-05 Richard Biener <rguenther@suse.de>
29784 * doc/install.texi: Refer to ISL, not PPL.
29786 2013-02-05 Jan Hubicka <jh@suse.cz>
29788 PR tree-optimization/55789
29789 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
29791 2013-02-05 Jan Hubicka <jh@suse.cz>
29793 PR tree-optimization/55789
29794 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
29795 the dead call anyway.
29797 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
29800 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
29802 2013-02-04 Alexander Potapenko <glider@google.com>
29803 Jack Howarth <howarth@bromo.med.uc.edu>
29804 Jakub Jelinek <jakub@redhat.com>
29807 * config/darwin.c (sort_ctor_records): Stabilized qsort
29808 on constructor priority by using original position.
29809 (finalize_ctors): New routine to sort constructors by
29810 priority before use in assemble_integer.
29811 (machopic_asm_out_constructor): Use finalize_ctors if needed.
29813 2013-02-04 Jakub Jelinek <jakub@redhat.com>
29816 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
29817 about visibility on artificial decls.
29818 * config/sol2.c (solaris_assemble_visibility): Likewise.
29820 2013-02-04 Kai Tietz <ktietz@redhat.com>
29823 * config/i386/i386.c (function_value_ms_64): Add additional valtype
29824 argument and improve checking of return-argument types for 16-byte
29826 (ix86_function_value_1): Add additional valtype argument on call
29827 of function_value_64.
29828 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
29829 handling infunction_value_64 function.
29831 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
29833 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
29835 2013-02-04 Richard Biener <rguenther@suse.de>
29837 PR tree-optimization/56188
29838 * tree-ssa-structalias.c (label_visit): Consider case with
29839 initially non-empty points-to set.
29840 (perform_var_substitution): Dump node mapping and clean up.
29842 2013-02-04 Richard Guenther <rguenther@suse.de>
29845 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
29846 node prevail as last resort.
29847 (lto_symtab_merge_decls): Remove guard on LTRANS here.
29848 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
29850 2013-02-04 Richard Biener <rguenther@suse.de>
29852 PR tree-optimization/56113
29853 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
29855 (equiv_class_lookup_or_add): ... this.
29856 (label_visit): Adjust and fix error in previous patch.
29857 (perform_var_substitution): Adjust.
29859 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
29861 * config/sh/divtab.c: Fix formatting and comments throughout the file.
29862 * config/sh/sh4-300.md: Likewise.
29863 * config/sh/sh4a.md: Likewise.
29864 * config/sh/constraints.md: Likewise.
29865 * config/sh/sh.md: Likewise.
29866 * config/sh/netbsd-elf.h: Likewise.
29867 * config/sh/predicates.md: Likewise.
29868 * config/sh/sh-protos.h: Likewise.
29869 * config/sh/ushmedia.h: Likewise.
29870 * config/sh/linux.h: Likewise.
29871 * config/sh/sh.c: Likewise.
29872 * config/sh/superh.h: Likewise.
29873 * config/sh/elf.h: Likewise.
29874 * config/sh/sh4.md: Likewise.
29875 * config/sh/sh.h: Likewise.
29877 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29879 * config/pa/constraints.md: Adjust unused letters. Change "T"
29880 constraint to match_test floating_point_store_memory_operand().
29881 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
29882 (base14_operand): New.
29883 (floating_point_store_memory_operand): New.
29884 (integer_store_memory_operand): Revise to use base14_operand and
29885 reg_plus_base_memory_operand.
29886 (move_dest_operand): Allow symbolic_memory_operands.
29887 (symbolic_memory_operand): Check for LO_SOM.
29888 (symbolic_operand): Change default case to break.
29889 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
29890 CONST_DOUBLE values to be reloaded by putting them into memory when
29891 the destination is a floating point register.
29892 (movdf): Remove code to handle CONST_DOUBLE.
29894 (reload_indf_r1): New.
29895 (reload_insf_r1): New.
29896 Consistently use "Q" and "T" constraints with integer and floating
29897 point move instructions, respectively.
29898 (movdi): Remove FAIL.
29899 Change predicate for source operand unamed DImode move from
29900 general_operand to move_src_operand.
29901 (umulsidi3): Change predicate for destination operand to
29903 Likewise for similar unamed patterns.
29904 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
29905 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
29906 (hppa_legitimize_address): Simplify mask calculation.
29907 (pa_emit_move_sequence): Revised handling of secondary reloads from
29908 REG+D addresses for floating point loads and stores. Directly handle
29909 loading CONST0_RTX (mode) to a floating point register.
29910 (pa_secondary_reload): Handle reloading DF and SFmode constant values
29911 to floating point registers. Don't restrict secondary reloads to
29912 floating point registers to integer modes. Revise some comments and
29914 (TARGET_LEGITIMATE_ADDRESS_P): Define.
29915 (pa_legitimate_address_p): New.
29916 (pa_legitimize_reload_address): New.
29917 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
29918 (STRICT_REG_OK_FOR_BASE_P): New.
29919 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
29920 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
29922 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
29923 Andrew Dixie <andrewd@gentrack.com>
29925 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
29928 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
29930 * expmed.c (extract_bit_field_1): Pass the full width of the
29931 structure to get_best_reg_extraction_insn.
29933 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
29936 * configure.ac (use_cxa_atexit): Add AIX.
29937 * configure: Regenerate.
29939 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
29941 2013-02-01 Jakub Jelinek <jakub@redhat.com>
29944 * final.c (need_profile_function): New variable.
29945 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
29946 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
29947 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
29948 notes, targetm.asm_out.function_prologue doesn't emit anything,
29949 HAVE_prologue and profiler should be emitted before prologue,
29950 set need_profile_function instead of emitting it.
29951 (final_scan_insn): If need_profile_function, emit
29952 profile_function on the first NOTE_INSN_BASIC_BLOCK or
29953 NOTE_INSN_FUNCTION_BEG note.
29955 2013-02-01 Richard Henderson <rth@redhat.com>
29957 * config/rs6000/rs6000.md (smulditi3): New.
29960 * config/alpha/alpha.md (umulditi3): New.
29962 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
29964 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
29965 (ASM_OUTPUT_ALIGNED_LOCAL): New.
29967 2013-02-01 Richard Biener <rguenther@suse.de>
29969 PR tree-optimization/56113
29970 * tree-ssa-structalias.c (label_visit): Reduce work for
29971 single-predecessor nodes.
29973 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
29975 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
29976 range isn't testing for zero.
29978 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
29980 PR middle-end/56113
29981 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
29983 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
29984 Nick Clifton <nickc@redhat.com>
29986 * config/v850/constraints.md (Q): Define as a memory constraint.
29987 * config/v850/predicates.md (label_ref_operand): New predicate.
29988 (e3v5_shift_operand): New predicate.
29989 (ior_operator): New predicate.
29990 * config/v850/t-v850: Add e3v5 multilib.
29991 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
29992 (v850_gen_movdi): Prototype.
29993 * config/v850/v850.c: Add support for e3v5 architecture.
29994 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
29996 (construct_save_jarl): Add e3v5 long JARL support.
29997 (v850_adjust_insn_length): New function. Adjust length of call
29998 insns when using e3v5 instructions.
29999 (v850_gen_movdi): New function: Generate instructions to move a
30001 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
30002 (CPP_SPEC): Define __v850e3v5__ as appropriate.
30003 (TARGET_USE_FPU): Enable for e3v5.
30004 (CONST_OK_FOR_W): New macro.
