1 2014-04-25 Tom de Vries <tom@codesourcery.com>
3 * expr.c (clobber_reg_mode): New function.
4 * expr.h (clobber_reg): New function.
6 2014-04-25 Tom de Vries <tom@codesourcery.com>
8 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
11 2014-04-25 Radovan Obradovic <robradovic@mips.com>
12 Tom de Vries <tom@codesourcery.com>
14 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
16 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
17 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
18 new argument to find_all_hard_reg_sets call.
20 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
23 Use HOST_WIDE_INT_C for mask literal.
24 (aarch_rev16_shleft_mask_imm_p): Likewise.
26 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
29 * config/sparc/sparc.md (ashlsi3_extend): Delete.
31 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
34 * config/i386/i386-c.c (ix86_target_macros): Define
35 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
37 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
39 * configure.ac (tga_func): Remove.
40 (LIB_TLS_SPEC): Remove.
41 * configure: Regenerate.
42 * config.in: Regenerate.
43 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
45 2014-04-25 Richard Biener <rguenther@suse.de>
48 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
49 call stmt use/clobber sets during stmt walk instead of
50 walking the possibly incomplete set of caller edges.
52 2014-04-25 Richard Biener <rguenther@suse.de>
55 * passes.c (apply_ipa_transforms): Inline into only caller ...
56 (execute_one_pass): ... here. Properly bring in function
57 bodies for nodes we want to apply IPA transforms to.
59 2014-04-24 Cong Hou <congh@google.com>
61 PR tree-optimization/60896
62 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
63 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
64 (vect_mark_pattern_stmts): Set the def type of all statements in
65 PATTERN_DEF_SEQ as vect_internal_def.
67 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
69 * doc/extend.texi (PowerPC Built-in Functions): Document new
70 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
72 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
74 * config/rs6000/predicates.md (const_0_to_3_operand): New
75 predicate to match 0..3 integer constants.
77 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
78 to support adding miscellaneous builtin functions.
79 (BU_DFP_MISC_2): Likewise.
80 (BU_P7_MISC_1): Likewise.
81 (BU_P7_MISC_2): Likewise.
82 (BU_P8V_MISC_3): Likewise.
83 (BU_MISC_1): Likewise.
84 (BU_MISC_2): Likewise.
85 (DIVWE): Add extended divide builtin functions.
93 (DXEX): Add decimal floating-point builtin functions.
105 (CDTBCD): Add new BCD builtin functions.
109 (BCDADD_LT): Likewise.
110 (BCDADD_EQ): Likewise.
111 (BCDADD_GT): Likewise.
112 (BCDADD_OV): Likewise.
114 (BCDSUB_LT): Likewise.
115 (BCDSUB_EQ): Likewise.
116 (BCDSUB_GT): Likewise.
117 (BCDSUB_OV): Likewise.
118 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
119 (UNPACK_TD): Likewise.
121 (UNPACK_TF): Likewise.
122 (UNPACK_TF_0): Likewise.
123 (UNPACK_TF_1): Likewise.
124 (PACK_V1TI): Likewise.
125 (UNPACK_V1TI): Likewise.
127 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
128 support for decimal floating point builtin functions.
129 (rs6000_expand_ternop_builtin): Add checks for the new builtin
130 functions that take constant arguments.
131 (rs6000_invalid_builtin): Add decimal floating point builtin
133 (rs6000_init_builtins): Setup long double, _Decimal64, and
134 _Decimal128 types for new builtin functions.
135 (builtin_function_type): Set the unsigned flags appropriately for
136 the new builtin functions.
137 (rs6000_opt_masks): Add support for decimal floating point builtin
140 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
141 floating point builtin functions.
142 (RS6000_BTM_COMMON): Likewise.
143 (RS6000_BTI_long_double): Likewise.
144 (RS6000_BTI_dfloat64): Likewise.
145 (RS6000_BTI_dfloat128): Likewise.
146 (long_double_type_internal_node): Likewise.
147 (dfloat64_type_internal_node): Likewise.
148 (dfloat128_type_internal_node): Likewise.
150 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
151 2.07 bcd arithmetic instructions.
152 (UNSPEC_BCDSUB): Likewise.
153 (UNSPEC_BCD_OVERFLOW): Likewise.
154 (UNSPEC_BCD_ADD_SUB): Likewise.
155 (bcd_add_sub): Likewise.
156 (BCD_TEST): Likewise.
157 (bcd<bcd_add_sub>): Likewise.
158 (bcd<bcd_add_sub>_test): Likewise.
159 (bcd<bcd_add_sub>_test2): Likewise.
160 (bcd<bcd_add_sub>_<code>): Likewise.
161 (peephole2 for combined bcd ops): Likewise.
163 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
164 decimal floating point builtin functions.
165 (UNSPEC_DENBCD): Likewise.
166 (UNSPEC_DXEX): Likewise.
167 (UNSPEC_DIEX): Likewise.
168 (UNSPEC_DSCLI): Likewise.
169 (UNSPEC_DSCRI): Likewise.
170 (D64_D128): Likewise.
171 (dfp_suffix): Likewise.
172 (dfp_ddedpd_<mode>): Likewise.
173 (dfp_denbcd_<mode>): Likewise.
174 (dfp_dxex_<mode>): Likewise.
175 (dfp_diex_<mode>): Likewise.
176 (dfp_dscli_<mode>): Likewise.
177 (dfp_dscri_<mode>): Likewise.
179 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
181 (UNSPEC_CDTBCD): Likewise.
182 (UNSPEC_CBCDTD): Likewise.
183 (UNSPEC_DIVE): Add support for new extended divide builtin
185 (UNSPEC_DIVEO): Likewise.
186 (UNSPEC_DIVEU): Likewise.
187 (UNSPEC_DIVEUO): Likewise.
188 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
189 pack/unpack 128-bit types.
190 (UNSPEC_PACK_128BIT): Likewise.
191 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
192 (udiv<mode>3): Use idiv_ldiv mode attribute.
193 (div<mode>3): Likewise.
194 (addg6s): Add new BCD builtin functions.
197 (UNSPEC_DIV_EXTEND): Add support for new extended divide
199 (div_extend): Likewise.
200 (div<div_extend>_<mode>"): Likewise.
201 (FP128_64): Add support for new builtin functions to pack/unpack
203 (unpack<mode>): Likewise.
204 (unpacktf_0): Likewise.
205 (unpacktf_1): Likewise.
206 (unpack<mode>_dm): Likewise.
207 (unpack<mode>_nodm): Likewise.
208 (pack<mode>): Likewise.
209 (unpackv1ti): Likewise.
210 (packv1ti): Likewise.
212 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
214 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
217 2014-04-24 Jakub Jelinek <jakub@redhat.com>
219 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
220 * gimplify.c (omp_is_private): Change last argument's type to int.
221 Only diagnose lastprivate if the simd argument is 1, only diagnose
222 linear if the simd argument is 2.
223 (gimplify_omp_for): Adjust omp_is_private callers. When adding
224 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
225 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
226 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
227 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
228 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
229 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
230 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
231 * tree-nested.c (convert_nonlocal_omp_clauses,
232 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
234 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
237 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
240 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
242 * flag-types.h (enum ivar_visibility): Add.
244 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
246 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
249 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
251 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
253 2014-04-24 Radovan Obradovic <robradovic@mips.com>
254 Tom de Vries <tom@codesourcery.com>
256 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
257 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
259 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
260 * emit-rtl.c (try_split): Same.
262 2014-04-24 Radovan Obradovic <robradovic@mips.com>
263 Tom de Vries <tom@codesourcery.com>
265 * common.opt (fuse-caller-save): New option.
267 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
269 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
272 2014-04-24 Richard Biener <rguenther@suse.de>
274 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
275 during TER and instead use the sepops interface for expanding
276 non-GIMPLE_SINGLE_RHS.
278 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
280 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
281 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
283 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
285 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
286 assembler 64-bit option.
287 * configure: Regenerate.
289 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
291 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
292 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
293 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
294 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
295 (TARGET_CRYPTO): Take TARGET_SIMD into account.
297 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
299 * config/aarch64/aarch64-builtins.c
300 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
301 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
302 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
303 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
305 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
306 (Vrevsuff): New mode attribute.
308 2014-04-24 Terry Guo <terry.guo@arm.com>
310 * config/arm/arm.h (machine_function): Define variable
311 after_arm_reorg here.
312 * config/arm/arm.c (after_arm_reorg): Remove the definition.
313 (arm_split_constant): Update the way to access variable
316 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
318 2014-04-23 Tom de Vries <tom@codesourcery.com>
320 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
322 2014-04-23 David Malcolm <dmalcolm@redhat.com>
324 * is-a.h: Update comments to reflect the following changes to the
325 "pointerness" of the API, making the template parameter match the
326 return type, allowing use of is-a.h with typedefs of pointers.
327 (is_a_helper::cast): Return a T rather then a pointer to a T, so
328 that the return type matches the parameter to the is_a_helper.
330 (dyn_cast): Likewise.
332 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
333 pointer from the is-a.h API.
335 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
336 (is_a_helper <cgraph_node *>::test): ...this, matching change to
338 (is_a_helper <varpool_node>::test): Likewise, convert to...
339 (is_a_helper <varpool_node *>::test): ...this.
341 (varpool_first_variable): Update for removal of implicit pointer
343 (varpool_next_variable): Likewise.
344 (varpool_first_static_initializer): Likewise.
345 (varpool_next_static_initializer): Likewise.
346 (varpool_first_defined_variable): Likewise.
347 (varpool_next_defined_variable): Likewise.
348 (cgraph_first_defined_function): Likewise.
349 (cgraph_next_defined_function): Likewise.
350 (cgraph_first_function): Likewise.
351 (cgraph_next_function): Likewise.
352 (cgraph_first_function_with_gimple_body): Likewise.
353 (cgraph_next_function_with_gimple_body): Likewise.
354 (cgraph_alias_target): Likewise.
355 (varpool_alias_target): Likewise.
356 (cgraph_function_or_thunk_node): Likewise.
357 (varpool_variable_node): Likewise.
358 (symtab_real_symbol_p): Likewise.
359 * cgraphunit.c (referred_to_p): Likewise.
360 (analyze_functions): Likewise.
361 (handle_alias_pairs): Likewise.
362 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
363 * gimple-ssa.h (gimple_vuse_op): Likewise.
364 (gimple_vdef_op): Likewise.
365 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
366 * gimple.c (gimple_build_asm_1): Likewise.
367 (gimple_build_try): Likewise.
368 (gimple_build_resx): Likewise.
369 (gimple_build_eh_dispatch): Likewise.
370 (gimple_build_omp_for): Likewise.
371 (gimple_omp_for_set_clauses): Likewise.
373 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
374 (is_a_helper <gimple_statement_asm *>::test): ...this.
375 (is_a_helper <gimple_statement_bind>::test): Convert to...
376 (is_a_helper <gimple_statement_bind *>::test): ...this.
377 (is_a_helper <gimple_statement_call>::test): Convert to...
378 (is_a_helper <gimple_statement_call *>::test): ...this.
379 (is_a_helper <gimple_statement_catch>::test): Convert to...
380 (is_a_helper <gimple_statement_catch *>::test): ...this.
381 (is_a_helper <gimple_statement_resx>::test): Convert to...
382 (is_a_helper <gimple_statement_resx *>::test): ...this.
383 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
384 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
385 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
386 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
387 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
388 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
389 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
390 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
391 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
392 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
393 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
394 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
395 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
396 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
397 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
398 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
399 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
400 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
401 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
402 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
403 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
404 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
405 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
406 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
407 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
408 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
409 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
410 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
411 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
412 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
413 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
414 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
415 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
416 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
417 (is_a_helper <gimple_statement_phi>::test): Convert to...
418 (is_a_helper <gimple_statement_phi *>::test): ...this.
419 (is_a_helper <gimple_statement_transaction>::test): Convert to...
420 (is_a_helper <gimple_statement_transaction *>::test): ...this.
421 (is_a_helper <gimple_statement_try>::test): Convert to...
422 (is_a_helper <gimple_statement_try *>::test): ...this.
423 (is_a_helper <gimple_statement_wce>::test): Convert to...
424 (is_a_helper <gimple_statement_wce *>::test): ...this.
425 (is_a_helper <const gimple_statement_asm>::test): Convert to...
426 (is_a_helper <const gimple_statement_asm *>::test): ...this.
427 (is_a_helper <const gimple_statement_bind>::test): Convert to...
428 (is_a_helper <const gimple_statement_bind *>::test): ...this.
429 (is_a_helper <const gimple_statement_call>::test): Convert to...
430 (is_a_helper <const gimple_statement_call *>::test): ...this.
431 (is_a_helper <const gimple_statement_catch>::test): Convert to...
432 (is_a_helper <const gimple_statement_catch *>::test): ...this.
433 (is_a_helper <const gimple_statement_resx>::test): Convert to...
434 (is_a_helper <const gimple_statement_resx *>::test): ...this.
435 (is_a_helper <const gimple_statement_eh_dispatch>::test):
437 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
439 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
441 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
442 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
444 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
446 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
448 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
450 (is_a_helper <const gimple_statement_omp_return>::test): Convert
452 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
453 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
455 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
456 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
458 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
459 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
460 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
461 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
463 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
464 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
466 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
467 (is_a_helper <const gimple_statement_omp_target>::test): Convert
469 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
470 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
472 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
473 (is_a_helper <const gimple_statement_omp_single>::test): Convert
475 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
476 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
478 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
479 (is_a_helper <const gimple_statement_omp_task>::test): Convert
481 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
482 (is_a_helper <const gimple_statement_phi>::test): Convert to...
483 (is_a_helper <const gimple_statement_phi *>::test): ...this.
484 (is_a_helper <const gimple_statement_transaction>::test): Convert
486 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
487 (is_a_helper <const gimple_statement_with_ops>::test): Convert
489 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
490 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
491 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
492 (is_a_helper <const gimple_statement_with_memory_ops>::test):
494 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
496 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
498 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
500 (gimple_use_ops): Update for removal of implicit pointer from the
502 (gimple_set_use_ops): Likewise.
503 (gimple_vuse): Likewise.
504 (gimple_vdef): Likewise.
505 (gimple_vuse_ptr): Likewise.
506 (gimple_vdef_ptr): Likewise.
507 (gimple_set_vuse): Likewise.
508 (gimple_set_vdef): Likewise.
509 (gimple_omp_return_set_lhs): Likewise.
510 (gimple_omp_return_lhs): Likewise.
511 (gimple_omp_return_lhs_ptr): Likewise.
512 (gimple_call_fntype): Likewise.
513 (gimple_call_set_fntype): Likewise.
514 (gimple_call_set_internal_fn): Likewise.
515 (gimple_call_use_set): Likewise.
516 (gimple_call_clobber_set): Likewise.
517 (gimple_bind_vars): Likewise.
518 (gimple_bind_set_vars): Likewise.
519 (gimple_bind_body_ptr): Likewise.
520 (gimple_bind_set_body): Likewise.
521 (gimple_bind_add_stmt): Likewise.
522 (gimple_bind_block): Likewise.
523 (gimple_bind_set_block): Likewise.
524 (gimple_asm_ninputs): Likewise.
525 (gimple_asm_noutputs): Likewise.
526 (gimple_asm_nclobbers): Likewise.
527 (gimple_asm_nlabels): Likewise.
528 (gimple_asm_input_op): Likewise.
529 (gimple_asm_input_op_ptr): Likewise.
530 (gimple_asm_output_op): Likewise.
531 (gimple_asm_output_op_ptr): Likewise.
532 (gimple_asm_set_output_op): Likewise.
533 (gimple_asm_clobber_op): Likewise.
534 (gimple_asm_set_clobber_op): Likewise.
535 (gimple_asm_label_op): Likewise.
536 (gimple_asm_set_label_op): Likewise.
537 (gimple_asm_string): Likewise.
538 (gimple_catch_types): Likewise.
539 (gimple_catch_types_ptr): Likewise.
540 (gimple_catch_handler_ptr): Likewise.
541 (gimple_catch_set_types): Likewise.
542 (gimple_catch_set_handler): Likewise.
543 (gimple_eh_filter_types): Likewise.
544 (gimple_eh_filter_types_ptr): Likewise.
545 (gimple_eh_filter_failure_ptr): Likewise.
546 (gimple_eh_filter_set_types): Likewise.
547 (gimple_eh_filter_set_failure): Likewise.
548 (gimple_eh_must_not_throw_fndecl): Likewise.
549 (gimple_eh_must_not_throw_set_fndecl): Likewise.
550 (gimple_eh_else_n_body_ptr): Likewise.
551 (gimple_eh_else_e_body_ptr): Likewise.
552 (gimple_eh_else_set_n_body): Likewise.
553 (gimple_eh_else_set_e_body): Likewise.
554 (gimple_try_eval_ptr): Likewise.
555 (gimple_try_cleanup_ptr): Likewise.
556 (gimple_try_set_eval): Likewise.
557 (gimple_try_set_cleanup): Likewise.
558 (gimple_wce_cleanup_ptr): Likewise.
559 (gimple_wce_set_cleanup): Likewise.
560 (gimple_phi_capacity): Likewise.
561 (gimple_phi_num_args): Likewise.
562 (gimple_phi_result): Likewise.
563 (gimple_phi_result_ptr): Likewise.
564 (gimple_phi_set_result): Likewise.
565 (gimple_phi_arg): Likewise.
566 (gimple_phi_set_arg): Likewise.
567 (gimple_resx_region): Likewise.
568 (gimple_resx_set_region): Likewise.
569 (gimple_eh_dispatch_region): Likewise.
570 (gimple_eh_dispatch_set_region): Likewise.
571 (gimple_omp_critical_name): Likewise.
572 (gimple_omp_critical_name_ptr): Likewise.
573 (gimple_omp_critical_set_name): Likewise.
574 (gimple_omp_for_clauses): Likewise.
575 (gimple_omp_for_clauses_ptr): Likewise.
576 (gimple_omp_for_set_clauses): Likewise.
577 (gimple_omp_for_collapse): Likewise.
578 (gimple_omp_for_index): Likewise.
579 (gimple_omp_for_index_ptr): Likewise.
580 (gimple_omp_for_set_index): Likewise.
581 (gimple_omp_for_initial): Likewise.
582 (gimple_omp_for_initial_ptr): Likewise.
583 (gimple_omp_for_set_initial): Likewise.
584 (gimple_omp_for_final): Likewise.
585 (gimple_omp_for_final_ptr): Likewise.
586 (gimple_omp_for_set_final): Likewise.
587 (gimple_omp_for_incr): Likewise.
588 (gimple_omp_for_incr_ptr): Likewise.
589 (gimple_omp_for_set_incr): Likewise.
590 (gimple_omp_for_pre_body_ptr): Likewise.
591 (gimple_omp_for_set_pre_body): Likewise.
592 (gimple_omp_parallel_clauses): Likewise.
593 (gimple_omp_parallel_clauses_ptr): Likewise.
594 (gimple_omp_parallel_set_clauses): Likewise.
595 (gimple_omp_parallel_child_fn): Likewise.
596 (gimple_omp_parallel_child_fn_ptr): Likewise.
597 (gimple_omp_parallel_set_child_fn): Likewise.
598 (gimple_omp_parallel_data_arg): Likewise.
599 (gimple_omp_parallel_data_arg_ptr): Likewise.
600 (gimple_omp_parallel_set_data_arg): Likewise.
601 (gimple_omp_task_clauses): Likewise.
602 (gimple_omp_task_clauses_ptr): Likewise.
603 (gimple_omp_task_set_clauses): Likewise.
604 (gimple_omp_task_child_fn): Likewise.
605 (gimple_omp_task_child_fn_ptr): Likewise.
606 (gimple_omp_task_set_child_fn): Likewise.
607 (gimple_omp_task_data_arg): Likewise.
608 (gimple_omp_task_data_arg_ptr): Likewise.
609 (gimple_omp_task_set_data_arg): Likewise.
610 (gimple_omp_taskreg_clauses): Likewise.
611 (gimple_omp_taskreg_clauses_ptr): Likewise.
612 (gimple_omp_taskreg_set_clauses): Likewise.
613 (gimple_omp_taskreg_child_fn): Likewise.
614 (gimple_omp_taskreg_child_fn_ptr): Likewise.
615 (gimple_omp_taskreg_set_child_fn): Likewise.
616 (gimple_omp_taskreg_data_arg): Likewise.
617 (gimple_omp_taskreg_data_arg_ptr): Likewise.
618 (gimple_omp_taskreg_set_data_arg): Likewise.
619 (gimple_omp_task_copy_fn): Likewise.
620 (gimple_omp_task_copy_fn_ptr): Likewise.
621 (gimple_omp_task_set_copy_fn): Likewise.
622 (gimple_omp_task_arg_size): Likewise.
623 (gimple_omp_task_arg_size_ptr): Likewise.
624 (gimple_omp_task_set_arg_size): Likewise.
625 (gimple_omp_task_arg_align): Likewise.
626 (gimple_omp_task_arg_align_ptr): Likewise.
627 (gimple_omp_task_set_arg_align): Likewise.
628 (gimple_omp_single_clauses): Likewise.
629 (gimple_omp_single_clauses_ptr): Likewise.
630 (gimple_omp_single_set_clauses): Likewise.
631 (gimple_omp_target_clauses): Likewise.
632 (gimple_omp_target_clauses_ptr): Likewise.
633 (gimple_omp_target_set_clauses): Likewise.
634 (gimple_omp_target_child_fn): Likewise.
635 (gimple_omp_target_child_fn_ptr): Likewise.
636 (gimple_omp_target_set_child_fn): Likewise.
637 (gimple_omp_target_data_arg): Likewise.
638 (gimple_omp_target_data_arg_ptr): Likewise.
639 (gimple_omp_target_set_data_arg): Likewise.
640 (gimple_omp_teams_clauses): Likewise.
641 (gimple_omp_teams_clauses_ptr): Likewise.
642 (gimple_omp_teams_set_clauses): Likewise.
643 (gimple_omp_sections_clauses): Likewise.
644 (gimple_omp_sections_clauses_ptr): Likewise.
645 (gimple_omp_sections_set_clauses): Likewise.
646 (gimple_omp_sections_control): Likewise.
647 (gimple_omp_sections_control_ptr): Likewise.
648 (gimple_omp_sections_set_control): Likewise.
649 (gimple_omp_for_set_cond): Likewise.
650 (gimple_omp_for_cond): Likewise.
651 (gimple_omp_atomic_store_set_val): Likewise.
652 (gimple_omp_atomic_store_val): Likewise.
653 (gimple_omp_atomic_store_val_ptr): Likewise.
654 (gimple_omp_atomic_load_set_lhs): Likewise.
655 (gimple_omp_atomic_load_lhs): Likewise.
656 (gimple_omp_atomic_load_lhs_ptr): Likewise.
657 (gimple_omp_atomic_load_set_rhs): Likewise.
658 (gimple_omp_atomic_load_rhs): Likewise.
659 (gimple_omp_atomic_load_rhs_ptr): Likewise.
660 (gimple_omp_continue_control_def): Likewise.
661 (gimple_omp_continue_control_def_ptr): Likewise.
662 (gimple_omp_continue_set_control_def): Likewise.
663 (gimple_omp_continue_control_use): Likewise.
664 (gimple_omp_continue_control_use_ptr): Likewise.
665 (gimple_omp_continue_set_control_use): Likewise.
666 (gimple_transaction_body_ptr): Likewise.
667 (gimple_transaction_label): Likewise.
668 (gimple_transaction_label_ptr): Likewise.
669 (gimple_transaction_set_body): Likewise.
670 (gimple_transaction_set_label): Likewise.
672 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
673 * ipa-inline-analysis.c (inline_write_summary): Likewise.
674 * ipa-ref.c (ipa_record_reference): Likewise.
675 * ipa-reference.c (analyze_function): Likewise.
676 (ipa_reference_write_optimization_summary): Likewise.
677 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
678 (address_taken_from_non_vtable_p): Likewise.
679 (comdat_can_be_unshared_p_1): Likewise.
680 * lto-cgraph.c (lto_output_ref): Likewise.
681 (add_references): Likewise.
682 (compute_ltrans_boundary): Likewise.
683 (output_symtab): Likewise.
684 (input_ref): Likewise.
685 (input_cgraph_1): Likewise.
686 (output_cgraph_opt_summary): Likewise.
687 * lto-streamer-out.c (lto_output): Likewise.
688 (output_symbol_p): Likewise.
689 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
690 (lsei_start_function_in_partition): Likewise.
691 (lsei_next_variable_in_partition): Likewise.
692 (lsei_start_variable_in_partition): Likewise.
693 * symtab.c (insert_to_assembler_name_hash): Likewise.
694 (unlink_from_assembler_name_hash): Likewise.
695 (symtab_unregister_node): Likewise.
696 (symtab_remove_node): Likewise.
697 (dump_symtab_node): Likewise.
698 (verify_symtab_base): Likewise.
699 (verify_symtab_node): Likewise.
700 (symtab_make_decl_local): Likewise.
701 (symtab_alias_ultimate_target): Likewise.
702 (symtab_resolve_alias): Likewise.
703 (symtab_get_symbol_partitioning_class): Likewise.
704 * tree-phinodes.c (allocate_phi_node): Likewise.
705 (reserve_phi_args_for_new_edge): Likewise.
706 (remove_phi_args): Likewise.
707 * varpool.c (varpool_node_for_asm): Likewise.
708 (varpool_remove_unreferenced_decls): Likewise.
710 2014-04-23 Jeff Law <law@redhat.com>
712 PR tree-optimization/60902
713 * tree-ssa-threadedge.c
714 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
715 invalidate outputs from statements that do not produce useful
716 outputs for threading.
718 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
720 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
721 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
722 machine descriptions for Stack Smashing Protector.
724 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
726 * aarch64.md (<optab>_rol<mode>3): New pattern.
727 (<optab>_rolsi3_uxtw): Likewise.
728 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
730 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
732 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
733 (arm_cortex_a12_tune): Likewise.
735 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
737 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
739 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
741 * config/arm/arm.md (arm_rev16si2): New pattern.
742 (arm_rev16si2_alt): Likewise.
743 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
745 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
747 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
748 (rev16<mode>2_alt): Likewise.
749 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
750 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
751 (aarch_rev16_shleft_mask_imm_p): Likewise.
752 (aarch_rev16_p_1): Likewise.
753 (aarch_rev16_p): Likewise.
754 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
755 (aarch_rev16_shright_mask_imm_p): Likewise.
756 (aarch_rev16_shleft_mask_imm_p): Likewise.
758 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
760 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
761 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
763 (cortex_a53_extra_costs): Likewise.
764 (cortex_a57_extra_costs): Likewise.
765 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
766 (cortexa7_extra_costs): Likewise.
767 (cortexa8_extra_costs): Likewise.
768 (cortexa12_extra_costs): Likewise.
769 (cortexa15_extra_costs): Likewise.
770 (v7m_extra_costs): Likewise.
771 (arm_new_rtx_costs): Handle BSWAP.
773 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
775 * config/arm/arm.c (cortexa8_extra_costs): New table.
776 (arm_cortex_a8_tune): New tuning struct.
777 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
779 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
781 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
783 2014-04-23 Richard Biener <rguenther@suse.de>
785 * Makefile.in (OBJS): Remove loop-unswitch.o.
786 * loop-unswitch.c: Delete.
787 * tree-pass.h (make_pass_rtl_unswitch): Remove.
788 * passes.def (pass_rtl_unswitch): Likewise.
789 * loop-init.c (gate_rtl_unswitch): Likewise.
790 (rtl_unswitch): Likewise.
791 (pass_data_rtl_unswitch): Likewise.
792 (pass_rtl_unswitch): Likewise.
793 (make_pass_rtl_unswitch): Likewise.
794 * rtl.h (reversed_condition): Likewise.
795 (compare_and_jump_seq): Likewise.
796 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
798 * loop-unroll.c (compare_and_jump_seq): Likewise.
800 2014-04-23 Richard Biener <rguenther@suse.de>
802 PR tree-optimization/60903
803 * tree-ssa-loop-im.c (analyze_memory_references): Remove
804 commented code block.
805 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
806 loop flags to newly created BBs and edges.
808 2014-04-23 Nick Clifton <nickc@redhat.com>
810 * config/msp430/msp430.c (msp430_handle_option): Move function
812 (msp430_option_override): Simplify mcu and mcpu option handling.
813 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
814 support for -mhwmult command line option.
815 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
816 -mhwmult command line option.
817 (msp430_hwmult_enabled): Delete.
818 (msp43o_output_labelref): Add support for -mhwmult command line
820 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
821 (umulsidi3): Likewise.
822 * config/msp430/msp430.opt (mmcu): Add Report attribute.
823 (mcpu, mlarge, msmall): Likewise.
824 (mhwmult): New option.
825 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
827 (msp430_is_f5_mcu): Remove prototype.
828 (msp430_use_f5_series_hwmult): Add prototype.
829 * config/msp430/msp430-opts.h: New file.
830 * common/config/msp430: New directory.
831 * common/config/msp430/msp430-common.c: New file.
832 * config.gcc (msp430): Remove target_has_targetm_common.
833 * doc/invoke.texi: Document -mhwmult command line option.
835 2014-04-23 Nick Clifton <nickc@redhat.com>
837 * config/i386/cygwin.h (ENDFILE_SPEC): Include
838 default-manifest.o if it can be found in the search path.
839 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
841 2014-04-23 Terry Guo <terry.guo@arm.com>
843 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
845 2014-04-23 Richard Biener <rguenther@suse.de>
848 * tree-inline.c (declare_return_variable): Use mark_addressable.
850 2014-04-23 Richard Biener <rguenther@suse.de>
853 * loop-init.c (loop_optimizer_init): Make sure to apply
854 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
856 2014-04-22 Jakub Jelinek <jakub@redhat.com>
859 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
861 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
862 if flag_sanitize_undefined_trap_on_error.
863 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
864 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
865 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
866 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
867 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
868 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
869 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
870 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
871 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
872 * ubsan.c (ubsan_instrument_unreachable): Return
873 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
874 (ubsan_expand_null_ifn): Emit __builtin_trap ()
875 if flag_sanitize_undefined_trap_on_error and
876 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
877 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
878 instrument_bool_enum_load): Emit __builtin_trap () if
879 flag_sanitize_undefined_trap_on_error and
880 __builtin_handle_*_abort () if !flag_sanitize_recover.
881 * doc/invoke.texi (-fsanitize-recover,
882 -fsanitize-undefined-trap-on-error): Document.
884 2014-04-22 Christian Bruel <christian.bruel@st.com>
886 * config/sh/sh.md (mov<mode>): Replace movQIHI.
887 Force immediates to SImode.
889 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
891 * config/nios2/nios2.md (UNSPEC_ROUND): New.
893 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
894 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
895 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
896 (nios2_fpu_insn): Add entry for round.
897 (N2FPU_NO_ERRNO_P): Define.
898 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
900 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
902 2014-04-22 Richard Henderson <rth@redhat.com>
904 * config/aarch64/aarch64 (addti3, subti3): New expanders.
905 (add<GPI>3_compare0): Remove leading * from name.
906 (add<GPI>3_carryin): Likewise.
907 (sub<GPI>3_compare0): Likewise.
908 (sub<GPI>3_carryin): Likewise.
909 (<su_optab>mulditi3): New expander.
910 (multi3): New expander.
911 (madd<GPI>): Remove leading * from name.
913 2014-04-22 Martin Jambor <mjambor@suse.cz>
915 * cgraphclones.c (cgraph_function_versioning): Copy
916 ipa_transforms_to_apply instead of asserting it is empty.
918 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
921 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
922 on count_exp to get mode.
924 2014-04-22 Andrew Pinski <apinski@cavium.com>
926 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
927 Handle TLS for ILP32.
928 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
929 (tlsie_small_<mode>): this and handle PTR.
930 (tlsie_small_sidi): New pattern.
931 (tlsle_small): Change to an expand to handle ILP32.
932 (tlsle_small_<mode>): New pattern.
933 (tlsdesc_small): Rename to ...
934 (tlsdesc_small_<mode>): this and handle PTR.
936 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
938 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
940 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
942 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
943 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
944 (aarch64_types_signed_poly_qualifiers): Likewise.
945 (aarch64_types_unsigned_signed_qualifiers): Likewise.
946 (aarch64_types_poly_signed_qualifiers): Likewise.
947 (TYPES_REINTERP_SS): Type macro added.
948 (TYPES_REINTERP_SU): Likewise.
949 (TYPES_REINTERP_SP): Likewise.
950 (TYPES_REINTERP_US): Likewise.
951 (TYPES_REINTERP_PS): Likewise.
952 (aarch64_fold_builtin): New expression folding added.
953 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
954 Declarations removed.
955 (REINTERP_SS): Declarations added.
956 (REINTERP_US): Likewise.