30005 (ADJUST_INSN_LENGTH): Define.
30006 * config/v850/v850.md (UNSPEC_LOOP): Define.
30007 (attr cpu): Add v850e3v5.
30008 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
30009 (movdi): New pattern.
30010 (movdi_internal): New pattern.
30011 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
30012 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
30013 (cstoresf4): Likewise.
30014 (cstoredf4): Likewise.
30015 (insv): New pattern.
30016 (rotlso3_a): New pattern.
30017 (rotlsi3_b): New pattern
30018 (rotlsi3_v850e3v5): New pattern.
30019 (doloop_begin): New pattern.
30020 (fix_loop_counter): New pattern.
30021 (doloop_end): New pattern.
30022 (branch_normal): Add e3v5 long branch support.
30023 (branch_invert): Likewise.
30024 (branch_z_normal): Likewise.
30025 (branch_z_invert): Likewise.
30026 (branch_nz_normal): Likewise.
30027 (branch_nz_invert): Likewise.
30028 (call_internal_short): Add e3v5 register-indirect JARL support.
30029 (call_internal_long): Likewise.
30030 (call_value_internal_short): Likewise.
30031 (call_value_internal_long): Likewise.
30032 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
30033 (mloop): New option.
30034 * config.gcc: Add support for configuring v840e3v5 target.
30035 * doc/invoke.texi: Document new v850 specific command line options.
30037 2013-01-31 Paul Koning <ni1d@arrl.net>
30041 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
30042 children if parent is a class.
30043 (prune_unused_types_prune): Don't add DW_AT_declaration.
30045 2013-01-31 Richard Biener <rguenther@suse.de>
30047 PR tree-optimization/56157
30048 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
30049 match up operand with SLP child.
30051 2013-01-31 Jason Merrill <jason@redhat.com>
30054 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
30055 parameters the first time.
30056 (gen_scheduled_generic_parms_dies): Check completeness here.
30058 2013-01-31 Richard Biener <rguenther@suse.de>
30060 PR middle-end/53073
30061 * common.opt (faggressive-loop-optimizations): New flag,
30062 enabled by default.
30063 * doc/invoke.texi (faggressive-loop-optimizations): Document.
30064 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
30065 infer_loop_bounds_from_undefined by it.
30067 2013-01-31 Richard Biener <rguenther@suse.de>
30069 PR tree-optimization/56150
30070 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
30071 visit virtual operands.
30072 (find_uses_to_rename_bb): Likewise.
30074 2013-01-31 Richard Biener <rguenther@suse.de>
30076 PR tree-optimization/56150
30077 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
30078 mixed store non-store stmts.
30080 2013-01-30 Jakub Jelinek <jakub@redhat.com>
30083 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
30084 LIBASAN_EARLY_SPEC is defined.
30085 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
30086 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
30088 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
30091 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
30092 invalid args instead of ICEing on it.
30093 (ix86_valid_target_attribute_tree): Return error_mark_node if
30094 ix86_valid_target_attribute_inner_p failed.
30095 (ix86_valid_target_attribute_p): Return false only if
30096 ix86_valid_target_attribute_tree returned error_mark_node. Allow
30097 target("default") attribute.
30098 (sorted_attr_string): Change argument from const char * to tree,
30099 merge in all target attribute arguments rather than just one.
30100 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
30101 instead of free. Avoid using strcat.
30102 (ix86_mangle_function_version_assembler_name): Mangle
30103 target("default") as if no target attribute is present. Adjust
30104 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
30105 instead of xmalloc and XDELETEVEC instead of free.
30106 (ix86_function_versions): Don't return true if one of the decls
30107 doesn't have target attribute. If they don't and one of the decls
30108 is DECL_FUNCTION_VERSIONED, report an error. Adjust
30109 sorted_attr_string caller. Use XDELETEVEC instead of free.
30110 (ix86_supports_function_versions): Remove.
30111 (make_name): Fix up formatting.
30112 (make_dispatcher_decl): Remove resolver_name and its initialization.
30113 Avoid leaking memory.
30114 (is_function_default_version): Return true if there is
30115 target("default") attribute rather than no target attribute at all.
30116 (make_resolver_func): Avoid leaking memory.
30117 (ix86_generate_version_dispatcher_body): Likewise.
30118 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
30119 * target.def (supports_function_versions): Remove.
30120 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
30121 * doc/tm.texi: Regenerated.
30123 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
30125 PR rtl-optimization/56144
30126 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
30127 for values with side effects.
30129 2013-01-30 Richard Biener <rguenther@suse.de>
30131 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
30132 (sparseset_pop): Likewise.
30133 * cfganal.c (compute_idf): Likewise. Increase work-stack size
30134 to be able to use quick_push in the worker loop.
30136 2013-01-30 Marek Polacek <polacek@redhat.com>
30138 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
30140 2013-01-30 Richard Biener <rguenther@suse.de>
30143 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
30145 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
30147 PR tree-optimization/56064
30148 * fixed-value.c (fixed_from_double_int): New function.
30149 * fixed-value.h (fixed_from_double_int): New prototype.
30150 (const_fixed_from_double_int): New static inline function.
30151 * fold-const.c (native_interpret_fixed): New static function.
30152 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
30153 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
30154 (native_encode_fixed): New static function.
30155 (native_encode_expr) <FIXED_CST>: Use it.
30156 (native_interpret_int): Move double_int worker code to...
30157 * double-int.c (double_int::from_buffer): ...this new static method.
30158 * double-int.h (double_int::from_buffer): Prototype it.
30160 2013-01-30 Richard Biener <rguenther@suse.de>
30162 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
30163 New pointer-map and obstack.
30164 (init_alias_vars): Allocate pointer-map and obstack.
30165 (delete_points_to_sets): Free them.
30166 (find_what_var_points_to): Cache result.
30167 (find_what_p_points_to): Adjust for changed interface of
30168 find_what_var_points_to.
30169 (compute_points_to_sets): Likewise.
30170 (ipa_pta_execute): Likewise.
30172 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
30174 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
30175 * configure: Regenerate.
30176 * config.in: Regenerate.
30177 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
30178 #nobits/#progbits if supported.
30180 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
30183 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
30184 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
30185 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
30187 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30189 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
30190 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
30192 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30194 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
30195 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
30197 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30199 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
30201 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
30202 * config/arm/cortex-a7.md: New bypasses using
30203 arm_mac_accumulator_is_result.
30205 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30207 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
30208 (cortex_a7_neon_mla): Likewise.
30209 (cortex_a7_fpfmad): New reservation.
30210 (cortex_a7_fpmacs): Use ffmas and update required units.
30211 (cortex_a7_fpmuld): Update required units and latency.
30212 (cortex_a7_fpmacd): Likewise.
30213 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
30214 (cortex_a7_neon). Likewise.
30215 (bypass) Update participating units.
30217 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
30219 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
30220 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
30222 * config/arm/vfp11.md (vfp_farith): Use ffmas.
30223 (vfp_fmul): Use ffmad.
30224 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
30225 (cortex_r4_fmacd): Use ffmad.
30226 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
30227 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
30228 (cortex_a9_fmacd): Use ffmad.
30229 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
30230 (cortex_a8_vfp_macd): Use ffmad.
30231 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
30232 (cortex_a5_fpmacd): Use ffmad.
30233 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
30234 (cortex_a15_vfp_macd): Use ffmad.