957 (REINTERP_PS): Likewise.
958 (REINTERP_SU): Likewise.
959 (REINTERP_SP): Likewise.
960 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
961 (vreinterpretq_p8_f64): Likewise.
962 (vreinterpret_p16_f64): Likewise.
963 (vreinterpretq_p16_f64): Likewise.
964 (vreinterpret_f32_f64): Likewise.
965 (vreinterpretq_f32_f64): Likewise.
966 (vreinterpret_f64_f32): Likewise.
967 (vreinterpret_f64_p8): Likewise.
968 (vreinterpret_f64_p16): Likewise.
969 (vreinterpret_f64_s8): Likewise.
970 (vreinterpret_f64_s16): Likewise.
971 (vreinterpret_f64_s32): Likewise.
972 (vreinterpret_f64_s64): Likewise.
973 (vreinterpret_f64_u8): Likewise.
974 (vreinterpret_f64_u16): Likewise.
975 (vreinterpret_f64_u32): Likewise.
976 (vreinterpret_f64_u64): Likewise.
977 (vreinterpretq_f64_f32): Likewise.
978 (vreinterpretq_f64_p8): Likewise.
979 (vreinterpretq_f64_p16): Likewise.
980 (vreinterpretq_f64_s8): Likewise.
981 (vreinterpretq_f64_s16): Likewise.
982 (vreinterpretq_f64_s32): Likewise.
983 (vreinterpretq_f64_s64): Likewise.
984 (vreinterpretq_f64_u8): Likewise.
985 (vreinterpretq_f64_u16): Likewise.
986 (vreinterpretq_f64_u32): Likewise.
987 (vreinterpretq_f64_u64): Likewise.
988 (vreinterpret_s64_f64): Likewise.
989 (vreinterpretq_s64_f64): Likewise.
990 (vreinterpret_u64_f64): Likewise.
991 (vreinterpretq_u64_f64): Likewise.
992 (vreinterpret_s8_f64): Likewise.
993 (vreinterpretq_s8_f64): Likewise.
994 (vreinterpret_s16_f64): Likewise.
995 (vreinterpretq_s16_f64): Likewise.
996 (vreinterpret_s32_f64): Likewise.
997 (vreinterpretq_s32_f64): Likewise.
998 (vreinterpret_u8_f64): Likewise.
999 (vreinterpretq_u8_f64): Likewise.
1000 (vreinterpret_u16_f64): Likewise.
1001 (vreinterpretq_u16_f64): Likewise.
1002 (vreinterpret_u32_f64): Likewise.
1003 (vreinterpretq_u32_f64): Likewise.
1005 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1007 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1008 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1009 (vreinterpret_p8_s8): Likewise.
1010 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1011 (vreinterpret_p8_s16): Likewise.
1012 (vreinterpret_p8_s32): Likewise.
1013 (vreinterpret_p8_s64): Likewise.
1014 (vreinterpret_p8_f32): Likewise.
1015 (vreinterpret_p8_u8): Likewise.
1016 (vreinterpret_p8_u16): Likewise.
1017 (vreinterpret_p8_u32): Likewise.
1018 (vreinterpret_p8_u64): Likewise.
1019 (vreinterpret_p8_p16): Likewise.
1020 (vreinterpretq_p8_s8): Likewise.
1021 (vreinterpretq_p8_s16): Likewise.
1022 (vreinterpretq_p8_s32): Likewise.
1023 (vreinterpretq_p8_s64): Likewise.
1024 (vreinterpretq_p8_f32): Likewise.
1025 (vreinterpretq_p8_u8): Likewise.
1026 (vreinterpretq_p8_u16): Likewise.
1027 (vreinterpretq_p8_u32): Likewise.
1028 (vreinterpretq_p8_u64): Likewise.
1029 (vreinterpretq_p8_p16): Likewise.
1030 (vreinterpret_p16_s8): Likewise.
1031 (vreinterpret_p16_s16): Likewise.
1032 (vreinterpret_p16_s32): Likewise.
1033 (vreinterpret_p16_s64): Likewise.
1034 (vreinterpret_p16_f32): Likewise.
1035 (vreinterpret_p16_u8): Likewise.
1036 (vreinterpret_p16_u16): Likewise.
1037 (vreinterpret_p16_u32): Likewise.
1038 (vreinterpret_p16_u64): Likewise.
1039 (vreinterpret_p16_p8): Likewise.
1040 (vreinterpretq_p16_s8): Likewise.
1041 (vreinterpretq_p16_s16): Likewise.
1042 (vreinterpretq_p16_s32): Likewise.
1043 (vreinterpretq_p16_s64): Likewise.
1044 (vreinterpretq_p16_f32): Likewise.
1045 (vreinterpretq_p16_u8): Likewise.
1046 (vreinterpretq_p16_u16): Likewise.
1047 (vreinterpretq_p16_u32): Likewise.
1048 (vreinterpretq_p16_u64): Likewise.
1049 (vreinterpretq_p16_p8): Likewise.
1050 (vreinterpret_f32_s8): Likewise.
1051 (vreinterpret_f32_s16): Likewise.
1052 (vreinterpret_f32_s32): Likewise.
1053 (vreinterpret_f32_s64): Likewise.
1054 (vreinterpret_f32_u8): Likewise.
1055 (vreinterpret_f32_u16): Likewise.
1056 (vreinterpret_f32_u32): Likewise.
1057 (vreinterpret_f32_u64): Likewise.
1058 (vreinterpret_f32_p8): Likewise.
1059 (vreinterpret_f32_p16): Likewise.
1060 (vreinterpretq_f32_s8): Likewise.
1061 (vreinterpretq_f32_s16): Likewise.
1062 (vreinterpretq_f32_s32): Likewise.
1063 (vreinterpretq_f32_s64): Likewise.
1064 (vreinterpretq_f32_u8): Likewise.
1065 (vreinterpretq_f32_u16): Likewise.
1066 (vreinterpretq_f32_u32): Likewise.
1067 (vreinterpretq_f32_u64): Likewise.
1068 (vreinterpretq_f32_p8): Likewise.
1069 (vreinterpretq_f32_p16): Likewise.
1070 (vreinterpret_s64_s8): Likewise.
1071 (vreinterpret_s64_s16): Likewise.
1072 (vreinterpret_s64_s32): Likewise.
1073 (vreinterpret_s64_f32): Likewise.
1074 (vreinterpret_s64_u8): Likewise.
1075 (vreinterpret_s64_u16): Likewise.
1076 (vreinterpret_s64_u32): Likewise.
1077 (vreinterpret_s64_u64): Likewise.
1078 (vreinterpret_s64_p8): Likewise.
1079 (vreinterpret_s64_p16): Likewise.
1080 (vreinterpretq_s64_s8): Likewise.
1081 (vreinterpretq_s64_s16): Likewise.
1082 (vreinterpretq_s64_s32): Likewise.
1083 (vreinterpretq_s64_f32): Likewise.
1084 (vreinterpretq_s64_u8): Likewise.
1085 (vreinterpretq_s64_u16): Likewise.
1086 (vreinterpretq_s64_u32): Likewise.
1087 (vreinterpretq_s64_u64): Likewise.
1088 (vreinterpretq_s64_p8): Likewise.
1089 (vreinterpretq_s64_p16): Likewise.
1090 (vreinterpret_u64_s8): Likewise.
1091 (vreinterpret_u64_s16): Likewise.
1092 (vreinterpret_u64_s32): Likewise.
1093 (vreinterpret_u64_s64): Likewise.
1094 (vreinterpret_u64_f32): Likewise.
1095 (vreinterpret_u64_u8): Likewise.
1096 (vreinterpret_u64_u16): Likewise.
1097 (vreinterpret_u64_u32): Likewise.
1098 (vreinterpret_u64_p8): Likewise.
1099 (vreinterpret_u64_p16): Likewise.
1100 (vreinterpretq_u64_s8): Likewise.
1101 (vreinterpretq_u64_s16): Likewise.
1102 (vreinterpretq_u64_s32): Likewise.
1103 (vreinterpretq_u64_s64): Likewise.
1104 (vreinterpretq_u64_f32): Likewise.
1105 (vreinterpretq_u64_u8): Likewise.
1106 (vreinterpretq_u64_u16): Likewise.
1107 (vreinterpretq_u64_u32): Likewise.
1108 (vreinterpretq_u64_p8): Likewise.
1109 (vreinterpretq_u64_p16): Likewise.
1110 (vreinterpret_s8_s16): Likewise.
1111 (vreinterpret_s8_s32): Likewise.
1112 (vreinterpret_s8_s64): Likewise.
1113 (vreinterpret_s8_f32): Likewise.
1114 (vreinterpret_s8_u8): Likewise.
1115 (vreinterpret_s8_u16): Likewise.
1116 (vreinterpret_s8_u32): Likewise.
1117 (vreinterpret_s8_u64): Likewise.
1118 (vreinterpret_s8_p8): Likewise.
1119 (vreinterpret_s8_p16): Likewise.
1120 (vreinterpretq_s8_s16): Likewise.
1121 (vreinterpretq_s8_s32): Likewise.
1122 (vreinterpretq_s8_s64): Likewise.
1123 (vreinterpretq_s8_f32): Likewise.
1124 (vreinterpretq_s8_u8): Likewise.
1125 (vreinterpretq_s8_u16): Likewise.
1126 (vreinterpretq_s8_u32): Likewise.
1127 (vreinterpretq_s8_u64): Likewise.
1128 (vreinterpretq_s8_p8): Likewise.
1129 (vreinterpretq_s8_p16): Likewise.
1130 (vreinterpret_s16_s8): Likewise.
1131 (vreinterpret_s16_s32): Likewise.
1132 (vreinterpret_s16_s64): Likewise.
1133 (vreinterpret_s16_f32): Likewise.
1134 (vreinterpret_s16_u8): Likewise.
1135 (vreinterpret_s16_u16): Likewise.
1136 (vreinterpret_s16_u32): Likewise.
1137 (vreinterpret_s16_u64): Likewise.
1138 (vreinterpret_s16_p8): Likewise.
1139 (vreinterpret_s16_p16): Likewise.
1140 (vreinterpretq_s16_s8): Likewise.
1141 (vreinterpretq_s16_s32): Likewise.
1142 (vreinterpretq_s16_s64): Likewise.
1143 (vreinterpretq_s16_f32): Likewise.
1144 (vreinterpretq_s16_u8): Likewise.
1145 (vreinterpretq_s16_u16): Likewise.
1146 (vreinterpretq_s16_u32): Likewise.
1147 (vreinterpretq_s16_u64): Likewise.
1148 (vreinterpretq_s16_p8): Likewise.
1149 (vreinterpretq_s16_p16): Likewise.
1150 (vreinterpret_s32_s8): Likewise.
1151 (vreinterpret_s32_s16): Likewise.
1152 (vreinterpret_s32_s64): Likewise.
1153 (vreinterpret_s32_f32): Likewise.
1154 (vreinterpret_s32_u8): Likewise.
1155 (vreinterpret_s32_u16): Likewise.
1156 (vreinterpret_s32_u32): Likewise.
1157 (vreinterpret_s32_u64): Likewise.
1158 (vreinterpret_s32_p8): Likewise.
1159 (vreinterpret_s32_p16): Likewise.
1160 (vreinterpretq_s32_s8): Likewise.
1161 (vreinterpretq_s32_s16): Likewise.
1162 (vreinterpretq_s32_s64): Likewise.
1163 (vreinterpretq_s32_f32): Likewise.
1164 (vreinterpretq_s32_u8): Likewise.
1165 (vreinterpretq_s32_u16): Likewise.
1166 (vreinterpretq_s32_u32): Likewise.
1167 (vreinterpretq_s32_u64): Likewise.
1168 (vreinterpretq_s32_p8): Likewise.
1169 (vreinterpretq_s32_p16): Likewise.
1170 (vreinterpret_u8_s8): Likewise.
1171 (vreinterpret_u8_s16): Likewise.
1172 (vreinterpret_u8_s32): Likewise.
1173 (vreinterpret_u8_s64): Likewise.
1174 (vreinterpret_u8_f32): Likewise.
1175 (vreinterpret_u8_u16): Likewise.
1176 (vreinterpret_u8_u32): Likewise.
1177 (vreinterpret_u8_u64): Likewise.
1178 (vreinterpret_u8_p8): Likewise.
1179 (vreinterpret_u8_p16): Likewise.
1180 (vreinterpretq_u8_s8): Likewise.
1181 (vreinterpretq_u8_s16): Likewise.
1182 (vreinterpretq_u8_s32): Likewise.
1183 (vreinterpretq_u8_s64): Likewise.
1184 (vreinterpretq_u8_f32): Likewise.
1185 (vreinterpretq_u8_u16): Likewise.
1186 (vreinterpretq_u8_u32): Likewise.
1187 (vreinterpretq_u8_u64): Likewise.
1188 (vreinterpretq_u8_p8): Likewise.
1189 (vreinterpretq_u8_p16): Likewise.
1190 (vreinterpret_u16_s8): Likewise.
1191 (vreinterpret_u16_s16): Likewise.
1192 (vreinterpret_u16_s32): Likewise.
1193 (vreinterpret_u16_s64): Likewise.
1194 (vreinterpret_u16_f32): Likewise.
1195 (vreinterpret_u16_u8): Likewise.
1196 (vreinterpret_u16_u32): Likewise.
1197 (vreinterpret_u16_u64): Likewise.
1198 (vreinterpret_u16_p8): Likewise.
1199 (vreinterpret_u16_p16): Likewise.
1200 (vreinterpretq_u16_s8): Likewise.
1201 (vreinterpretq_u16_s16): Likewise.
1202 (vreinterpretq_u16_s32): Likewise.
1203 (vreinterpretq_u16_s64): Likewise.
1204 (vreinterpretq_u16_f32): Likewise.
1205 (vreinterpretq_u16_u8): Likewise.
1206 (vreinterpretq_u16_u32): Likewise.
1207 (vreinterpretq_u16_u64): Likewise.
1208 (vreinterpretq_u16_p8): Likewise.
1209 (vreinterpretq_u16_p16): Likewise.
1210 (vreinterpret_u32_s8): Likewise.
1211 (vreinterpret_u32_s16): Likewise.
1212 (vreinterpret_u32_s32): Likewise.
1213 (vreinterpret_u32_s64): Likewise.
1214 (vreinterpret_u32_f32): Likewise.
1215 (vreinterpret_u32_u8): Likewise.
1216 (vreinterpret_u32_u16): Likewise.
1217 (vreinterpret_u32_u64): Likewise.
1218 (vreinterpret_u32_p8): Likewise.
1219 (vreinterpret_u32_p16): Likewise.
1220 (vreinterpretq_u32_s8): Likewise.
1221 (vreinterpretq_u32_s16): Likewise.
1222 (vreinterpretq_u32_s32): Likewise.
1223 (vreinterpretq_u32_s64): Likewise.
1224 (vreinterpretq_u32_f32): Likewise.
1225 (vreinterpretq_u32_u8): Likewise.
1226 (vreinterpretq_u32_u16): Likewise.
1227 (vreinterpretq_u32_u64): Likewise.
1228 (vreinterpretq_u32_p8): Likewise.
1229 (vreinterpretq_u32_p16): Likewise.
1231 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1233 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1235 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1238 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1239 (vqnegd_s64): Likewise.
1240 (vqabs_s64): Likewise.
1241 (vqabsd_s64): Likewise.
1243 2014-04-22 Richard Henderson <rth@redhat.com>
1245 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1246 computation to the top of the loop.
1248 2014-04-22 Renlin <renlin.li@arm.com>
1249 Jiong Wang <jiong.wang@arm.com>
1251 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1252 * config/aarch64/aarch64.c (aarch64_layout_frame)
1253 (aarch64_initial_elimination_offset): Likewise.
1255 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1257 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1260 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1262 * machmode.h (bitwise_mode_for_mode): Declare.
1263 * stor-layout.h (bitwise_type_for_mode): Likewise.
1264 * stor-layout.c (bitwise_mode_for_mode): New function.
1265 (bitwise_type_for_mode): Likewise.
1266 * builtins.c (fold_builtin_memory_op): Use it instead of
1267 int_mode_for_mode and build_nonstandard_integer_type.
1269 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1271 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1272 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1273 (*-*-solaris2*): Simplify.
1274 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1275 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1276 *-*-solaris2.9* handling.
1278 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1280 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1281 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1283 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1284 * configure: Regenerate.
1286 * config/i386/sol2-9.h: Remove.
1288 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1289 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1290 Remove Solaris 9 references.
1292 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1294 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1295 (floatuns<GPI:mode><GPF:mode>2): Remove.
1296 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1297 and floatuns conversions.
1298 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1299 and floatuns conversions.
1300 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1301 (w1,w2): New mode attributes for inequal width conversions.
1303 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1305 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1306 the output asm format.
1308 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1310 * config/aarch64/aarch64-simd.md
1311 (aarch64_cm<optab>di): Always split.
1312 (*aarch64_cm<optab>di): New.
1313 (aarch64_cmtstdi): Always split.
1314 (*aarch64_cmtstdi): New.
1316 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1318 PR tree-optimization/60823
1319 * omp-low.c (ipa_simd_modify_function_body): Go through
1320 all SSA_NAMEs and for those refering to vector arguments
1321 which are going to be replaced adjust SSA_NAME_VAR and,
1322 if it is a default definition, change it into a non-default
1323 definition assigned at the beginning of function from new_decl.
1324 (ipa_simd_modify_stmt_ops): Rewritten.
1325 * tree-dfa.c (set_ssa_default_def): When removing default def,
1326 check for NULL loc instead of NULL *loc.
1328 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1330 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1331 restrictions on core registers for DImode values in Thumb2.
1333 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1335 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1336 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1338 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1340 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1341 (*iordi_notzesidi_di): Likewise.
1342 (*iordi_notsesidi_di): Likewise.
1344 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1346 * config/arm/arm-protos.h (tune_params): New struct members.
1347 * config/arm/arm.c: Initialise tune_params per processor.
1348 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1349 for speed, based on new tune_params.
1351 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1353 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1355 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1357 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1359 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1360 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1361 (vrnda_f64): Likewise.
1362 (vrndi_f64): Likewise.
1363 (vrndm_f64): Likewise.
1364 (vrndn_f64): Likewise.
1365 (vrndp_f64): Likewise.
1366 (vrndx_f64): Likewise.
1368 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1370 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1371 GET_MODE_SIZE argument is enum machine_mode.
1373 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1376 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1377 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1379 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1382 * asan.c (asan_emit_stack_protection): Force the base to align to
1383 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1384 appropriate bits if STRICT_ALIGNMENT.
1385 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1387 (expand_used_vars): Leave a space in the stack frame for alignment
1388 if STRICT_ALIGNMENT.
1390 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1392 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1394 (gimple_store_p): Likewise.
1395 (gimple_assign_load_p): Likewise.
1396 (gimple_assign_cast_p): Likewise.
1397 (gimple_clobber_p): Likewise.
1399 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1400 rather than a gimple.
1401 (gimple_assign_cast_p): Likewise.
1403 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1406 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1407 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1409 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1410 more debug information for E500 if -mdebug=reg.
1412 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1415 * config/i386/i386.c (ix86_expand_builtin)
1416 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1417 register for target RTX.
1418 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1420 2014-04-18 Cong Hou <congh@google.com>
1422 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1423 the widen-mult pattern by handling two operands with different sizes,
1424 and operands whose size is smaller than half of the result type.
1426 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1428 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1429 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1430 (do_estimate_edge_time): Compute it.
1431 * ipa-inline.c (want_inline_small_function_p): Bypass
1432 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1434 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1436 * ipa-inline.c (spec_rem): New static variable.
1437 (dump_overall_stats): New function.
1438 (dump_inline_stats): New function.
1440 2014-04-18 Richard Henderson <rth@redhat.com>
1442 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1443 to GET_MODE_SIZE, not a reg_class_t.
1445 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1447 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1448 (vsx_xxmrglw_<mode>): Likewise.
1450 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1453 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1454 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1455 (rs6000_init_hard_regno_mode_ok): Likewise.
1457 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1459 * ipa-inline.c (inline_small_functions): Account only non-cold
1461 * doc/invoke.texi (inline-unit-growth): Update documentation.
1463 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1465 * config/rs6000/rs6000.md (addti3, subti3): New.
1467 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1470 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1471 comment. Check optimize_insn_for_size_p instead of
1472 optimize_insn_for_speed_p.
1474 2014-04-17 Martin Jambor <mjambor@suse.cz>
1476 * gimple-iterator.c (gsi_start_edge): New function.
1477 * gimple-iterator.h (gsi_start_edge): Declare.
1478 * tree-sra.c (single_non_eh_succ): New function.
1479 (disqualify_ops_if_throwing_stmt): Renamed to
1480 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1481 having one non-EH successor BB.
1482 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1483 generate loads into replacements.
1484 (sra_modify_assign): Likewise and and also use the simple path for
1486 (sra_modify_function_body): Commit statements on edges.
1488 2014-04-17 Richard Biener <rguenther@suse.de>
1491 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1492 comparison results and add clarifying comment.
1494 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1496 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1497 (blank_mode): Initialize it.
1498 (emit_mode_size_inline, emit_mode_nunits_inline,
1499 emit_mode_inner_inline): New functions.
1500 (emit_insn_modes_h): Call them and surround their output with
1501 #if GCC_VERSION >= 4001 ... #endif.
1502 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1503 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1504 mode_* arrays if the argument is __builtin_constant_p.
1505 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1506 is enum machine_mode.
1508 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1510 * passes.c (opt_pass::execute): Adjust.
1511 (pass_manager::execute_pass_mode_switching): Likewise.
1512 (early_local_passes::execute): Likewise.
1513 (execute_one_pass): Pass cfun to the pass's execute method.
1514 * tree-pass.h (opt_pass::execute): Add function * argument.
1515 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1516 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1517 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1518 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1519 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1520 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1521 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1522 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1523 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1524 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1525 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1526 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1527 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1528 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1529 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1530 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1531 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1532 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1533 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1534 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1535 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1536 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1537 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1538 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1539 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1540 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1541 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1542 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1543 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1546 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1548 * passes.c (opt_pass::gate): Take function * argument.
1549 (gate_all_early_local_passes): Merge into
1550 (early_local_passes::gate): this.
1551 (gate_all_early_optimizations): Merge into
1552 (all_early_optimizations::gate): this.
1553 (gate_all_optimizations): Mege into
1554 (all_optimizations::gate): this.
1555 (gate_all_optimizations_g): Merge into
1556 (all_optimizations_g::gate): this.
1557 (gate_rest_of_compilation): Mege into
1558 (rest_of_compilation::gate): this.
1559 (gate_postreload): Merge into
1560 (postreload::gate): this.
1561 (dump_one_pass): Pass cfun to the pass's gate method.
1562 (execute_ipa_summary_passes): Likewise.
1563 (execute_one_pass): Likewise.
1564 (ipa_write_summaries_2): Likewise.
1565 (ipa_write_optimization_summaries_1): Likewise.
1566 (ipa_read_summaries_1): Likewise.
1567 (ipa_read_optimization_summaries_1): Likewise.
1568 (execute_ipa_stmt_fixups): Likewise.
1569 * tree-pass.h (opt_pass::gate): Add function * argument.
1570 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1571 combine-stack-adj.c, combine.c, compare-elim.c,
1572 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1573 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1574 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1575 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1576 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1577 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1578 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1579 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1580 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1581 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1582 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1583 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1584 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1585 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1586 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1587 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1588 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1589 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1590 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1591 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1592 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1593 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1594 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1595 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1596 var-tracking.c, vtable-verify.c, web.c: Adjust.
1598 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1600 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1601 * configure: Regenerate.
1603 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1605 * passes.c (dump_one_pass): don't check pass->has_gate.
1606 (execute_ipa_summary_passes): Likewise.
1607 (execute_one_pass): Likewise.
1608 (ipa_write_summaries_2): Likewise.
1609 (ipa_write_optimization_summaries_1): Likewise.
1610 (ipa_read_optimization_summaries_1): Likewise.
1611 (execute_ipa_stmt_fixups): Likewise.
1612 * tree-pass.h (pass_data::has_gate): Remove.
1613 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1614 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1615 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1616 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1617 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1618 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1619 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1620 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1621 gimple-low.c, gimple-ssa-isolate-paths.c,
1622 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1623 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1624 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1625 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1626 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1627 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1628 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1629 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1630 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1631 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1632 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1633 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1634 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1635 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1636 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1637 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1638 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1639 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1640 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1641 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1642 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1643 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1646 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1648 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1650 * passes.c (pass_manager::register_dump_files_1): Merge into
1651 (pass_manager::register_dump_files): this, and remove its handling of
1652 properties since the pass always has the properties anyway.
1653 (pass_manager::pass_manager): Adjust.
1655 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1657 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1658 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1659 dealing with properties.
1660 (pass_manager::register_dump_files): Adjust.
1662 2014-03-20 Mark Wielaard <mjw@redhat.com>
1664 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1665 then represent the bound as normal constant value.
1667 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1670 Forward port from 4.8 branch
1671 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1673 * config/i386/bmiintrin.h (_blsi_u32): New.
1677 (_blsmsk_u32): Ditto.
1678 (_blsmsk_u64): Ditto.
1679 (_tzcnt_u32): Ditto.
1680 (_tzcnt_u64): Ditto.
1682 2014-04-17 Kito Cheng <kito@0xlab.org>
1684 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1686 2014-04-17 Richard Biener <rguenther@suse.de>
1689 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1690 boolean results for comparisons.
1692 2014-04-17 Richard Biener <rguenther@suse.de>
1694 PR tree-optimization/60836
1695 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1696 initial PHI args to be gimple values.
1698 2014-04-17 Richard Biener <rguenther@suse.de>
1700 PR tree-optimization/60841
1701 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1702 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1703 of stmts to SLP build.
1704 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1705 (vect_analyze_slp): Likewise.
1706 (vect_analyze_slp_instance): Likewise.
1707 (vect_build_slp_tree): Limit overall SLP tree growth.
1708 * tree-vectorizer.h (vect_analyze_data_refs,
1709 vect_analyze_slp): Adjust prototypes.
1711 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1713 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1716 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1718 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1719 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1720 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1721 for TARGET_SLOW_PSHUFB
1723 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1725 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1726 * config/i386/i386.c (intel_cost): Ditto.
1728 2014-04-17 Joey Ye <joey.ye@arm.com>
1730 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1732 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1734 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1737 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1739 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1740 (type_all_derivations_known_p): New predicate.
1741 (type_all_ctors_visible_p): New predicate.
1742 (type_possibly_instantiated_p): New predicate.
1743 (get_odr_type): Compute all_derivations_known.
1744 (dump_odr_type): Dump the flag.
1745 (maybe_record_type): Cleanup.
1746 (record_target_from_binfo): Add bases_to_consider array;
1747 record bases for types w/o instances and skip CXX destructor.
1748 (possible_polymorphic_call_targets_1): Add bases_to_consider
1749 and consider_construction parameters; check if type may have instance.
1750 (get_polymorphic_call_info): Set maybe_in_construction to true
1751 when we know nothing.
1752 (record_targets_from_bases): Skip CXX destructors; they are
1753 never called for types in construction.
1754 (possible_polymorphic_call_targets): Do not record target when
1755 type may not have instance.
1757 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1760 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1761 external aliases alive, too.
1763 2014-04-16 Andrew Pinski <apinski@cavium.com>
1765 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1768 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1770 * final.c (compute_alignments): Do not apply loop alignment to a block
1771 falling through to the exit.
1773 2014-04-16 Catherine Moore <clm@codesourcery.com>
1775 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1776 Adjust constraints for microMIPS store patterns.
1778 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1780 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1782 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1784 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1785 (append_use): Run at -O0.
1786 (append_vdef): Likewise.
1787 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1788 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1790 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1792 PR tree-optimization/60844
1793 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1794 (propagate_op_to_single_use, remove_visited_stmt_chain,
1795 linearize_expr, repropagate_negates, reassociate_bb): Use it
1796 instead of gsi_remove.
1798 2014-04-16 Martin Jambor <mjambor@suse.cz>
1800 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1801 ipa_transforms_to_apply.
1802 (cgraph_function_versioning): Assert that old_node has empty
1803 ipa_transforms_to_apply.
1804 * trans-mem.c (ipa_tm_create_version): Likewise.
1805 * tree-inline.c (tree_function_versioning): Do not duplicate
1806 ipa_transforms_to_apply.
1808 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1811 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
1813 Pass necessary as flags on 64-bit Solaris/x86.
1814 Use lowercase relocs for x86_64-*-*.
1815 * configure: Regenerate.
1817 2014-04-15 Jan Hubicka <jh@suse.cz>
1819 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
1820 (maybe_record_node, likely_target_p): Use it.
1822 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1825 Revert following patch
1827 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1830 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
1831 software floating point or no floating point registers, do not
1832 allow any type in the FPRs. Eliminate a test for SPE SIMD types
1833 in GPRs that occurs after we tested for GPRs that would never be
1836 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
1837 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
1838 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
1839 specifically allow DDmode, since that does not use the SPE SIMD
1842 2014-03-21 Mark Wielaard <mjw@redhat.com>
1844 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
1845 as unsigned or int depending on type and value used.
1847 2014-04-15 Richard Biener <rguenther@suse.de>
1849 PR rtl-optimization/56965
1850 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
1851 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
1853 * alias.c (true_dependence_1): Do not call
1854 nonoverlapping_component_refs_p.
1855 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
1856 nonoverlapping_component_refs_p.
1857 (indirect_refs_may_alias_p): Likewise.
1859 2014-04-15 Teresa Johnson <tejohnson@google.com>
1861 * cfg.c (dump_bb_info): Fix flags check.
1862 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
1864 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1866 PR rtl-optimization/60663
1867 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
1870 2014-04-15 Richard Biener <rguenther@suse.de>
1872 * lto-streamer.h (LTO_major_version): Bump to 4.
1874 2014-04-15 Richard Biener <rguenther@suse.de>
1876 * common.opt (lto_partition_model): New enum.
1877 (flto-partition=): Merge separate options with a single with argument,
1878 add -flto-partition=one support.
1879 * flag-types.h (enum lto_partition_model): Declare.
1880 * opts.c (finish_options): Remove duplicate -flto-partition=
1882 * lto-wrapper.c (run_gcc): Adjust.
1884 2014-04-15 Richard Biener <rguenther@suse.de>
1886 * alias.c (ncr_compar): New function.
1887 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
1889 2014-04-15 Richard Biener <rguenther@suse.de>
1891 * alias.c (record_component_aliases): Do not walk BINFOs.
1893 2014-04-15 Richard Biener <rguenther@suse.de>
1895 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1896 Add struct function argument and adjust.
1897 (find_func_aliases_for_call): Likewise.
1898 (find_func_aliases): Likewise.
1899 (find_func_clobbers): Likewise.
1900 (intra_create_variable_infos): Likewise.
1901 (compute_points_to_sets): Likewise.
1902 (ipa_pta_execute): Adjust. Do not push/pop cfun.
1904 2014-04-15 Richard Biener <rguenther@suse.de>
1906 * tree.c (iterative_hash_expr): Use enum tree_code_class
1907 to store TREE_CODE_CLASS.
1908 (tree_block): Likewise.
1909 (tree_set_block): Likewise.
1910 * tree.h (fold_build_pointer_plus_loc): Use
1911 convert_to_ptrofftype_loc.
1913 2014-04-15 Jakub Jelinek <jakub@redhat.com>
1916 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
1919 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
1921 * cfgloop.h (struct loop): Move force_vectorize down.
1922 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
1923 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
1924 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
1925 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
1926 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
1927 * tree-core.h (enum annot_expr_kind): Add new kind values.
1928 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
1929 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
1931 * tree.def (ANNOTATE_EXPR): Tweak comment.
1933 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1935 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
1938 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
1940 * tree.h (TYPE_IDENTIFIER): Declare.
1941 * tree.c (subrange_type_for_debug_p): Use it.
1942 * godump.c (go_format_type): Likewise.
1943 * dwarf2out.c (is_cxx_auto, modified_type_die,
1944 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
1945 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
1947 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1950 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
1952 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
1954 * config/i386/i386.c (examine_argument): Return bool. Return true if
1955 parameter should be passed in memory.
1956 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
1957 (construct_container): Update calls to examine_argument.
1958 (function_arg_advance_64): Ditto.
1959 (return_in_memory_32): Merge with ix86_return_in_memory.
1960 (return_in_memory_64): Ditto.
1961 (return_in_memory_ms_64): Ditto.
1963 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1965 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
1966 * coverage.c (coverage_compute_profile_id): Handle externally visible
1969 2014-04-14 Martin Jambor <mjambor@suse.cz>
1971 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
1972 DECL_DISREGARD_INLINE_LIMITS functions.
1974 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
1977 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
1979 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
1982 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
1983 optimize_insn_for_speed_p instead of
1984 optimize_function_for_speed_p.