30235 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
30237 2013-01-29 Jason Merrill <jason@redhat.com>
30240 * varasm.c (default_assemble_visibility): Don't warn about
30241 visibility on artificial decls.
30243 2013-01-29 Richard Biener <rguenther@suse.de>
30245 PR tree-optimization/56113
30246 * tree-ssa-structalias.c (equiv_class_lookup): Also return
30248 (label_visit): Free duplicate bitmaps and record the leader instead.
30249 (perform_var_substitution): Adjust.
30251 2013-01-29 Richard Biener <rguenther@suse.de>
30253 PR tree-optimization/55270
30254 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
30255 the CFG, schedule loops for fixup.
30257 2013-01-29 Nick Clifton <nickc@redhat.com>
30259 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
30262 2013-01-28 Leif Ekblad <leif@rdos.net>
30264 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
30265 * config/i386/i386.h (TARGET_RDOS): New macro.
30266 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
30267 * config/i386/i386.c (ix86_option_override_internal): For 64bit
30268 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
30269 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
30270 DEFAULT_LARGE_SECTION_THRESHOLD.
30271 * config/i386/i386.md (R14_REG, R15_REG): New constants.
30272 * config/i386/rdos.h: New file.
30273 * config/i386/rdos64.h: New file.
30275 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
30278 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
30281 2013-01-28 Jakub Jelinek <jakub@redhat.com>
30283 PR rtl-optimization/56117
30284 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
30285 call cselib_lookup_from_insn on the MEM before calling
30286 add_insn_mem_dependence.
30288 2013-01-28 Richard Biener <rguenther@suse.de>
30290 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
30291 to a stmt that didn't have one.
30292 (copy_phis_for_bb): Likewise for PHI arguments.
30293 (copy_debug_stmt): Likewise for debug stmts.
30295 2013-01-28 Richard Biener <rguenther@suse.de>
30297 PR tree-optimization/56034
30298 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
30299 (partition_builtin_p): Adjust.
30300 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
30301 it is the last partition.
30302 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
30303 up the vertex for the definition.
30304 (classify_partition): Classify whether a partition is a
30305 PKIND_REDUCTION, thus has uses outside of the loop.
30306 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
30307 Merge all PKIND_REDUCTION partitions into the last partition.
30308 (tree_loop_distribution): Seed partitions from reductions as well.
30310 2013-01-28 Jakub Jelinek <jakub@redhat.com>
30312 PR tree-optimization/56125
30313 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
30314 pow(x,c) into sqrt(x) * powi(x, n/2) or
30315 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
30316 optimizing for size.
30317 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
30318 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
30321 PR tree-optimization/56094
30322 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
30323 to UNKNOWN_LOCATION while gimplifying expr.
30325 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
30328 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
30329 operand 0 in movabs insn template for -masm=intel asm alternative.
30330 (*movabs<mode>_2): Ditto for operand 1.
30332 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
30335 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
30338 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30340 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
30343 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
30345 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
30346 Undef to avoid warning.
30348 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
30350 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
30351 * configure: Regenerate.
30353 2013-01-25 Jakub Jelinek <jakub@redhat.com>
30355 PR tree-optimization/56098
30356 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
30357 for stmts with volatile ops.
30358 (cond_store_replacement): Don't optimize if assign has volatile ops.
30359 (cond_if_else_store_replacement_1): Don't optimize if either
30360 then_assign or else_assign have volatile ops.
30361 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
30364 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30366 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
30368 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30370 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
30371 missing ':' in asm example.
30373 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
30375 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
30376 entries into lane and laneq entries.
30377 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
30378 Remove AdvSIMD scalar modes.
30379 (aarch64_sq<r>dmulh_laneq<mode>): New.
30380 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
30382 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
30383 builtin implementations to relfect changes in RTL in aarch64-simd.md.
30384 * config/aarch64/iterators.md (VCOND): New.
30387 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
30390 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
30391 Add NULL LIBNAME argument to existing definitions.
30392 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
30393 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
30394 * config/avr/avr.c (DEF_BUILTIN): Same.
30395 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
30396 (avr_expand_builtin): Expand to a vanilla call if a libgcc
30397 implementation is available (DECL_ASSEMBLER_NAME is set).
30398 (avr_fold_absfx): New static function.
30399 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
30400 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
30401 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
30402 AVR_BUILTIN_ABSLLK.
30403 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
30404 (abshk, absk, abslk, absllk): Provide as static inline functions.
30406 2013-01-25 Marek Polacek <polacek@redhat.com>
30408 PR tree-optimization/56035
30409 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
30411 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
30413 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
30414 (*movtf_internal_rex64): Add (!o,C) alternative
30415 (*movxf_internal_rex64): Ditto.
30416 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
30418 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
30420 * doc/invoke.texi: fix typo.
30421 * doc/objc.texi: fix typo.
30423 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
30425 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
30426 for the first two alternatives.
30428 2013-01-24 Diego Novillo <dnovillo@google.com>
30430 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
30431 (ggc-zone.o): Remove.
30432 * configure.ac: Remove option --with-gc.
30433 * configure: Re-generate.
30434 * doc/install.texi: Remove documentation for --with-gc.
30435 * gengtype.c (write_enum_defn): Remove. Update all users.
30436 (write_Types_process_field): Remove generation of gt_e_* argument.
30437 (output_type_enum): Remove. Update all users.
30438 (write_enum_defn): Remove. Update all users.
30439 (enum alloc_zone): Remove. Update all users.
30440 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
30441 * ggc-common.c (ggc_splay_alloc): Remove first argument.
30442 Update all callers.
30443 (struct ptr_data): Remove field TYPE. Update all users.
30444 (gt_pch_note_object): Remove argument TYPE. Update all users.
30445 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
30447 * ggc-none.c (ggc_alloc_typed_stat): Remove.
30448 (struct alloc_zone): Remove.
30449 (ggc_internal_alloc_zone_stat): Remove.
30450 (ggc_internal_cleared_alloc_zone_stat): Remove.
30451 * ggc-page.c (ggc_alloc_typed_stat): Remove.
30452 (ggc_pch_count_object): Remove last argument. Update all users.
30453 (ggc_pch_alloc_object): Remove last argument. Update all users.
30454 (struct alloc_zone): Remove.
30455 * ggc-zone.c: Remove.
30456 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
30457 (struct alloc_zone): Remove.
30458 (ggc_alloc_typed_stat): Remove.
30459 (ggc_alloc_typed): Remove.
30460 (ggc_splay_alloc): Remove first argument.
30461 (rtl_zone): Remove. Update all users.
30462 (tree_zone): Remove. Update all users.
30463 (tree_id_zone): Remove. Update all users.
30464 (ggc_internal_zone_alloc_stat): Remove. Update all users.
30465 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
30466 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
30467 * tree-ssanames.c: Remove references to zone allocator in comments.
30469 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
30471 * config/avr/avr.c (avr_out_fract): Make register numbers that
30472 might be outside of source operand signed.
30474 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
30476 * config/i386/constraints.md (Yf): New constraint.
30477 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
30478 of f constraint to conditionaly disable x87 register preferences.
30479 (*movdf_internal): Ditto.
30480 (*movsf_internal): Ditto.
30482 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
30484 PR inline-asm/55934
30485 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
30486 that have operands with impossible constraints.
30487 Add a FIXME for a speed-up opportunity.
30488 * lra-constraints.c (process_alt_operands): Verify that a class
30489 selected from constraints on asms is valid for the operand mode.
30490 (curr_insn_transform): Remove incorrect comment.