1986 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
1988 * doc/invoke.texi (free): Document AArch64.
1990 2014-04-14 Richard Biener <rguenther@suse.de>
1992 PR tree-optimization/60042
1993 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
1994 (insert_into_preds_of_block): Do not prevent PHI insertion
1995 for REFERENCE exprs here ...
1996 (eliminate_dom_walker::before_dom_children): ... but prevent
1997 their use here under similar conditions when applied to the
1998 IL after PRE optimizations.
2000 2014-04-14 Richard Biener <rguenther@suse.de>
2002 * passes.def: Move early points-to after early SRA.
2004 2014-04-14 Richard Biener <rguenther@suse.de>
2006 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2007 check for which sign-changes we allow when forwarding
2008 a converted value into a switch.
2010 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2012 * stor-layout.c (place_field): Finalize non-constant offset for the
2015 2014-04-14 Richard Biener <rguenther@suse.de>
2017 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2019 (expand_switch_using_bit_tests_p): Likewise.
2020 (process_switch): Compute and pass on speed_p based on the
2022 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2023 optimize_bb_for_speed_p.
2025 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2027 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2028 * function.h (struct function): Rename has_force_vect_loops into
2029 has_force_vectorize_loops.
2030 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2031 (input_struct_function_base): Likewise.
2032 * lto-streamer-out.c (output_cfg): Likewise.
2033 (output_struct_function_base): Likewise.
2034 * omp-low.c (expand_omp_simd): Likewise.
2035 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2036 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2037 (version_loop_for_if_conversion): Likewise.
2038 (tree_if_conversion): Likewise.
2039 (main_tree_if_conversion): Likewise.
2040 (gate_tree_if_conversion): Likewise.
2041 * tree-inline.c (copy_loops): Likewise.
2042 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2043 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2044 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2045 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2046 * tree-vectorizer.c (vectorize_loops): Likewise.
2047 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2049 2014-04-14 Richard Biener <rguenther@suse.de>
2052 * lto-streamer-out.c (wrap_refs): New function.
2053 (lto_output): Wrap symbol references in global initializes in
2054 type-preserving MEM_REFs.
2056 2014-04-14 Christian Bruel <christian.bruel@st.com>
2058 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2060 2014-04-14 Christian Bruel <christian.bruel@st.com>
2062 * config/sh/sh.md (setmemqi): New expand pattern.
2063 * config/sh/sh.h (CLEAR_RATIO): Define.
2064 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2065 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2067 2014-04-14 Richard Biener <rguenther@suse.de>
2070 * fold-const.c (negate_expr_p): Don't negate directional rounding
2072 (fold_negate_expr): Likewise.
2074 2014-04-14 Richard Biener <rguenther@suse.de>
2076 PR tree-optimization/59817
2077 PR tree-optimization/60453
2078 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2079 recursion to catch all CHRECs in the scalar evolution and restrict
2080 the predicate for the remains appropriately.
2082 2014-04-12 Catherine Moore <clm@codesourcery.com>
2084 * config/mips/constraints.md: Add new register constraint "kb".
2085 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2086 (*movhi_internal): Likewise.
2087 (*movqi_internal): Likewise.
2088 * config/mips/mips.h (M16_STORE_REGS): New register class.
2089 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2090 (REG_CLASS_CONTENTS): Likewise.
2091 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2093 2014-04-11 Tobias Burnus <burnus@net-b.de>
2096 * doc/invoke.texi (-Wformat-signedness): Document it.
2097 (Wformat=2): Mention that this enables -Wformat-signedness.
2099 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2101 * common/config/epiphany/epiphany-common.c
2102 (epiphany_option_optimization_table): Enable section anchors by
2103 default at -O1 or higher.
2104 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2105 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2106 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2107 carries no extra cost.
2108 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2109 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2110 * config/epiphany/predicates.md (memclob_operand): New predicate.
2111 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2112 Use memclob_operand predicate and X constraint for operand 3.
2114 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2116 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2117 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2120 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2122 PR rtl-optimization/60651
2123 * mode-switching.c (optimize_mode_switching): Make sure to emit
2124 sets of a lower numbered entity before sets of a higher numbered
2125 entity to a mode of the same or lower priority.
2126 When creating a seginfo for a basic block that starts with a code
2127 label, move the insertion point past the code label.
2128 (new_seginfo): Document and enforce requirement that
2129 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2130 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2131 * doc/tm.texi: Regenerate.
2133 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2136 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2138 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2140 * BASE-VER: Set to 4.10.0.
2142 2014-04-11 Tobias Burnus <burnus@net-b.de>
2145 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2146 * doc/gcc.texi (Service): Update description in the @menu
2147 * doc/invoke.texi (Option Summary): Remove misplaced and
2150 2014-04-11 Steve Ellcey <sellcey@mips.com>
2151 Jakub Jelinek <jakub@redhat.com>
2154 * expr.c (convert_move): Use emit_store_flag_force instead of
2155 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2158 2014-04-11 Richard Biener <rguenther@suse.de>
2161 * varasm.c (assemble_alias): Avoid endless error reporting
2162 recursion by setting TREE_ASM_WRITTEN.
2164 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2166 * config/s390/s390.md: Add a splitter for NOT rtx.
2168 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2170 PR rtl-optimization/60663
2171 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2173 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2174 Jakub Jelinek <jakub@redhat.com>
2177 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2178 flag from decl_node to node.
2180 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2183 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2184 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2185 ameliorating the cases where it can be.
2187 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2190 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2192 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2193 (loadsync_<mode>): Change mode.
2194 (load_quadpti, store_quadpti): New.
2195 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2196 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2197 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2199 2014-04-09 Cong Hou <congh@google.com>
2202 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2205 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2207 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2208 instead of vnor to exploit possible fusion opportunity in the
2210 (altivec_expand_vec_perm_const_le): Likewise.
2212 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2214 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2215 (loadsync_<mode>): Change mode.
2216 (load_quadpti, store_quadpti): New.
2217 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2218 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2220 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2223 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2224 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2225 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2227 2014-04-08 Richard Biener <rguenther@suse.de>
2230 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2231 a 64bit widest int print double-int similar to on HWI64 hosts.
2233 2014-04-08 Richard Biener <rguenther@suse.de>
2235 PR tree-optimization/60785
2236 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2237 default defs properly.
2239 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2241 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2242 (Weffc++): Likewise.
2244 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2246 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2247 set completep to false rather than true.
2249 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2252 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2253 ARM_TARGET2_DWARF_FORMAT.
2255 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2258 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2259 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2262 2014-04-07 Richard Biener <rguenther@suse.de>
2264 PR tree-optimization/60766
2265 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2266 (may_eliminate_iv): Convert cand_value_at result to desired type.
2268 2014-04-07 Jason Merrill <jason@redhat.com>
2271 * common.opt (-fno-gnu-unique): Add.
2272 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2274 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2276 * haifa-sched.c: Fix outdated function reference and minor
2277 grammar errors in introductory comment.
2279 2014-04-07 Richard Biener <rguenther@suse.de>
2282 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2284 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2286 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2289 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2290 size accounting for thunks.
2291 (pa_asm_output_mi_thunk): Use final_start_function() and
2292 final_end_function() to output function start and end directives.
2294 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2296 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2297 device specific ISA/ feature information. Remove short_sp and
2298 errata_skip ds. Add avr_device_specific_features enum to have device
2300 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2301 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2302 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2303 updated device specific info.
2304 * config/avr/avr-mcus.def: Merge device specific details to
2305 dev_attribute field.
2306 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2308 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2309 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2310 assembler if RMW isa supported by current device.
2311 * config/avr/genmultilib.awk: Update as device info structure changed.
2312 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2314 2014-04-04 Cong Hou <congh@google.com>
2316 PR tree-optimization/60656
2317 * tree-vect-stmts.c (supportable_widening_operation):
2318 Fix a bug that elements in a vector with vect_used_by_reduction
2319 property are incorrectly reordered when the operation on it is not
2320 consistant with the one in reduction operation.
2322 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2324 PR rtl-optimization/60155
2325 * gcse.c (record_set_data): New function.
2326 (single_set_gcse): New function.
2327 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2328 (hoist_code): Likewise.
2329 (get_pressure_class_and_nregs): Likewise.
2331 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2333 * explow.c (probe_stack_range): Emit a final optimization blockage.
2335 2014-04-04 Anthony Green <green@moxielogic.com>
2337 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2340 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2343 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2344 flags are set only during streaming.
2345 * ipa.c (process_references, walk_polymorphic_call_targets,
2346 symtab_remove_unreachable_nodes): Drop bodies of always inline
2347 after early inlining.
2348 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2350 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2351 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2354 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2357 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2360 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2362 (cortex_a53_fdivd): Likewise.
2364 2014-04-04 Martin Jambor <mjambor@suse.cz>
2367 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2369 * cgraph.c (clone_of_p): Also return true if thunks match.
2370 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2371 cgraph_function_or_thunk_node and an obsolete comment.
2372 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2374 (build_function_decl_skip_args): Likewise.
2375 (set_new_clone_decl_and_node_flags): New function.
2376 (duplicate_thunk_for_node): Likewise.
2377 (redirect_edge_duplicating_thunks): Likewise.
2378 (cgraph_clone_node): New parameter args_to_skip, pass it to
2379 redirect_edge_duplicating_thunks which is called instead of
2380 cgraph_redirect_edge_callee.
2381 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2382 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2384 2014-04-04 Jeff Law <law@redhat.com>
2387 * config/arm/predicates.md (const_int_I_operand): New predicate.
2388 (const_int_M_operand): Similarly.
2389 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2391 (insv_t2, extv_reg, extzv_t2): Likewise.
2392 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2393 (pop_multiple_with_writeback_and_return): Likewise.
2394 (vfp_pop_multiple_with_writeback): Likewise
2396 2014-04-04 Richard Biener <rguenther@suse.de>
2399 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2400 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2402 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2403 * gimplify.c (gimple_add_tmp_var_fn): New function.
2404 * gimple-expr.h (create_tmp_reg_fn): Declare.
2405 * gimple-expr.c (create_tmp_reg_fn): New function.
2406 * gimple-low.c (record_vars_into): Don't change cfun.
2407 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2408 code generation without cfun.
2410 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2413 * Makefile.in (install-driver): Fix shell scripting.
2415 2014-04-03 Cong Hou <congh@google.com>
2417 PR tree-optimization/60505
2418 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2419 threshold of number of iterations below which no vectorization
2421 * tree-vect-loop.c (new_loop_vec_info):
2422 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2423 * tree-vect-loop.c (vect_analyze_loop_operations):
2424 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2425 * tree-vect-loop.c (vect_transform_loop):
2426 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2427 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2428 of iterations of the loop and see if we should build the epilogue.
2430 2014-04-03 Richard Biener <rguenther@suse.de>
2432 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2433 (streamer_tree_cache_create): Adjust.
2434 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2435 to allow optional nodes array.
2436 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2437 (streamer_tree_cache_append): Likewise.
2438 (streamer_tree_cache_create): Create nodes array optionally
2439 as specified by parameter.
2440 * lto-streamer-out.c (create_output_block): Avoid maintaining
2441 the node array in the writer cache.
2442 (DFS_write_tree): Remove assertion.
2443 (produce_asm_for_decls): Free the out decl state hash table early.
2444 * lto-streamer-in.c (lto_data_in_create): Adjust for
2445 streamer_tree_cache_create prototype change.
2447 2014-04-03 Richard Biener <rguenther@suse.de>
2449 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2450 set TREE_CHAIN to NULL_TREE.
2452 2014-04-03 Richard Biener <rguenther@suse.de>
2454 PR tree-optimization/60740
2455 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2456 over all GIMPLE_COND operands.
2458 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2460 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2461 (Weffc++): Remove Scott's numbering, merge lists and reference
2464 2014-04-03 Nick Clifton <nickc@redhat.com>
2466 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2469 2014-04-03 Martin Jambor <mjambor@suse.cz>
2471 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2472 mention gcc_unreachable before failing.
2473 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2476 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2479 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2480 inconsistent code and instead mark the context inconsistent.
2481 (possible_polymorphic_call_targets): For inconsistent contexts
2482 return empty complete list.
2484 2014-04-02 Anthony Green <green@moxielogic.com>
2486 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2487 (extendqisi2, extendhisi2): Define.
2488 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2489 (WCHAR_TYPE): Change to unsigned int.
2491 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2493 PR tree-optimization/60733
2494 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2495 insertion point for PHI candidates to be the end of the feeding
2496 block for the PHI argument.
2498 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2500 PR rtl-optimization/60650
2501 * lra-constraints.c (process_alt_operands): Decrease reject for
2502 earlyclobber matching.
2504 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2506 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2508 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2510 * config/spu/spu.c (pad_bb): Do not crash when the last
2511 insn is CODE_FOR_blockage.
2513 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2515 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2516 lies outside the target mode.
2518 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2521 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2522 software floating point or no floating point registers, do not
2523 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2524 in GPRs that occurs after we tested for GPRs that would never be
2527 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2528 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2529 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2530 specifically allow DDmode, since that does not use the SPE SIMD
2533 2014-04-02 Richard Biener <rguenther@suse.de>
2536 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2537 MODE_INTs. Properly use negv_optab.
2538 (expand_abs): Likewise.
2540 2014-04-02 Richard Biener <rguenther@suse.de>
2543 * Makefile.in (install-driver): Guard extra installs with special
2546 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2548 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2551 2014-04-01 Richard Henderson <rth@redhat.com>
2554 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2555 alternative enabled before register allocation.
2557 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2559 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2560 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2562 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2563 (nios2_got_address): Update nios2_large_got_address call site.
2564 (nios2_delegitimize_address): New function.
2565 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2566 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2567 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2569 2014-04-01 Martin Husemann <martin@duskware.de>
2571 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2574 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2576 PR rtl-optimization/60604
2577 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2578 check from register_operand.
2579 (register_operand): Redefine in terms of general_operand.
2580 (nonmemory_operand): Use register_operand for the non-constant cases.
2582 2014-04-01 Richard Biener <rguenther@suse.de>
2584 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2586 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2588 * doc/invoke.texi (mapp-regs): Clarify.
2590 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2592 * config/i386/avx512fintrin.h (__v32hi): Define type.
2593 (__v64qi): Likewise.
2594 (_mm512_set1_epi8): Define.
2595 (_mm512_set1_epi16): Define.
2596 (_mm512_set4_epi32): Define.
2597 (_mm512_set4_epi64): Define.
2598 (_mm512_set4_pd): Define.
2599 (_mm512_set4_ps): Define.
2600 (_mm512_setr4_epi64): Define.
2601 (_mm512_setr4_epi32): Define.
2602 (_mm512_setr4_pd): Define.
2603 (_mm512_setr4_ps): Define.
2604 (_mm512_setzero_epi32): Define.
2606 2014-03-31 Martin Jambor <mjambor@suse.cz>
2609 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2610 callsite_arguments_match_p. Updated all callers. Also check types of
2611 corresponding formal parameters and actual arguments.
2612 (not_all_callers_have_enough_arguments_p) Renamed to
2613 some_callers_have_mismatched_arguments_p.
2615 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2617 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2619 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2622 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2625 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2627 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2628 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2630 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2631 Use FMAMODE_NOVF512 mode iterator.
2632 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2633 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2634 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2635 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2637 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2638 Use VF_128_256 mode iterator.
2639 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2642 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2644 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2645 static chain if needed.
2647 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2650 * lra-constraints.c (index_part_to_reg): New.
2651 (process_address): Use it.
2653 2014-03-27 Jeff Law <law@redhat.com>
2654 Jakub Jelinek <jakub@redhat.com>
2657 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2658 gen_rtx_{PLUS,MULT} to build up the address expression.
2660 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2661 creating non-canonical RTL.
2663 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2666 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2667 functions; reorganize to make cheap checks first.
2668 (inline_small_functions): Do not estimate growth when dumping;
2670 * ipa-inline.h (inline_summary): Add min_size.
2671 (growth_likely_positive): New function.
2672 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2673 (set_cond_stmt_execution_predicate): Cleanup.
2674 (estimate_edge_size_and_time): Compute min_size.
2675 (estimate_calls_size_and_time): Likewise.
2676 (estimate_node_size_and_time): Likewise.
2677 (inline_update_overall_summary): Update min_size.
2678 (do_estimate_edge_time): Likewise.
2679 (do_estimate_edge_size): Update.
2680 (do_estimate_edge_hints): Update.
2681 (growth_likely_positive): New function.
2683 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2686 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2687 also if addr has VOIDmode.
2689 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2691 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2692 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2694 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2695 instructions as well as AdvancedSIMD loads.
2697 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2699 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2700 Use crypto_aese type.
2701 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2702 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2703 crypto_aese, crypto_aesmc. Move to types.md.
2704 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2706 * config/arm/iterators.md (crypto_type): Likewise.
2708 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2710 * cgraph.c: Include expr.h and tree-dfa.h.
2711 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2714 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2717 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2718 regs from checking multi-reg pseudos.
2720 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2722 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2724 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2726 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2727 if it would clobber the stack pointer, even temporarily.
2729 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2731 * mode-switching.c: Make small adjustments to the top comment.
2733 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2735 * config/rs6000/constraints.md (wD constraint): New constraint to
2736 match the constant integer to get the top DImode/DFmode out of a
2737 vector in a VSX register.
2739 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2740 match the constant integer to get the top DImode/DFmode out of a
2741 vector in a VSX register.
2743 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2746 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2749 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2750 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2752 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2753 Optimize vec_extract of 64-bit values, where the value being
2754 extracted is in the top word, where we can use scalar
2755 instructions. Add direct move and store support. Combine the big
2756 endian/little endian vector select load support into a single insn.
2757 (vsx_extract_<mode>_internal1): Likewise.
2758 (vsx_extract_<mode>_internal2): Likewise.
2759 (vsx_extract_<mode>_load): Likewise.
2760 (vsx_extract_<mode>_store): Likewise.
2761 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2762 combined into vsx_extract_<mode>_load.
2763 (vsx_extract_<mode>_one_le): Likewise.
2765 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2766 define the top 64-bit vector element.
2768 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2771 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2772 Document vec_vbpermq builtin.
2775 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2776 enable use of xxsldwi and xxpermdi builtin functions.
2777 (vec_xxpermdi): Likewise.
2779 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2780 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2782 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2784 PR rtl-optimization/60650
2785 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2787 (find_spills_for): New.
2788 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2789 Spill all pseudos on the second iteration.
2791 2014-03-27 Marek Polacek <polacek@redhat.com>
2794 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2797 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2799 * config/s390/s390.c (s390_can_use_return_insn): Check for
2800 call-saved FPRs on 31 bit.
2802 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2805 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2806 if they need regimplification, just drop them instead of
2807 calling gimple_regimplify_operands on them.
2809 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2812 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
2813 (aarch64_frame_pointer_required): Adjust logic.
2814 (aarch64_can_eliminate): Adjust logic.
2815 (aarch64_override_options_after_change): Adjust logic.
2817 2014-03-27 Dehao Chen <dehao@google.com>
2819 * ipa-inline.c (early_inliner): Update node's inline info.
2821 2014-03-26 Dehao Chen <dehao@google.com>
2823 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
2824 compiler inserted conditional jumps for NAN float check.
2826 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2828 * ubsan.h (ubsan_create_data): Change second argument's type
2829 to const location_t *.
2830 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
2832 (ubsan_create_data): Change second argument to const location_t *PLOC.
2833 Create Loc field whenever PLOC is non-NULL.
2834 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
2835 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
2839 * real.c (real_to_integer2): Change type of low to UHWI.
2841 2014-03-26 Tobias Burnus <burnus@net-b.de>
2843 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
2844 (CILK_SELF_SPECS): New define.
2845 (driver_self_specs): Use it.
2847 2014-03-26 Richard Biener <rguenther@suse.de>
2849 * tree-pretty-print.c (percent_K_format): Implement special
2850 case for LTO and its stripped down BLOCK tree.
2852 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2855 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
2857 * tree-vrp.c (simplify_internal_call_using_ranges): If only
2858 one range is range_int_cst_p, but not both, at least optimize
2859 addition/subtraction of 0 and multiplication by 0 or 1.
2860 * gimple-fold.c (gimple_fold_call): Fold
2861 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
2862 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
2863 INTEGER_CSTs, try to fold at least x * 0 and y - y.
2865 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
2867 PR rtl-optimization/60452
2868 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
2869 <case REG>: Return 1 for invalid offsets from the frame pointer.
2871 2014-03-26 Marek Polacek <polacek@redhat.com>
2874 * doc/extend.texi (C Extensions): Mention variable-length arrays in
2877 2014-03-26 Marek Polacek <polacek@redhat.com>
2880 * doc/extend.texi (Designated Inits): Describe what happens to omitted
2883 2014-03-26 Marek Polacek <polacek@redhat.com>
2886 * ira-color.c (update_conflict_hard_regno_costs): Perform the
2887 multiplication in unsigned type.
2889 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2891 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
2893 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2895 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
2897 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2900 * cif-code.def (UNREACHABLE) New code.
2901 * ipa-inline.c (inline_small_functions): Skip edges to
2902 __builtlin_unreachable.
2903 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
2904 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
2905 predicate to __bulitin_unreachable.
2906 (set_cond_stmt_execution_predicate): Fix issue when
2907 invert_tree_comparison returns ERROR_MARK.
2908 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
2909 propagate to inline clones.
2910 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
2912 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
2913 * cgraphclones.c (cgraph_clone_node): If call destination is already
2914 ureachable, do not redirect it back.
2915 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
2918 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2920 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
2921 Do not modify inline clones.
2923 2014-03-25 Jakub Jelinek <jakub@redhat.com>
2925 * config/i386/i386.md (general_sext_operand): New mode attr.
2926 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
2927 don't generate (sign_extend (const_int)).
2928 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
2929 operands[2]. Use We constraint instead of <i> and
2930 <general_sext_operand> predicate instead of <general_operand>.
2931 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
2932 * config/i386/constraints.md (We): New constraint.
2933 * config/i386/predicates.md (x86_64_sext_operand,
2934 sext_operand): New predicates.
2936 2014-03-25 Martin Jambor <mjambor@suse.cz>
2939 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
2940 inconsistent devirtualizations to __builtin_unreachable.
2942 2014-03-25 Marek Polacek <polacek@redhat.com>
2945 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
2947 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
2949 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
2950 order of elements for big-endian.
2952 2014-03-25 Richard Biener <rguenther@suse.de>
2955 * gimplify-me.c (gimple_regimplify_operands): Update the
2958 2014-03-25 Martin Jambor <mjambor@suse.cz>
2961 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
2962 (lto_output_varpool_node): Likewise.
2963 (input_overwrite_node): Likewise.
2964 (input_varpool_node): Likewise.
2966 2014-03-25 Richard Biener <rguenther@suse.de>
2968 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
2969 (run_gcc): Likewise.
2971 2014-03-25 Jakub Jelinek <jakub@redhat.com>
2973 * combine.c (simplify_compare_const): Add MODE argument.
2974 Handle mode_width 0 as very large mode_width.
2975 (try_combine, simplify_comparison): Adjust callers.
2977 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
2978 type to avoid signed integer overflow.
2979 * explow.c (plus_constant): Likewise.
2981 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
2983 * doc/generic.texi: Correct typos.
2985 2014-03-24 Tobias Burnus <burnus@net-b.de>
2987 * doc/invoke.texi (-flto): Expand section about
2988 using static libraries with LTO.
2990 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2992 PR rtl-optimization/60501
2993 * optabs.def (addptr3_optab): New optab.
2994 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
2995 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
2996 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
2998 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3000 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3002 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3004 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3007 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3008 (_mm256_undefined_ps): Define.
3009 (_mm256_undefined_pd): Define.
3010 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3011 (_mm_undefined_pd): Define.
3012 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3013 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3014 (_mm512_undefined_ps): Define.
3015 (_mm512_undefined_pd): Define.
3016 Use _mm*_undefined_*.
3017 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3019 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3021 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3022 (lshr_simd): DI mode added.
3023 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3024 (aarch64_ushr_simddi): Likewise.
3025 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3026 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3027 (vshrd_n_u64): Likewise.
3029 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3031 * Makefile.in (s-macro_list): Depend on cc1.
3033 2014-03-23 Teresa Johnson <tejohnson@google.com>
3035 * ipa-utils.c (ipa_print_order): Use specified dump file.
3037 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3039 PR rtl-optimization/60601
3040 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3042 * gcc.c (eval_spec_function): Initialize save_growing_value.
3044 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3047 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3048 code == MINUS_EXPR, never swap op0 with op1.
3050 * toplev.c (init_local_tick): Avoid signed integer multiplication
3052 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3053 shift by first operand's bitsize.
3055 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3058 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3060 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3061 TARGET_ISA_64BIT_P(x).
3063 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3065 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3066 pattern for vector nor instead of subtract from splat(-1).
3067 (altivec_expand_vec_perm_const_le): Likewise.
3069 2014-03-21 Richard Henderson <rth@twiddle.net>
3072 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3073 related insns after epilogue_completed.
3075 2014-03-21 Martin Jambor <mjambor@suse.cz>
3078 * cgraph.h (symtab_node): New flag body_removed.
3079 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3080 when removing bodies.
3081 * symtab.c (dump_symtab_base): Dump body_removed flag.
3082 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3083 had their bodies removed.
3085 2014-03-21 Martin Jambor <mjambor@suse.cz>
3088 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3091 2014-03-21 Richard Biener <rguenther@suse.de>
3093 PR tree-optimization/60577
3094 * tree-core.h (struct tree_base): Document nothrow_flag use
3095 in VAR_DECL_NONALIASED.
3096 * tree.h (VAR_DECL_NONALIASED): New.
3097 (may_be_aliased): Adjust.
3098 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3100 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3102 * expr.c (expand_expr_real_1): Remove outdated comment.
3104 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3107 * ira.c (adjust_cleared_regs): Call copy_rtx on
3108 *reg_equiv[REGNO (loc)].src_p before passing it to
3109 simplify_replace_fn_rtx.
3112 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3113 into CONST, put pic register as first operand of PLUS. Use
3114 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3116 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3118 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3120 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3122 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3123 around for store forwarding issue in the FPU on the UT699.
3124 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3125 loads and operations if -mfix-ut699 is specified.
3126 (divtf3_hq): Tweak attribute.
3127 (sqrttf2_hq): Likewise.
3129 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3131 * calls.c (store_one_arg): Remove incorrect const qualification on the
3132 type of the temporary.
3133 * cfgexpand.c (expand_return): Likewise.
3134 * expr.c (expand_constructor): Likewise.
3135 (expand_expr_real_1): Likewise.
3137 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3139 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3142 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3145 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3147 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3149 * config/arm/aarch-common-protos.h
3150 (alu_cost_table): Fix spelling of "extend".
3151 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3153 2014-03-19 Richard Biener <rguenther@suse.de>
3156 * tree-core.h (tree_type_common): Re-order pointer members
3157 to reduce recursion depth during GC walks.
3159 2014-03-19 Marek Polacek <polacek@redhat.com>
3162 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3163 before accessing it.
3165 2014-03-19 Richard Biener <rguenther@suse.de>
3168 * lto-streamer-in.c (input_function): In WPA stage do not drop
3171 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3173 PR tree-optimization/60559
3174 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3175 with build_zero_cst assignment.
3177 2014-03-18 Kai Tietz <ktietz@redhat.com>
3179 PR rtl-optimization/56356
3180 * sdbout.c (sdbout_parms): Verify that parms'
3181 incoming argument is valid.
3182 (sdbout_reg_parms): Likewise.
3184 2014-03-18 Richard Henderson <rth@redhat.com>
3187 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3188 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3189 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3191 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3193 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3194 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3195 Italicize plugin event names in description. Explain that
3196 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3197 Remind that no GCC functions should be called after PLUGIN_FINISH.
3198 Explain what pragmas with expansion are.
3200 2014-03-18 Martin Liska <mliska@suse.cz>
3202 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3203 gimple call statement is update.
3204 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3205 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3207 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3210 * ubsan.c (ubsan_instrument_unreachable): Call
3211 initialize_sanitizer_builtins.
3212 (ubsan_pass): Likewise.
3215 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3216 varpool_finalize_decl instead of rest_of_decl_compilation.
3218 2014-03-18 Richard Biener <rguenther@suse.de>
3220 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3221 by using bitmap_and_compl instead of bitmap_and_compl_into.
3222 (df_rd_transfer_function): Likewise.
3224 2014-03-18 Richard Biener <rguenther@suse.de>
3226 * doc/lto.texi (fresolution): Fix typo.
3228 2014-03-18 Richard Biener <rguenther@suse.de>
3230 * doc/invoke.texi (flto): Update for changes in 4.9.
3232 2014-03-18 Richard Biener <rguenther@suse.de>
3234 * doc/loop.texi: Remove section on the removed lambda framework.
3235 Update loop docs with recent changes in preserving loop structure.
3237 2014-03-18 Richard Biener <rguenther@suse.de>
3239 * doc/lto.texi (-fresolution): Document.
3241 2014-03-18 Richard Biener <rguenther@suse.de>
3243 * doc/contrib.texi: Adjust my name.
3245 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3248 * internal-fn.c: Include diagnostic-core.h.
3249 (expand_BUILTIN_EXPECT): New function.
3250 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3251 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3252 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3253 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3255 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3256 Revert 3 argument __builtin_expect code.
3257 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3258 * gimple-fold.c (gimple_fold_call): Likewise.
3259 * tree.h (fold_builtin_expect): New prototype.
3260 * builtins.c (build_builtin_expect_predicate): Add predictor
3261 argument, if non-NULL, create 3 argument __builtin_expect.
3262 (fold_builtin_expect): No longer static. Add ARG2 argument,
3263 pass it through to build_builtin_expect_predicate.
3264 (fold_builtin_2): Adjust caller.
3265 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3266 * internal-fn.def (BUILTIN_EXPECT): New.
3268 2014-03-18 Tobias Burnus <burnus@net-b.de>
3271 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3272 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3273 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3275 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3278 * predict.c (combine_predictions_for_bb): Fix up formatting.
3279 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3280 fill what it points to if non-NULL.
3281 (tree_predict_by_opcode): Adjust caller, use the predictor.
3282 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3284 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3286 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3287 proper constant for the store mode.
3289 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3291 * symtab.c (change_decl_assembler_name): Fix transparent alias
3294 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3296 * config/aarch64/aarch64.c: Correct the comments about the
3297 aarch64 stack layout.
3299 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3301 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3302 check for GF_OMP_FOR_KIND_FOR.
3304 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3306 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3307 ymm and zmm register names.
3309 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3312 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3313 note creation for the 2010-08-31 changes.
3315 2014-03-17 Marek Polacek <polacek@redhat.com>
3318 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3319 as -fno-tree-loop-vectorize.
3320 (expand_omp_simd): Likewise.
3322 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3324 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3325 (eligible_for_call_delay): New prototype.
3326 * config/sparc/sparc.c (tls_call_delay): Rename into...
3327 (eligible_for_call_delay): ...this. Return false if the instruction
3328 cannot be put in the delay slot of a branch.
3329 (eligible_for_restore_insn): Simplify.
3330 (eligible_for_return_delay): Return false if the instruction cannot be
3331 put in the delay slot of a branch and simplify.
3332 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3333 put in the delay slot of a branch.
3334 * config/sparc/sparc.md (fix_ut699): New attribute.
3335 (tls_call_delay): Delete.
3336 (in_call_delay): Reimplement.
3337 (eligible_for_sibcall_delay): Rename into...
3338 (in_sibcall_delay): ...this.
3339 (eligible_for_return_delay): Rename into...
3340 (in_return_delay): ...this.
3341 (in_branch_delay): Reimplement.
3342 (in_uncond_branch_delay): Delete.
3343 (in_annul_branch_delay): Delete.
3345 2014-03-14 Richard Henderson <rth@redhat.com>
3348 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3349 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3350 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3351 (floathi splitters): Remove.
3352 (float<SWI48x>xf2): New pattern.
3353 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3354 code that tried to handle DImode for 32-bit, but which was excluded
3355 by the pattern's condition. Drop allocation of stack temporary.
3356 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3357 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3358 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3359 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3360 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3361 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3362 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3363 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3364 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3365 (*float<SWI48x><X87MODEF>2_i387): Remove.
3366 (all float _with_temp splitters): Remove.
3367 (*float<SWI48x><MODEF>2_i387): New pattern.
3368 (*float<SWI48><MODEF>2_sse): New pattern.
3369 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3370 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3372 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3373 Marek Polacek <polacek@redhat.com>
3376 * common.opt (dump_base_name_prefixed): New Variable.
3377 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3378 if x_dump_base_name_prefixed is already set, set it at the end.