30492 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
30494 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
30495 TOC operand is a valid symbol ref in the constant pool.
30497 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30499 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
30501 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
30504 * config/avr/stdfix.h: New file.
30505 * t-avr (stdfix-gcc.h): New rule to build it.
30506 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
30508 2013-01-23 Kostya Serebryany <kcc@google.com>
30510 * config/darwin.h: remove dependency on
30511 CoreFoundation (asan on Mac OS).
30513 2013-01-23 Jakub Jelinek <jakub@redhat.com>
30516 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
30517 instead of cmpdi_operand for first comparison operand.
30518 Don't assert that comparison operands aren't both constants.
30520 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
30522 * doc/install.texi (Downloading the Source): Update references to
30523 downloading separate components.
30525 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
30527 * doc/extend.texi (__int128): Improve grammar.
30529 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
30532 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
30533 alternative to (o,r).
30534 (*movdi_internal_rex64): Remove (!o,n) alternative.
30535 (DImode immediate->memory splitter): Remove.
30536 (DImode immediate->memory peephole2): Remove.
30537 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
30538 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
30539 alternative to (!o,*r).
30540 (*movtf_internal_sse): New pattern.
30541 (*movxf_internal_rex64): New pattern.
30542 (*movxf_internal): Disable for TARGET_64BIT.
30543 (*movdf_internal_rex64): Remove (!o,F) alternative.
30545 2013-01-22 Jakub Jelinek <jakub@redhat.com>
30547 PR middle-end/56074
30548 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
30549 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
30550 * tree-vect-loop-manip.c (find_loop_location): Also ignore
30551 stmt locations where LOCATION_LOCUS of the stmt location is
30552 UNKNOWN_LOCATION or BUILTINS_LOCATION.
30555 * config/i386/i386.md (UNSPEC_STOS): New.
30556 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
30557 *strsetqi_1): Add UNSPEC_STOS.
30559 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
30562 * doc/invoke.texi: Remove left over -Wsynth example.
30564 2013-01-21 Jakub Jelinek <jakub@redhat.com>
30566 PR tree-optimization/56051
30567 * fold-const.c (fold_binary_loc): Don't fold
30568 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
30569 a narrowing conversion, or widening conversion from signed
30572 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
30574 PR rtl-optimization/56023
30575 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
30576 dependent on debug instruction.
30578 2013-01-21 Martin Jambor <mjambor@suse.cz>
30580 PR middle-end/56022
30581 * function.c (allocate_struct_function): Call
30582 invoke_set_current_function_hook earlier.
30584 2013-01-21 Jakub Jelinek <jakub@redhat.com>
30586 * reload1.c (init_reload): Only initialize reload_obstack
30587 during the first call.
30589 2013-01-21 Marek Polacek <polacek@redhat.com>
30591 * cfgloop.c (verify_loop_structure): Fix up grammar.
30593 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
30595 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
30596 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
30598 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
30601 * config/arm/marvell-pj4.md: Update copyright year.
30602 Fix up use of alu to alu_reg and simple_alu_imm.
30604 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
30606 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
30608 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
30611 * lra-constraints.c (curr_insn_transform): Don't reuse original
30612 insn for secondary memory move when memory mode should be different.
30614 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30616 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
30617 atomic_storedi_1): New patterns.
30619 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
30621 btver2 pipeline descriptions.
30622 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
30624 * config/i386/i386.md (btver2_decode): New type attributes.
30625 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
30627 * config/i386/btver2.md: New file describing btver2 pipelines.
30629 2013-01-19 Andrew Pinski <apinski@cavium.com>
30631 PR tree-optimization/52631
30632 * tree-ssa-sccvn (visit_use): Before looking up the original
30633 statement, try looking up the simplified expression.
30635 2013-01-19 Anthony Green <green@moxielogic.com>
30637 * config/moxie/moxie.c (moxie_expand_prologue): Set
30638 current_function_static_stack_size.
30640 2013-01-18 Jakub Jelinek <jakub@redhat.com>
30642 PR tree-optimization/56029
30643 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
30644 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
30646 2013-01-18 Sharad Singhai <singhai@google.com>
30648 PR tree-optimization/55995
30649 * dumpfile.c (dump_loc): Print location only if available.
30650 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
30652 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
30655 * lra-constraints.c (curr_insn_transform): Reuse original insn for
30656 secondary memory move.
30657 (inherit_reload_reg): Use rclass instead of cl for
30658 check_secondary_memory_needed_p.
30660 2013-01-18 Jakub Jelinek <jakub@redhat.com>
30662 PR middle-end/56015
30663 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
30664 the case where writing real complex part of target modifies op1.
30666 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
30668 * config/aarch64/aarch64-simd.md
30669 (aarch64_vcond_internal<mode>): Handle unordered cases.
30670 * config/aarch64/iterators.md (v_cmp_result): New.
30672 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
30673 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
30675 * config/arm/marvell-pj4.md: New file.
30676 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
30677 * config/arm/arm.md (generic_sched): Add marvell_pj4.
30678 (generic_vfp): Likewise.
30679 * config/arm/arm-cores.def: Add marvell-pj4.
30680 * config/arm/arm-tune.md: Regenerate.
30681 * config/arm/arm-tables.opt: Regenerate.
30682 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
30683 * doc/invoke.texi: Document marvell-pj4.
30685 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
30687 * config/aarch64/arm_neon.h: Map scalar types to standard types.
30689 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
30694 * var-tracking.c (negative_power_of_two_p): New.
30695 (global_get_addr_cache, local_get_addr_cache): New.
30696 (get_addr_from_global_cache, get_addr_from_local_cache): New.
30697 (vt_canonicalize_addr): Rewrite using the above. Adjust the
30699 (vt_stack_offset_p): Remove.
30700 (vt_canon_true_dep): Always canonicalize loc's address.
30701 (clobber_overlapping_mems): Make sure we have a MEM.
30702 (local_get_addr_clear_given_value): New.
30703 (val_reset): Clear local cached entries.
30704 (compute_bb_dataflow): Create and release the local cache.
30705 Disable duplicate MEMs clobbering.
30706 (emit_notes_in_bb): Clobber MEMs likewise.
30707 (vt_emit_notes): Create and release the local cache.
30708 (vt_initialize, vt_finalize): Create and release the global
30709 cache, respectively.
30710 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
30712 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
30714 PR libmudflap/53359
30715 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
30716 not found in the symtab.
30718 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
30721 PR rtl-optimization/55547
30722 PR rtl-optimization/53827
30725 * alias.c (offset_overlap_p): New, factored out of...
30726 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
30727 the conservative special case for symbolic constants. Don't
30728 adjust zero sizes on alignment.
30730 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
30732 PR rtl-optimization/52573
30733 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
30734 REG_UNUSED for the same register.
30736 2013-01-17 Richard Biener <rguenther@suse.de>
30737 Marek Polacek <polacek@redhat.com>
30739 PR rtl-optimization/55833
30740 * loop-unswitch.c (unswitch_loops): Move loop verification...
30741 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
30742 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
30743 Set it to true when we're removing a loop from hierarchy tree in
30744 an irreducible region.
30745 (fix_bb_placements): Adjust caller.
30746 (fix_loop_placements): Likewise.
30748 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
30750 * config/avr/builtins.def (DEF_BUILTIN): Factor out
30751 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
30752 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
30753 Remove ID. Adjust comments.
30754 * config/avr/avr-c.c (avr_builtin_name): Remove.