3380 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3382 PR rtl-optimization/60508
3383 * lra-constraints.c (get_reload_reg): Add new parameter
3385 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3386 Pass the new parameter values.
3388 2014-03-14 Richard Biener <rguenther@suse.de>
3390 * common.opt: Revert unintented changes from r205065.
3393 2014-03-14 Richard Biener <rguenther@suse.de>
3396 * cfghooks.c (split_block): Properly adjust all loops the
3397 block was a latch of.
3399 2014-03-14 Martin Jambor <mjambor@suse.cz>
3402 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3405 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3408 * config/avr/avr.c (avr_set_current_function): Pass function name
3409 through default_strip_name_encoding before sanity checking instead
3410 of skipping the first char of the assembler name.
3412 2014-03-13 Richard Henderson <rth@redhat.com>
3415 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3416 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3417 * config/i386/i386-protos.h: Likewise.
3418 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3419 in the expander instead of a splitter.
3420 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3421 any possibility of requiring a memory.
3422 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3423 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3424 (fp branch splitters): Update for ix86_split_fp_branch.
3425 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3426 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3427 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3428 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3429 (*fop_<MODEF>_3_i387): Likewise.
3430 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3431 (splitters for the fop_* register patterns): Remove.
3432 (fscalexf4_i387): Rename from *fscalexf4_i387.
3433 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3435 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3437 PR tree-optimization/59779
3438 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3439 type for bitsize and maxsize instead of HOST_WIDE_INT.
3441 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3443 PR rtl-optimization/57320
3444 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3445 the CFG after thread_prologue_and_epilogue_insns.
3447 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3449 PR rtl-optimization/57189
3450 * lra-constraints.c (process_alt_operands): Disfavor spilling
3453 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3455 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3457 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3459 PR tree-optimization/59025
3461 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3462 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3464 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3467 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3468 calls of avr_out_plus_1.
3470 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3472 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3473 BB's single pred and update the father loop's latch info later.
3475 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3477 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3481 (VEC_base): Likewise.
3482 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3483 registers, we need to swap double words in little endian mode.
3485 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3486 to be a container mode for 128-bit integer operations added in ISA
3487 2.07. Unlike TImode and PTImode, the preferred register set is
3488 the Altivec/VMX registers for the 128-bit operations.
3490 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3492 (rs6000_split_128bit_ok_p): Likewise.
3494 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3495 macros for creating ISA 2.07 normal and overloaded builtin
3496 functions with 3 arguments.
3497 (BU_P8V_OVERLOAD_3): Likewise.
3498 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3499 for use as overloaded functions.
3500 (VPERM_1TI_UNS): Likewise.
3501 (VSEL_1TI): Likewise.
3502 (VSEL_1TI_UNS): Likewise.
3503 (ST_INTERNAL_1ti): Likewise.
3504 (LD_INTERNAL_1ti): Likewise.
3505 (XXSEL_1TI): Likewise.
3506 (XXSEL_1TI_UNS): Likewise.
3507 (VPERM_1TI): Likewise.
3508 (VPERM_1TI_UNS): Likewise.
3509 (XXPERMDI_1TI): Likewise.
3510 (SET_1TI): Likewise.
3511 (LXVD2X_V1TI): Likewise.
3512 (STXVD2X_V1TI): Likewise.
3513 (VEC_INIT_V1TI): Likewise.
3514 (VEC_SET_V1TI): Likewise.
3515 (VEC_EXT_V1TI): Likewise.
3516 (EQV_V1TI): Likewise.
3517 (NAND_V1TI): Likewise.
3518 (ORC_V1TI): Likewise.
3519 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3520 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3522 (VADDUQM): Likewise.
3523 (VSUBCUQ): Likewise.
3524 (VADDEUQM): Likewise.
3525 (VADDECUQ): Likewise.
3526 (VSUBEUQM): Likewise.
3527 (VSUBECUQ): Likewise.
3529 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3530 __int128_t and __uint128_t types.
3531 (__uint128_type): Likewise.
3532 (altivec_categorize_keyword): Add support for vector __int128_t,
3533 vector __uint128_t, vector __int128, and vector unsigned __int128
3534 as a container type for TImode operations that need to be done in
3535 VSX/Altivec registers.
3536 (rs6000_macro_to_expand): Likewise.
3537 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3538 to support 128-bit integer instructions vaddcuq, vadduqm,
3539 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3540 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3542 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3543 for V1TImode, and set up preferences to use VSX/Altivec registers.
3544 Setup VSX reload handlers.
3545 (rs6000_debug_reg_global): Likewise.
3546 (rs6000_init_hard_regno_mode_ok): Likewise.
3547 (rs6000_preferred_simd_mode): Likewise.
3548 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3549 (easy_altivec_constant): Likewise.
3550 (output_vec_const_move): Likewise.
3551 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3553 (rs6000_expand_vector_extract): Likewise.
3554 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3556 (rs6000_const_vec): Add support for V1TImode.
3557 (rs6000_emit_le_vsx_load): Swap double words when loading or
3558 storing TImode/V1TImode.
3559 (rs6000_emit_le_vsx_store): Likewise.
3560 (rs6000_emit_le_vsx_move): Likewise.
3561 (rs6000_emit_move): Add support for V1TImode.
3562 (altivec_expand_ld_builtin): Likewise.
3563 (altivec_expand_st_builtin): Likewise.
3564 (altivec_expand_vec_init_builtin): Likewise.
3565 (altivec_expand_builtin): Likewise.
3566 (rs6000_init_builtins): Add support for V1TImode type. Add
3567 support for ISA 2.07 128-bit integer builtins. Define type names
3568 for the VSX/Altivec vector types.
3569 (altivec_init_builtins): Add support for overloaded vector
3570 functions with V1TImode type.
3571 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3572 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3574 (rs6000_split_128bit_ok_p): Likewise.
3575 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3576 __int128_t and vector __uint128_t.
3578 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3579 and mode attributes.
3586 (VS_scalar): Likewise.
3587 (VS_double): Likewise.
3588 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3590 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3591 we support the ISA 2.07 128-bit integer arithmetic instructions.
3592 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3593 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3594 and TImode types for use with the builtin functions.
3595 (V1TI_type_node): Likewise.
3596 (unsigned_V1TI_type_node): Likewise.
3597 (intTI_type_internal_node): Likewise.
3598 (uintTI_type_internal_node): Likewise.
3600 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3601 128-bit builtin functions.
3602 (UNSPEC_VADDEUQM): Likewise.
3603 (UNSPEC_VADDECUQ): Likewise.
3604 (UNSPEC_VSUBCUQ): Likewise.
3605 (UNSPEC_VSUBEUQM): Likewise.
3606 (UNSPEC_VSUBECUQ): Likewise.
3607 (VM): Add V1TImode to vector mode iterators.
3609 (VI_unit): Likewise.
3610 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3611 (altivec_vaddcuq): Likewise.
3612 (altivec_vsubuqm): Likewise.
3613 (altivec_vsubcuq): Likewise.
3614 (altivec_vaddeuqm): Likewise.
3615 (altivec_vaddecuq): Likewise.
3616 (altivec_vsubeuqm): Likewise.
3617 (altivec_vsubecuq): Likewise.
3619 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3621 (BOOL_128): Likewise.
3622 (BOOL_REGS_OUTPUT): Likewise.
3623 (BOOL_REGS_OP1): Likewise.
3624 (BOOL_REGS_OP2): Likewise.
3625 (BOOL_REGS_UNARY): Likewise.
3626 (BOOL_REGS_AND_CR0): Likewise.
3628 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3629 128-bit integer builtin support.
3630 (vec_vadduqm): Likewise.
3631 (vec_vaddecuq): Likewise.
3632 (vec_vaddeuqm): Likewise.
3633 (vec_vsubecuq): Likewise.
3634 (vec_vsubeuqm): Likewise.
3635 (vec_vsubcuq): Likewise.
3636 (vec_vsubuqm): Likewise.
3638 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3639 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3640 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3641 128-bit integer add/subtract to ISA 2.07.
3643 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3645 * config/arc/arc.c (arc_predicate_delay_insns):
3646 Fix third argument passed to conditionalize_nonjump.
3648 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3650 * config/aarch64/aarch64-builtins.c
3651 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3652 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3653 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3654 instead of __builtin_lfloor.
3655 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3657 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3659 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3660 (tree_ssa_ifcombine_bb_1): New function.
3661 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3662 is an empty forwarder block to then_bb or vice versa and then_bb
3663 and else_bb are effectively swapped.
3665 2014-03-12 Christian Bruel <christian.bruel@st.com>
3668 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3669 REG_CFA_DEF_CFA note.
3670 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3671 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3673 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3675 PR tree-optimization/60454
3676 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3678 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3680 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3681 Do not define target_cpu_default2 to generic.
3682 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3683 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3684 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3686 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3687 Marc Glisse <marc.glisse@inria.fr>
3689 PR tree-optimization/60502
3690 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3691 instead of build_low_bits_mask.
3693 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3696 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3697 if there are multiple uses, but op doesn't live on E edge.
3698 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3699 clobber stmts before __builtin_unreachable.
3701 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3703 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3704 hard_frame_pointer_rtx.
3705 * cse.c (cse_insn): Remove volatile check.
3706 * cselib.c (cselib_process_insn): Likewise.
3707 * dse.c (scan_insn): Likewise.
3709 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3711 * config/arc/arc.c (conditionalize_nonjump): New function,
3713 (arc_ifcvt): ... this.
3714 (arc_predicate_delay_insns): Use it.
3716 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3718 * config/arc/predicates.md (extend_operand): During/after reload,
3719 allow const_int_operand.
3720 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3721 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3722 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3724 (umulsi3_highpart_i): Likewise.
3726 2014-03-11 Richard Biener <rguenther@suse.de>
3728 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3729 Add asserts to guard possible wrong-code bugs.
3731 2014-03-11 Richard Biener <rguenther@suse.de>
3733 PR tree-optimization/60429
3734 PR tree-optimization/60485
3735 * tree-ssa-structalias.c (set_union_with_increment): Properly
3736 take into account all fields that overlap the shifted vars.
3737 (do_sd_constraint): Likewise.
3738 (do_ds_constraint): Likewise.
3739 (get_constraint_for_ptr_offset): Likewise.
3741 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3743 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3744 (nios2_compute_frame_layout):
3745 Add calculation of cfun->machine->fp_save_offset.
3746 (nios2_expand_prologue): Correct setting of frame pointer register
3748 (nios2_expand_epilogue): Update recovery of stack pointer from
3749 frame pointer accordingly.
3750 (nios2_initial_elimination_offset): Update calculation of offset
3751 for eliminating to HARD_FRAME_POINTER_REGNUM.
3753 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3756 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3757 cgraph_get_create_node on VAR_DECLs.
3759 2014-03-10 Richard Biener <rguenther@suse.de>
3762 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3764 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3766 * config/vms/vms.opt (vms_float_format): New variable.
3768 2014-03-08 Tobias Burnus <burnus@net-b.de>
3770 * doc/invoke.texi (-fcilkplus): Update implementation status.
3772 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3773 Richard Biener <rguenther@suse.de>
3775 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3776 consistently accross all TUs.
3777 (run_gcc): Enable -fshort-double automatically at link at link-time
3778 and disallow override.
3780 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3783 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3784 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3785 if they can't be used.
3787 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3789 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3790 for Solaris 11/x86 ld.
3791 * configure: Regenerate.
3793 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3795 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3796 (LIB_TLS_SPEC): Save as ld_tls_libs.
3797 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3798 (HAVE_AS_IX86_TLSLDM): New test.
3799 * configure, config.in: Regenerate.
3800 * config/i386/i386.c (legitimize_tls_address): Fall back to
3801 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3802 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3803 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3804 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3806 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3808 * common.opt (fira-loop-pressure): Mark as optimization.
3810 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3812 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
3813 an OpenMP mappable type.
3815 2014-03-06 Matthias Klose <doko@ubuntu.com>
3817 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
3818 MULTILIB_OSDIRNAMES is not defined.
3820 2014-03-06 Jakub Jelinek <jakub@redhat.com>
3821 Meador Inge <meadori@codesourcery.com>
3824 * config/arm/arm.c (arm_tls_symbol_p): Remove.
3825 (arm_legitimize_address): Call legitimize_tls_address for any
3826 arm_tls_referenced_p expression, handle constant addend. Call it
3827 before testing for !TARGET_ARM.
3828 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
3830 2014-03-06 Richard Biener <rguenther@suse.de>
3836 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3838 * tree-streamer.c (record_common_node): Assert we don't record
3839 nodes with type double.
3840 (preload_common_node): Skip type double, complex double and double
3841 pointer since it is now frontend dependent due to fshort-double option.
3843 2014-03-06 Richard Biener <rguenther@suse.de>
3845 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
3846 or -fno-lto is specified and the linker has full plugin support.
3847 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
3848 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
3849 * lto-wrapper.c (merge_and_complain): Merge compile-time
3850 optimization levels.
3851 (run_gcc): And pass it through to the link options.
3853 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
3857 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3859 * cselib.c (remove_useless_values): Skip to avoid quadratic
3860 behavior if the condition moved from...
3861 (cselib_process_insn): ... here holds.
3863 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3866 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
3867 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
3870 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
3871 (TM_H): Add x86-tune.def.
3873 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3875 * config/aarch64/aarch64.c (generic_tunings):
3876 Use cortexa57_extra_costs.
3878 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3881 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
3882 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
3883 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
3886 2014-03-04 Heiher <r@hev.cc>
3888 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
3889 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
3891 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
3893 * config/i386/predicates.md (const2356_operand): Change to ...
3894 (const2367_operand): ... this.
3895 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
3897 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
3898 (*avx512pf_scatterpf<mode>sf): Ditto.
3899 (avx512pf_scatterpf<mode>df): Ditto.
3900 (*avx512pf_scatterpf<mode>df_mask): Ditto.
3901 (*avx512pf_scatterpf<mode>df): Ditto.
3902 * config/i386/i386.c (ix86_expand_builtin): Update
3903 incorrect hint operand error message.
3905 2014-03-04 Richard Biener <rguenther@suse.de>
3907 * lto-section-in.c (lto_get_section_data): Fix const cast.
3909 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3911 * tree-streamer.c (record_common_node): Assert we don't record
3912 nodes with type double.
3913 (preload_common_node): Skip type double, complex double and double
3914 pointer since it is now frontend dependent due to fshort-double option.
3916 2014-03-04 Richard Biener <rguenther@suse.de>
3919 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
3920 (lto_input_toplevel_asms): Likewise.
3921 * lto-section-in.c (lto_get_section_data): Instead do it here
3924 2014-03-04 Richard Biener <rguenther@suse.de>
3926 PR tree-optimization/60382
3927 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
3928 dead PHIs a reduction.
3930 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
3932 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
3934 (_mm_prefetch): Move out of GCC target("sse") pragma.
3935 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
3936 GCC target("prfchw") pragma.
3937 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
3939 * config/i386/i386.c (ix86_option_override_internal): Enable
3940 -mprfchw with -mprefetchwt1.
3942 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
3944 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
3947 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
3949 * opts.h (CL_PCH_IGNORE): Define.
3950 * targhooks.c (option_affects_pch_p):
3951 Return false for options that have CL_PCH_IGNORE set.
3952 * opt-functions.awk: Process PchIgnore.
3953 * doc/options.texi: Document PchIgnore.
3955 * config/arc/arc.opt (misize): Add PchIgnore property.
3957 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3959 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
3960 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
3961 constraint on constants to permit them being loaded into
3962 GENERAL_REGS or BASE_REGS.
3964 2014-03-03 Nick Clifton <nickc@redhat.com>
3966 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
3967 anti-cacnonical alternatives.
3968 (negandhi3_real): New pattern.
3969 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
3971 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
3973 * config/avr/avr-mcus.def: Remove atxmega16x1.
3974 * config/avr/avr-tables.opt: Regenerate.
3975 * config/avr/t-multilib: Regenerate.
3976 * doc/avr-mmcu.texi: Regenerate.
3978 2014-03-03 Tobias Grosser <tobias@grosser.es>
3979 Mircea Namolaru <mircea.namolaru@inria.fr>
3981 PR tree-optimization/58028
3982 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
3985 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3987 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
3988 not handled by recognizers.
3990 2014-03-03 Jakub Jelinek <jakub@redhat.com>
3993 * function.c (expand_function_end): Don't emit
3994 clobber_return_register sequence if clobber_after is a BARRIER.
3995 * cfgexpand.c (construct_exit_block): Append instructions before
3996 return_label to prev_bb.
3998 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4000 * config/rs6000/constraints.md: Document reserved use of "wc".