30755 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
30756 * config/avr/avr.c (avr_tolower): New static function.
30757 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
30758 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
30759 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
30762 2013-01-17 Jan Hubicka <jh@suse.cz>
30764 PR tree-optimization/55273
30765 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
30767 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
30770 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
30771 store through atomic_store<mode>_1.
30772 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
30774 2013-01-17 Martin Jambor <mjambor@suse.cz>
30776 PR tree-optimizations/55264
30777 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
30778 for virtual methods.
30779 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
30780 virtual methods before inlining is over.
30781 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
30783 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
30786 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
30788 PR rtl-optimization/56005
30789 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
30790 pending reads for prefetch.
30792 2013-01-16 Ian Bolton <ian.bolton@arm.com>
30794 * config/aarch64/aarch64.md
30795 (*cstoresi_neg_uxtw): New pattern.
30796 (*cmovsi_insn_uxtw): New pattern.
30797 (*<optab>si3_uxtw): New pattern.
30798 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
30799 (*<optab>si3_insn_uxtw): New pattern.
30800 (*bswapsi2_uxtw): New pattern.
30802 2013-01-16 Richard Biener <rguenther@suse.de>
30804 * tree-inline.c (tree_function_versioning): Remove set but
30805 never used variable.
30807 2013-01-16 Richard Biener <rguenther@suse.de>
30809 PR tree-optimization/55964
30810 * tree-flow.h (rename_variables_in_loop): Remove.
30811 (rename_variables_in_bb): Likewise.
30812 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
30813 (copy_loop_before): Adjust and delete update-ssa status.
30814 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
30815 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
30816 (rename_variables_in_loop): Remove.
30817 (slpeel_update_phis_for_duplicate_loop): Likewise.
30818 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
30819 use available cfg machinery instead of duplicating it.
30820 Update PHI nodes and perform poor-mans SSA update here.
30821 (slpeel_tree_peel_loop_to_edge): Adjust.
30823 2013-01-16 Richard Biener <rguenther@suse.de>
30825 PR tree-optimization/54767
30826 PR tree-optimization/53465
30827 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
30828 (vrp_visit_phi_node): For PHI arguments coming via backedges
30829 drop all symbolical range information.
30830 (execute_vrp): Compute backedges.
30832 2013-01-16 Richard Biener <rguenther@suse.de>
30834 * doc/install.texi: Update CLooG and ISL requirements to
30837 2013-01-16 Christian Bruel <christian.bruel@st.com>
30840 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
30841 (broken_move): Handle UNSPECV_SP_SWITCH_B.
30842 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
30844 2013-01-16 DJ Delorie <dj@redhat.com>
30846 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
30847 (UNSPECV_SP_SWITCH_E): New.
30848 (sp_switch_1): Change to an unspec.
30849 (sp_switch_2): Change to an unspec. Don't use post-inc when we
30852 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
30854 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
30855 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
30856 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
30857 (expand_mem_thread_fence): Ditto.
30858 (expand_mem_signal_fence): Ditto.
30859 (expand_atomic_load): Ditto.
30860 (expand_atomic_store): Ditto.
30862 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
30864 PR rtl-optimization/55547
30865 PR rtl-optimization/53827
30868 * alias.c (memrefs_conflict_p): Set sizes to negative after
30871 2013-01-15 Jakub Jelinek <jakub@redhat.com>
30874 * function.c (thread_prologue_and_epilogue_insns): Always
30875 add crtl->drap_reg to set_up_by_prologue.set, even if
30876 stack_realign_drap is false.
30878 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
30880 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
30881 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
30882 *call): Fix indention.
30884 2013-01-15 Tom de Vries <tom@codesourcery.com>
30887 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
30890 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
30892 PR rtl-optimization/55153
30893 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
30895 2013-01-15 Martin Jambor <mjambor@suse.cz>
30897 PR tree-optimization/55920
30898 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
30899 accesses as grp_to_be_debug_replaced.
30901 2013-01-15 Jakub Jelinek <jakub@redhat.com>
30903 PR tree-optimization/55920
30904 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
30905 there is non-useless type conversion needed from debug rhs to lhs,
30906 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
30908 2013-01-15 Joseph Myers <joseph@codesourcery.com>
30909 Mikael Pettersson <mikpe@it.uu.se>
30912 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
30914 (ASM_OUTPUT_CASE_LABEL): Remove.
30915 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
30916 * final.c (shorten_branches): Update alignment of labels before
30917 jump tables if CASE_VECTOR_SHORTEN_MODE.
30919 2013-01-15 Richard Biener <rguenther@suse.de>
30922 * system.h: Do not include gmp.h for building host tools.
30924 2013-01-15 Richard Biener <rguenther@suse.de>
30926 PR middle-end/55882
30927 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
30928 account for bitpos when computing alignment.
30930 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
30932 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
30933 (ix86_target_macros_internal): Likewise.
30935 * config/i386/i386.c (m_CORE2I7): Removed.
30936 (m_CORE_HASWELL): New macro.
30937 (m_CORE_ALL): Likewise.
30938 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
30939 (initial_ix86_arch_features): Likewise.
30940 (processor_target_table): Initializations for Core avx2.
30941 (cpu_names): New names "core-avx2".
30942 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
30943 PROCESSOR_CORE_HASWELL.
30944 (ix86_issue_rate): New case.
30945 (ia32_multipass_dfa_lookahead): Likewise.
30946 (ix86_sched_init_global): Likewise.
30948 * config/i386/i386.h (TARGET_HASWELL): New macro.
30949 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
30950 (processor_type): New PROCESSOR_HASWELL.
30952 2013-01-15 Jakub Jelinek <jakub@redhat.com>
30954 PR tree-optimization/55955
30955 * tree-vect-loop.c (vectorizable_reduction): Give up early on
30956 *SHIFT_EXPR and *ROTATE_EXPR codes.
30958 PR tree-optimization/48766
30959 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
30960 -ftrapv disable -fwrapv.
30962 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30965 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
30966 etc. to 1 and not to __flash.
30967 Use LL suffix for __INT24_MAX__ with -mint8.
30968 Use ULL suffix for __UINT24_MAX__ with -mint8.
30970 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30972 * config/avr/avr-arch.h
30973 (struct base_arch_s): Use typedef avr_arch_t instead.
30974 (struct arch_info_s): Use typedef avr_arch_info_t instead.
30975 (struct mcu_type_s): Use typedef avr_mcu_t instead.
30976 * config/avr/avr.c: Same.
30977 * config/avr/avr-devices.c: Same.
30978 * config/avr/driver-avr.c: Same.
30979 * config/avr/gen-avr-mmcu-texi.c: Same.
30980 * config/avr/avr-mcus.def: Adjust comment.
30982 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
30984 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
30985 * config/aarch64/iterators.md (VALLDI): New.
30987 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
30988 Andi Kleen <ak@linux.intel.com>
30991 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
30992 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
30995 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30997 * config/avr/avr-stdint.h: Remove trailing blanks.
30998 * config/avr/avr-log.h: Same.
30999 * config/avr/avr-arch.h: Same.
31000 * config/avr/avr-devices.c: Same.
31001 * config/avr/avr-dimode.md: Same.
31002 * config/avr/predicates.md: Same.
31003 * config/avr/avr-c.c: Same. And fix typo.
31005 * config/avr/avr-protos.h: Same. And:
31006 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
31007 (init_cumulative_args): Rename to avr_init_cumulative_args.
31008 (expand_prologue): Rename to avr_expand_prologue.
31009 (expand_epilogue): Rename to avr_expand_epilogue.
31010 (adjust_insn_length): Rename to avr_adjust_insn_length.
31011 (notice_update_cc): Rename to avr_notice_update_cc.
31012 (final_prescan_insn): Rename to avr_final_prescan_insn.
31013 * config/avr/avr.c: Same.
31014 * config/avr/avr.h: Same.
31015 * config/avr/avr.md: Remove trailing blanks.
31016 (prologue): Use avr_expand_prologue.
31017 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
31019 2013-01-14 Richard Biener <rguenther@suse.de>
31021 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
31022 verify_location, collect_subblocks): New functions.
31023 (verify_gimple_in_cfg): Verify that locations only reference
31024 BLOCKs in the functions BLOCK tree.
31026 2013-01-14 Richard Biener <rguenther@suse.de>
31028 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
31030 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
31032 (insert_out_of_ssa_copy_on_edge): Likewise.
31033 (rewrite_close_phi_out_of_ssa): Likewise.
31034 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
31036 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
31037 propagated constants.
31038 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
31041 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
31043 * config/avr/avr-modes.def: Add GPL copyright notice.
31045 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
31047 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
31048 MEMMODEL_MASK to determine memory model.
31049 (atomic_store<mode>): Ditto from operands[2].
31050 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
31052 2013-01-13 Jakub Jelinek <jakub@redhat.com>
31055 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
31056 (fold_gimple_assign): Don't call unshare_expr here.
31057 (fold_ctor_reference): Call unshare_expr.
31059 2013-01-13 Terry Guo <terry.guo@arm.com>
31061 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
31062 * doc/fragments.texi: Document MULTILIB_REUSE.
31063 * gcc.c (multilib_reuse): New internal spec.
31064 (set_multilib_dir): Also search multilib from multilib_reuse.
31065 * genmultilib (tmpmultilib3): Refactor code.
31066 (tmpmultilib4): Ditto.
31067 (multilib_reuse): New multilib argument.
31069 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
31071 * Makefile.in: Update copyright.
31073 2013-01-12 Tom de Vries <tom@codesourcery.com>
31075 PR middle-end/55890
31076 * calls.c (expand_call): Check if arg_nr is valid.
31078 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
31080 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
31081 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
31082 documentation. Add missing '__' in front of
31083 __builtin_ia32_packssdw256.
31085 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
31088 * config/s390/s390.c (s390_preferred_reload_class): Do not return
31089 NO_REGS for larl operands.
31090 (s390_reload_larl_operand): Use s390_load_address instead of
31093 2013-01-11 Richard Biener <rguenther@suse.de>
31095 * tree-cfg.c (verify_node_sharing_1): Split out from ...
31096 (verify_node_sharing): ... here.
31097 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
31099 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
31101 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
31102 Substitute TREECHECKING.
31103 * configure: Regenerate.
31104 * Makefile.in (TREECHECKING): New.
31106 2013-01-11 Richard Guenther <rguenther@suse.de>
31108 PR tree-optimization/44061
31109 * tree-vrp.c (extract_range_basic): Compute zero as
31110 value-range for __builtin_constant_p of function parameters.
31112 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
31114 Update copyright years.
31116 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
31118 PR rtl-optimization/55672
31119 * lra-eliminations.c (mark_not_eliminable): Permit addition with
31120 const to be eliminable.
31122 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
31124 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
31125 * configure: Regenerate.
31127 2013-01-10 Richard Biener <rguenther@suse.de>
31129 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
31131 2013-01-10 Richard Biener <rguenther@suse.de>
31134 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
31135 locations for virtual PHI arguments.
31136 (rewrite_update_phi_arguments): Likewise.
31138 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
31140 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
31143 2013-01-10 Jakub Jelinek <jakub@redhat.com>
31145 PR tree-optimization/55921
31146 * tree-complex.c (expand_complex_asm): New function.
31147 (expand_complex_operations_1): Call it for GIMPLE_ASM.
31149 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
31152 * config/s390/s390.c (s390_symref_operand_p)
31153 (s390_loadrelative_operand_p): Merge the two functions.
31154 (s390_check_qrst_address, print_operand_address): Add parameters
31155 to s390_loadrelative_operand_p invokation.
31156 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
31157 (s390_reload_larl_operand, s390_secondary_reload): Use
31158 s390_loadrelative_operand_p instead of s390_symref_operand_p.
31159 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
31161 2013-01-09 Mike Stump <mikestump@comcast.net>
31163 * dse.c (record_store): Remove unnecessary assert.
31165 2013-01-09 Jan Hubicka <jh@suse.cz>
31167 PR tree-optimization/55569
31168 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
31169 * cfgloop.h (scale_loop_profile): Likewise.
31171 2013-01-09 Jan Hubicka <jh@suse.cz>
31174 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
31176 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
31178 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
31180 PR middle-end/55114
31181 * expr.h (maybe_emit_group_store): Declare.
31182 * expr.c (maybe_emit_group_store): New function.
31183 * builtins.c (expand_builtin_int_roundingfn): Call it.
31184 (expand_builtin_int_roundingfn_2): Likewise.
31186 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
31188 PR rtl-optimization/55829
31189 * lra-constraints.c (match_reload): Add code for absent output.
31190 (curr_insn_transform): Add code for reloads of matched inputs
31193 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
31195 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
31196 attribute of movddup insn to DF.
31197 (*vec_interleave_lowv2df): Ditto.
31198 (vec_dupv2df): Ditto.
31200 2013-01-09 Jan Hubicka <jh@suse.cz>
31202 PR tree-optimiation/55875
31203 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
31204 EVERY_ITERATION parameter.
31205 (number_of_iterations_exit): Check if exit is executed every iteration.
31206 (idx_infer_loop_bounds): Similarly here.
31207 (n_of_executions_at_most): Simplify
31208 to only test for cases where statement is dominated by the
31209 particular bound; handle correctly the "postdominance" test.
31210 (scev_probably_wraps_p): Use max loop iterations info
31211 as a global bound first.
31213 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
31214 Nick Clifton <nickc@redhat.com>
31216 * config/v850/v850.md (cbranchsf4): New pattern.
31217 (cstoresf4): New pattern.
31218 (cbranchdf4): New pattern.
31219 (cstoredf4): New pattern.
31220 (movsicc): Disallow floating point comparisons.
31221 (cmpsf_le_insn): Fix order of operators.
31222 (cmpsf_lt_insn): Likewise.
31223 (cmpsf_eq_insn): Likewise.
31224 (cmpdf_le_insn): Likewise.
31225 (cmpdf_lt_insn): Likewise.
31226 (cmpdf_eq_insn): Likewise.
31227 (cmpsf_ge_insn): Use LE comparison.
31228 (cmpdf_ge_insn): Likewise.
31229 (cmpsf_gt_insn): Use LT comparison.
31230 (cmpdf_gt_insn): Likewise.
31231 (cmpsf_ne_insn): Delete pattern.
31232 (cmpdf_ne_insn): Delete pattern.
31233 * config/v850/v850.c (v850_gen_float_compare): Use
31234 gen_cmpdf_eq_insn for NE comparison.
31235 (v850_float_z_comparison_operator)
31236 (v850_float_nz_comparison_operator): Move from here ...
31237 * config/v850/predicates.md: ... to here. Move GT and GE
31238 comparisons into v850_float_z_comparison_operator.
31239 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
31241 (v850_float_nz_comparison_operator): Likewise.
31243 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31245 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
31246 with calls to gen_insvsi/gen_insvdi.
31248 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
31250 * config/i386/i386.c (initial_ix86_tune_features): Set up
31251 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
31253 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
31254 Jakub Jelinek <jakub@redhat.com>
31256 PR tree-optimization/48189
31257 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
31258 If nitercst is 0, don't predict the exit edge.
31260 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
31262 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
31263 in asm_fprintf with reg_names.
31264 (aarch64_print_operand_address): Likewise.
31265 (aarch64_return_addr): Likewise.
31266 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
31268 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31270 * config/pa/pa.h (VAL_U6_BITS_P): Define.
31271 (INT_U6_BITS): Likewise.
31272 * config/pa/predicates.md (uint6_operand): New predicate.
31273 (shift5_operand, shift6_operand): Likewise.
31274 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
31276 (lshrdi3): Use shift6_operand.
31277 (shrpsi4, shrpdi4): New insn patterns.
31278 (extzv): Delete expander.
31279 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
31280 predicates in unamed zero extract patterns. Tighten common constraint.
31281 (extv): Delete expander.
31282 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
31283 predicates in unamed sign extract patterns. Tighten common constraint.
31284 (insv): Delete expander.
31285 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
31286 predicates in unamed insert patterns. Tighten common constraint.
31287 Change uint32_operand predicate to uint6_operand predicate in unamed
31288 DImode pattern to insert constant values of type 1...1xxxx.
31290 2013-01-04 Jan Hubicka <jh@suse.cz>
31292 PR tree-optimization/55823
31293 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
31296 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31297 Uros Bizjak <ubizjak@gmail.com>
31299 PR rtl-optimization/55845
31300 * df-problems.c (can_move_insns_across): Stop scanning at
31301 volatile_insn_p source instruction or give up if
31302 across_from .. across_to range contains any volatile_insn_p
31305 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
31307 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
31308 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
31310 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
31311 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
31313 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31316 * asan.c (asan_clear_shadow): New function.
31317 (asan_emit_stack_protection): Use it.
31319 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
31321 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
31322 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
31323 with tab instead of space.
31325 2013-01-08 Nick Clifton <nickc@redhat.com>
31327 * config/rl78/rl78.c (rl78_expand_prologue): Always select
31328 register bank 0 at the start of an interrupt handler.
31329 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
31332 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
31334 * config/aarch64/aarch64-simd.md
31335 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
31336 (aarch64_simd_bsl): Likewise.
31337 (aarch64_vcond_internal<mode>): Likewise.
31338 (vcond<mode><mode>): Likewise.
31339 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
31340 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
31342 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
31344 * config/aarch64/aarch64-builtins.c
31345 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
31347 2013-01-08 Martin Jambor <mjambor@suse.cz>
31350 * tree-sra.c (analyze_access_subtree): Return true also after
31351 potentially creating a debug-only replacement.
31353 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31355 PR middle-end/55890
31356 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
31358 PR tree-optimization/54120
31359 * tree-vrp.c (range_fits_type_p): Don't allow
31360 src_precision < precision from signed vr to unsigned_p
31361 if vr->min or vr->max is negative.
31362 (simplify_float_conversion_using_ranges): Test can_float_p
31363 against CODE_FOR_nothing.
31365 2013-01-08 Jakub Jelinek <jakub@redhat.com>
31366 Richard Biener <rguenther@suse.de>
31368 PR middle-end/55851
31369 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
31370 types instead of just INTEGER_TYPE types.
31372 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
31374 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
31377 2013-01-07 Steve Ellcey <sellcey@mips.com>
31380 * config/mips/mips.opt: Change mad to mmad to match documentation.
31382 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31385 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
31386 .progmemx.data now.
31388 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31391 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
31392 (avr_addrspace_t): Add .section_name field.
31393 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
31395 (avr_addrspace): Same. Initialize .section_name. Remove last
31396 NULL entry. Put __memx into .progmemx.data.
31397 (progmem_section_prefix): Remove.
31398 (avr_asm_init_sections): No need to initialize progmem_section.
31399 (avr_asm_named_section): Use avr_addrspace[].section_name to get
31400 section name prefix.
31401 (avr_asm_select_section): Ditto. And use get_unnamed_section to
31402 retrieve the progmem section.
31403 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
31404 boundary to run over avr_addrspace[].
31405 (avr_register_target_pragmas): Ditto.
31407 2013-01-06 Jakub Jelinek <jakub@redhat.com>
31409 * varasm.c (output_constant_def_contents): For asan_protect_global
31410 protected strings, adjust DECL_ALIGN if needed, before testing for
31412 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
31413 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
31415 (output_object_block): For asan protected decls, emit asan padding
31416 after their contents.
31417 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
31418 (asan_finish_file): Test it here instead.
31420 2013-01-07 Nick Clifton <nickc@redhat.com>
31421 Matthias Klose <doko@debian.org>
31422 Doug Kwan <dougkwan@google.com>
31423 H.J. Lu <hongjiu.lu@intel.com>
31426 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
31428 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
31430 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
31432 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
31434 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
31436 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31439 * doc/install.texi (Cross-Compiler-Specific Options): Document
31442 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
31444 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
31445 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
31446 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
31447 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
31448 vqmovun_high_s64): Fix source operand number and update copyright.
31450 2013-01-07 Richard Biener <rguenther@suse.de>
31452 PR middle-end/55890
31453 * gimple.h (gimple_call_builtin_p): New overload.
31454 * gimple.c (validate_call): New function.
31455 (gimple_call_builtin_p): Likewise.
31456 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
31457 Use gimple_call_builtin_p.
31458 (find_func_clobbers): Likewise.
31459 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
31460 (strlen_optimize_stmt): Likewise.
31462 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
31464 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
31465 (vld1q_dup_*): Likewise.
31466 (vld1_*): Likewise.
31467 (vld1q_*): Likewise.
31468 (vld1_lane_*): Likewise.
31469 (vld1q_lane_*): Likewise.
31471 2013-01-07 Richard Biener <rguenther@suse.de>
31473 * lto-streamer.h (LTO_minor_version): Bump to 2.
31475 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
31477 * config/aarch64/aarch64-protos.h
31478 (aarch64_const_double_zero_rtx_p): Rename to...
31479 (aarch64_float_const_zero_rtx_p): ...this.
31480 (aarch64_float_const_representable_p): New.
31481 (aarch64_output_simd_mov_immediate): Likewise.
31482 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
31483 move immediate case.
31484 * config/aarch64/aarch64.c
31485 (aarch64_const_double_zero_rtx_p): Rename to...
31486 (aarch64_float_const_zero_rtx_p): ...this.
31487 (aarch64_print_operand): Allow printing of new constants.
31488 (aarch64_valid_floating_const): New.
31489 (aarch64_legitimate_constant_p): Check for valid floating-point
31491 (aarch64_simd_valid_immediate): Likewise.
31492 (aarch64_vect_float_const_representable_p): New.
31493 (aarch64_float_const_representable_p): Likewise.
31494 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
31495 (aarch64_output_simd_mov_immediate): New.
31496 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
31497 (*movdf_aarch64): Likewise.
31498 * config/aarch64/constraints.md (Ufc): New.
31499 (Y): call aarch64_float_const_zero_rtx.
31500 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
31502 2013-01-07 Richard Biener <rguenther@suse.de>
31504 PR tree-optimization/55888
31505 PR tree-optimization/55862
31506 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
31507 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
31508 not if it is contained therein.
31510 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31512 * config/avr/t-avr: Typo.
31514 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
31517 * config/avr/t-avr: Don't automatically rebuild
31518 $(srcdir)/config/avr/t-multilib
31519 $(srcdir)/config/avr/avr-tables.opt
31520 $(srcdir)/doc/avr-mmcu.texi
31521 (avr-mcus): New phony target to build them on request.
31522 (s-avr-mlib, s-avr-mmcu-texi): Remove.
31523 * avr/avr-mcus.def: Adjust comments.
31525 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
31527 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
31529 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
31531 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
31533 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
31535 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
31537 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
31539 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
31540 to generate profiling.
31541 * config/rs6000/aix64.h (LIB_SPEC): Same.
31543 2013-01-04 Andrew Pinski <apinski@cavium.com>
31545 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
31547 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
31549 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
31551 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
31553 (ix86_expand_move): Ditto.
31554 (ix86_zero_extend_to_Pmode): Ditto.
31555 (ix86_expand_call): Ditto.
31556 (ix86_expand_special_args_builtin): Ditto.
31557 (ix86_expand_builtin): Ditto.
31559 2013-01-04 Richard Biener <rguenther@suse.de>
31561 PR tree-optimization/55862
31562 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
31563 translating them through PHI nodes.
31565 2013-01-04 Martin Jambor <mjambor@suse.cz>
31567 PR tree-optimization/55755
31568 * tree-sra.c (sra_modify_assign): Do not check that an access has no
31569 children when trying to avoid producing a VIEW_CONVERT_EXPR.
31571 2013-01-04 Marek Polacek <polacek@redhat.com>
31573 PR middle-end/55859
31574 * opts.c (default_options_optimization): Clarify error message.
31576 2013-01-04 Richard Biener <rguenther@suse.de>
31578 PR middle-end/55863
31579 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
31582 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31585 * config/pa/pa.md (movsi): Revert previous change.
31586 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
31589 2013-01-03 Richard Henderson <rth@redhat.com>
31591 * config/i386/i386.c (ix86_expand_move): Always assign to op1
31592 after eliminating TLS symbols.
31594 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
31597 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
31598 * graphite-poly.c (debug_gmp_value): Likewise.
31600 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
31603 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
31604 selected code model, define __code_mode_small__, __code_model_medium__,
31605 __code_model_large__, __code_model_32__ or __code_model_kernel__.
31606 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
31607 xchg temporary register with %k. Declare temporary register as
31609 [__x86_64__]: For medium and large code models, preserve %rbx register.
31611 2013-01-03 Richard Biener <rguenther@suse.de>
31613 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
31614 (dump_subscript): Adjust.
31615 (finalize_ddr_dependent): Do not dump redundant info.
31616 (analyze_siv_subscript): Adjust.
31617 (subscript_dependence_tester): Likewise.
31618 (compute_affine_dependence): Likewise.
31620 2013-01-03 Richard Biener <rguenther@suse.de>
31623 2013-01-03 Richard Biener <rguenther@suse.de>
31625 PR tree-optimization/55857
31626 * tree-vect-stmts.c (vectorizable_load): Do not setup
31627 re-alignment for invariant loads.
31629 2013-01-02 Richard Biener <rguenther@suse.de>
31631 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
31632 invariant load do not generate a vector load from the scalar location.
31634 2013-01-03 Richard Biener <rguenther@suse.de>
31636 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
31637 for not vectorizing.
31638 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
31639 not build INDIRECT_REFs, call get_name once only.
31640 (vect_create_data_ref_ptr): Likewise. Dump base object kind
31641 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
31643 2013-01-03 Richard Biener <rguenther@suse.de>
31645 PR tree-optimization/55857
31646 * tree-vect-stmts.c (vectorizable_load): Do not setup
31647 re-alignment for invariant loads.
31649 2013-01-03 Richard Biener <rguenther@suse.de>
31652 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
31653 prefer a built-in decl.
31655 2013-01-03 Jakub Jelinek <jakub@redhat.com>
31657 * gcc.c (process_command): Update copyright notice dates.
31658 * gcov.c (print_version): Likewise.
31659 * gcov-dump.c (print_version): Likewise.
31661 PR rtl-optimization/55838
31662 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
31663 iv0.step, iv1.step and step.
31665 2013-01-03 Jakub Jelinek <jakub@redhat.com>
31666 Marc Glisse <marc.glisse@inria.fr>
31668 PR tree-optimization/55832
31669 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
31670 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
31671 integer_{one,zero}_node.
31673 2013-01-03 Jakub Jelinek <jakub@redhat.com>
31676 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
31677 * var-tracking.c (reverse_op): Don't add reverse ops to
31678 VALUEs that have already
31679 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
31681 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
31683 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
31685 2013-01-02 Teresa Johnson <tejohnson@google.com>
31687 * dumpfile.c (dump_loc): Print filename with location.
31688 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
31689 new location_t parameter to emit complete unroll message with
31690 new dump framework.
31691 (canonicalize_loop_induction_variables): Compute loops location
31692 and pass to try_unroll_loop_completely.
31693 * loop-unroll.c (report_unroll_peel): New function.
31694 (peel_loops_completely): Use new dump format with location
31695 for main dumpfile message, and invoke report_unroll_peel on success.
31696 (decide_unrolling_and_peeling): Ditto.
31697 (decide_peel_once_rolling): Remove old dumpfile message subsumed
31698 by report_unroll_peel.
31699 (decide_peel_completely): Ditto.
31700 (decide_unroll_constant_iterations): Ditto.
31701 (decide_unroll_runtime_iterations): Ditto.
31702 (decide_peel_simple): Ditto.
31703 (decide_unroll_stupid): Ditto.
31704 * cfgloop.c (get_loop_location): New function.
31705 * cfgloop.h (get_loop_location): Declare.
31707 2013-01-02 Sriraman Tallam <tmsriram@google.com>
31709 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
31712 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31714 PR middle-end/55198
31715 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
31716 BLKmode objects when EXPAND_MEMORY is specified.
31718 2013-01-02 Sriraman Tallam <tmsriram@google.com>
31720 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
31722 (fold_builtin_cpu): Do not share cpu model decls across statements.
31724 2013-01-02 Jason Merrill <jason@redhat.com>
31727 * tree.c (build_array_type_1): Revert earlier change.
31729 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
31731 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
31733 * config/aarch64/aarch64-tune.md: Re-generate.
31735 2013-01-02 Richard Biener <rguenther@suse.de>
31737 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
31738 invariant load do not generate a vector load from the scalar location.
31740 2013-01-02 Richard Biener <rguenther@suse.de>
31743 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
31744 * configure: Regenerate.
31746 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
31748 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
31749 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
31750 (expand_builtin_int_roundingfn_2): Keep the original target around
31751 for the fallback case.
31753 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
31755 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
31756 to be clear for sign changes.
31758 2013-01-01 Jan Hubicka <jh@suse.cz>
31760 * ipa-inline-analysis.c: Fix formatting.
31762 2013-01-01 Jakub Jelinek <jakub@redhat.com>
31764 PR tree-optimization/55831
31765 * tree-vect-loop.c (get_initial_def_for_induction): Use
31766 gsi_after_labels instead of gsi_start_bb.
31768 Copyright (C) 2013 Free Software Foundation, Inc.
31770 Copying and distribution of this file, with or without modification,
31771 are permitted in any medium without royalty provided the copyright
31772 notice and this notice are preserved.