4002 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4005 * ipa.c (function_and_variable_visibility): When dissolving comdat
4006 group, also set all symbols to local.
4008 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4013 2013-12-14 Jan Hubicka <jh@suse.cz>
4015 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4017 2014-03-02 Jon Beniston <jon@beniston.com>
4023 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4024 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4025 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4026 (simple_return, *simple_return): New patterns
4027 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4028 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4030 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4032 * dwarf2out.c (gen_subprogram_die): Tidy.
4034 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4037 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4038 (*mov_t_msb_neg_negc): ... this new insn.
4040 2014-02-28 Jason Merrill <jason@redhat.com>
4043 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4046 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4049 * dwarf2out.c (decltype_auto_die): New static.
4050 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4051 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4052 (is_cxx_auto): Likewise.
4054 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4056 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4057 we are not using general regs only.
4059 2014-02-28 Richard Biener <rguenther@suse.de>
4062 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4063 previous fix and only allow to remove trivial pre-headers
4064 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4065 (remove_forwarder_block): Properly update the latch of a loop.
4067 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4070 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4071 (cselib_preserved_hash_table): New.
4072 (preserve_constants_and_equivs): Move preserved vals to it.
4073 (cselib_find_slot): Look it up first.
4074 (cselib_init): Initialize it.
4075 (cselib_finish): Release it.
4076 (dump_cselib_table): Dump it.
4078 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4081 * cselib.c (remove_useless_values): Skip to avoid quadratic
4082 behavior if the condition moved from...
4083 (cselib_process_insn): ... here holds.
4085 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4088 * var-tracking.c (vt_initialize): Apply the same condition to
4089 preserve the CFA base value.
4091 2014-02-28 Joey Ye <joey.ye@arm.com>
4094 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4095 if reload in progress or completed.
4097 2014-02-28 Tobias Burnus <burnus@net-b.de>
4100 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4103 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4105 * doc/tm.texi.in (Condition Code Status): Update documention for
4106 relative locations of cc0-setter and cc0-user.
4108 2014-02-27 Jeff Law <law@redhat.com>
4110 PR rtl-optimization/52714
4111 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4112 into two independent simple sets, if I3 is a jump, ensure the
4113 pattern we place into I3 is a (set (pc) ...).
4115 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4116 Jeff Law <law@redhat.com>
4118 PR rtl-optimization/49847
4119 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4120 are in different blocks.
4121 * doc/tm.texi (Condition Code Status): Update documention for
4122 relative locations of cc0-setter and cc0-user.
4124 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4127 * lra.c (lra_emit_add): Check SUBREG too.
4129 2014-02-27 Andreas Schwab <schwab@suse.de>
4131 * config/m68k/m68k.c (m68k_option_override): Disable
4132 -flive-range-shrinkage for classic m68k.
4133 (m68k_override_options_after_change): Likewise.
4135 2014-02-27 Marek Polacek <polacek@redhat.com>
4138 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4139 -Wmaybe-uninitialized.
4141 2014-02-27 Alan Modra <amodra@gmail.com>
4144 * reload1.c (emit_input_reload_insns): When reload_override_in,
4145 set old to rl->in_reg when rl->in_reg is a subreg.
4147 2014-02-26 Richard Biener <rguenther@suse.de>
4150 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4152 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4154 * common/config/i386/predicates.md (const1256_operand): Remove.
4155 (const2356_operand): New.
4156 (const_1_to_2_operand): Remove.
4157 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4158 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4159 (*avx512pf_gatherpf<mode>sf): Ditto.
4160 (avx512pf_gatherpf<mode>df): Ditto.
4161 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4162 (*avx512pf_gatherpf<mode>df): Ditto.
4163 (avx512pf_scatterpf<mode>sf): Ditto.
4164 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4165 (*avx512pf_scatterpf<mode>sf): Ditto.
4166 (avx512pf_scatterpf<mode>df): Ditto.
4167 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4168 (*avx512pf_scatterpf<mode>df): Ditto.
4169 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4171 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4173 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4174 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4175 (_mm512_mask_testn_epi64_mask): Move to ...
4176 * config/i386/avx512cdintrin.h: Here.
4177 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4178 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4179 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4180 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4181 TARGET_AVX512F from TARGET_AVX512CD.
4183 2014-02-26 Richard Biener <rguenther@suse.de>
4186 * ipa.c (walk_polymorphic_call_targets): Properly guard
4187 call to inline_update_overall_summary.
4189 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4192 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4193 and latches only if requested. Fix latch if it is removed.
4194 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4195 LOOPS_HAVE_PREHEADERS.
4197 2014-02-25 Andrew Pinski <apinski@cavium.com>
4199 * builtins.c (expand_builtin_thread_pointer): Create a new target
4200 when the target is NULL.
4202 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4204 PR rtl-optimization/60317
4205 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4206 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4207 * lra-assigns.c: Include params.h.
4208 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4209 other reload pseudos considerations.
4211 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4213 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4214 to use canonical form for nor<mode>3.
4216 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4219 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4222 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4224 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4225 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4226 (ix86_handle_option): Handle OPT_mprefetchwt1.
4227 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4228 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4230 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4231 OPTION_MASK_ISA_PREFETCHWT1.
4232 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4233 (PTA_PREFETCHWT1): New.
4234 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4235 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4236 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4237 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4238 (*prefetch_avx512pf_<mode>_: Change into ...
4239 (*prefetch_prefetchwt1_<mode>: This.
4240 * config/i386/i386.opt (mprefetchwt1): New.
4241 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4242 (_mm_prefetch): Handle intent to write.
4243 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4245 2014-02-25 Richard Biener <rguenther@suse.de>
4248 * emit-rtl.c (mem_attrs_htab): Remove.
4249 (mem_attrs_htab_hash): Likewise.
4250 (mem_attrs_htab_eq): Likewise.
4251 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4252 (init_emit_once): Do not allocate mem_attrs_htab.
4254 2014-02-25 Richard Biener <rguenther@suse.de>
4257 * lto-opts.c (lto_write_options): Output non-explicit conservative
4258 -fwrapv, -fno-trapv and -fno-strict-overflow.
4259 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4260 (run_gcc): And pass them through.
4262 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4264 * sel-sched.c (calculate_new_fences): New parameter ptime.
4265 Calculate it as a maximum over all fence cycles.
4266 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4267 Print the final schedule timing when sched_verbose.
4269 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4271 PR rtl-optimization/60292
4272 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4273 bit fot the fence instruction.
4275 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4277 * calls.h: Fix typo in comment.
4279 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4281 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4282 adjusting offsetable addresses.
4284 2014-02-24 Guozhi Wei <carrot@google.com>
4286 * sparseset.h (sparseset_pop): Fix the wrong index.
4288 2014-02-24 Walter Lee <walt@tilera.com>
4290 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4291 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4293 * common/config/tilegx/tilegx-common.c
4294 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4295 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4297 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4298 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4299 (tilegx_gimplify_va_arg_expr): Handle big endian.
4300 (tilegx_expand_unaligned_load): Ditto.
4301 (tilegx_expand_unaligned_store): Ditto.
4302 (TARGET_RETURN_IN_MSB): New.
4303 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4304 (TARGET_ENDIAN_DEFAULT): New.
4305 (TARGET_BIG_ENDIAN): Handle big endian.
4306 (BYTES_BIG_ENDIAN): Ditto.
4307 (WORDS_BIG_ENDIAN): Ditto.
4308 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4311 * config/tilegx/tilegx.md (extv): Handle big endian.
4313 (insn_st<n>): Ditto.
4314 (insn_st<n>_add<bitsuffix>): Ditto.
4315 (insn_stnt<n>): Ditto.
4316 (insn_stnt<n>_add<bitsuffix>):Ditto.
4317 (vec_interleave_highv8qi): Handle big endian.
4318 (vec_interleave_highv8qi_be): New.
4319 (vec_interleave_highv8qi_le): New.
4320 (insn_v1int_h): Handle big endian.
4321 (vec_interleave_lowv8qi): Handle big endian.
4322 (vec_interleave_lowv8qi_be): New.
4323 (vec_interleave_lowv8qi_le): New.
4324 (insn_v1int_l): Handle big endian.
4325 (vec_interleave_highv4hi): Handle big endian.
4326 (vec_interleave_highv4hi_be): New.
4327 (vec_interleave_highv4hi_le): New.
4328 (insn_v2int_h): Handle big endian.
4329 (vec_interleave_lowv4hi): Handle big endian.
4330 (vec_interleave_lowv4hi_be): New.
4331 (vec_interleave_lowv4hi_le): New.
4332 (insn_v2int_l): Handle big endian.
4333 (vec_interleave_highv2si): Handle big endian.
4334 (vec_interleave_highv2si_be): New.
4335 (vec_interleave_highv2si_le): New.
4336 (insn_v4int_h): Handle big endian.
4337 (vec_interleave_lowv2si): Handle big endian.
4338 (vec_interleave_lowv2si_be): New.
4339 (vec_interleave_lowv2si_le): New.
4340 (insn_v4int_l): Handle big endian.
4341 * config/tilegx/tilegx.opt (mbig-endian): New option.
4342 (mlittle-endian): New option.
4343 * doc/install.texi: Document tilegxbe-linux.
4344 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4346 2014-02-24 Martin Jambor <mjambor@suse.cz>
4349 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4350 there are no parameter descriptors.
4352 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4354 PR rtl-optimization/60268
4355 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4356 initialization to ...
4357 (sched_rgn_init): ... here.
4358 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4360 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4362 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4365 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4367 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4370 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4372 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4373 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4375 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4377 * config/microblaze/predicates.md: Add cmp_op predicate.
4378 * config/microblaze/microblaze.md: Add branch_compare instruction
4379 which uses cmp_op predicate and emits cmp insn before branch.
4380 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4381 to microblaze_expand_conditional_branch and consolidate logic.
4382 (microblaze_expand_conditional_branch): emit branch_compare
4383 insn instead of handling cmp op separate from branch insn.
4385 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4387 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4390 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4392 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4393 define_insn with define_expand and new define_insn
4394 *altivec_lve<VI_char>x_internal.
4395 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4396 and new define_insn *altivec_stve<VI_char>x_internal.
4397 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4399 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4401 (altivec_expand_stvex_be): New function.
4403 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4405 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4406 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4407 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4408 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4410 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4413 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4414 instead of emit_move_insn.
4416 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4418 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4420 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4421 gen_altivec_vsumsws.
4423 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4425 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4426 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4427 (altivec_lvxl_<mode>): New define_expand incorporating
4428 -maltivec=be semantics where needed.
4429 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4430 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4431 semantics where needed.
4432 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4433 (altivec_stvx_<mode>): New define_expand incorporating
4434 -maltivec=be semantics where needed.
4435 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4436 VM2 iterator instead of V4SI.
4437 (altivec_stvxl_<mode>): New define_expand incorporating
4438 -maltivec=be semantics where needed.
4439 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4440 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4441 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4442 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4443 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4444 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4445 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4446 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4447 ALTIVEC_BUILTIN_STVXL.
4448 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4449 (altivec_expand_stvx_be): Likewise.
4450 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4451 (altivec_expand_lvx_be): Likewise.
4452 (altivec_expand_stvx_be): Likewise.
4453 (altivec_expand_builtin): Add cases for
4454 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4455 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4456 (altivec_init_builtins): Add definitions for
4457 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4458 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4460 2014-02-21 Catherine Moore <clm@codesourcery.com>
4462 * doc/invoke.texi (mvirt, mno-virt): Document.
4463 * config/mips/mips.opt (mvirt): New option.
4464 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4466 2014-02-21 Richard Biener <rguenther@suse.de>
4468 PR tree-optimization/60276
4469 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4470 (STMT_VINFO_MIN_NEG_DIST): New macro.
4471 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4472 STMT_VINFO_MIN_NEG_DIST.
4473 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4474 made for negative dependence distances still hold.
4476 2014-02-21 Richard Biener <rguenther@suse.de>
4479 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4480 DECL_INITIAL for globals not in the current function context.
4482 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4484 PR tree-optimization/56490
4485 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4486 * tree-ssa-uninit.c: Include params.h.
4487 (compute_control_dep_chain): Add num_calls argument, return false
4488 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4489 num_calls to recursive call.
4490 (find_predicates): Change dep_chain into normal array,
4491 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4492 variable and adjust compute_control_dep_chain caller.
4493 (find_def_preds): Likewise.
4495 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4497 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4498 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4500 2014-02-21 Nick Clifton <nickc@redhat.com>
4502 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4503 (pushhi1): Likewise.
4504 (popqi1): Add mode to pre_dec.
4507 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4509 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4510 mode for mask of V8SFmode permutation.
4512 2014-02-20 Richard Henderson <rth@redhat.com>
4515 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4516 a new pseudo for OLDVAL.
4518 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4521 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4522 gen_reg_rtx if d->testing_p.
4523 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4524 if d->testing_p and we will certainly return true.
4525 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4528 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4530 * emit-rtl.c (gen_reg_rtx): Assert that
4531 crtl->emit.regno_pointer_align_length is non-zero.
4533 2014-02-20 Richard Henderson <rth@redhat.com>
4536 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4537 on failure the store back into EXPECT.
4539 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4540 Sandra Loosemore <sandra@codesourcery.com>
4542 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4543 * config/nios2/nios2.c (nios2_function_profiler): Add
4544 -fPIC (flag_pic == 2) support.
4545 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4546 (nios2_large_offset_p): New function.
4547 (nios2_unspec_reloc_p): Move up position, update to use
4548 nios2_large_offset_p.
4549 (nios2_unspec_address): Remove function.
4550 (nios2_unspec_offset): New function.
4551 (nios2_large_got_address): New function.
4552 (nios2_got_address): Add large offset support.
4553 (nios2_legitimize_tls_address): Update usage of removed and new
4555 (nios2_symbol_binds_local_p): New function.
4556 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4557 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4558 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4559 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4560 processing for (const (unspec ...)).
4561 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4563 2014-02-20 Richard Biener <rguenther@suse.de>
4565 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4566 doing the substitution.
4567 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4569 2014-02-20 Martin Jambor <mjambor@suse.cz>
4572 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4573 info when checking whether lattices are bottom.
4575 2014-02-20 Richard Biener <rguenther@suse.de>
4578 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4581 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4584 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4585 parameter specifying the scaling.
4586 (inline_call): Update.
4587 (want_inline_recursively): Guard division by zero.
4588 (recursive_inlining): Update.
4589 * ipa-inline.h (clone_inlined_nodes): Update.
4591 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4594 * config/i386/i386.c (classify_argument): Pass structures of size
4595 64 bytes or less in register.
4597 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4598 Kirill Yukhin <kirill.yukhin@intel.com>
4600 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4601 (_mm_rcp28_round_ss): Ditto.
4602 (_mm_rsqrt28_round_sd): Ditto.
4603 (_mm_rsqrt28_round_ss): Ditto.
4604 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4605 (_mm_rcp14_round_ss): Ditto.
4606 (_mm_rsqrt14_round_sd): Ditto.
4607 (_mm_rsqrt14_round_ss): Ditto.
4608 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4609 the first input operand, get rid of match_dup.
4610 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4612 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4614 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4615 operand as the first input operand, set type attribute.
4616 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4618 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4619 operand as the first input operand, set type attribute.
4621 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4623 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4626 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4629 * config/i386/i386.c (construct_container): Remove TFmode check
4630 for X86_64_INTEGER_CLASS.
4632 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4635 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4636 only when -Wpsabi is enabled.
4638 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4641 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4642 passing arrays in registers are the same as for structs, so remove the
4643 special case for them.
4645 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4647 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4648 destination type, extract only the valid bits if the source type is not
4649 integral and has a different mode.
4651 2014-02-19 Richard Biener <rguenther@suse.de>
4654 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4657 2014-02-19 Richard Biener <rguenther@suse.de>
4660 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4661 (ipa_modify_call_arguments): Emit an argument load explicitely and
4662 preserve virtual SSA form there and for the replacement call.
4663 Do not update SSA form nor free dominance info.
4665 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4667 * ipa.c (function_and_variable_visibility): Also clear WEAK
4668 flag when disolving COMDAT_GROUP.
4670 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4672 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4673 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4675 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4676 do more sanity checks.
4677 (detect_type_change): Return true when giving up early.
4678 (compute_complex_assign_jump_func): Fix type parameter of
4679 ipa_set_ancestor_jf.
4680 (compute_complex_ancestor_jump_func): Likewise.
4681 (update_jump_functions_after_inlining): Fix updating of
4683 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4685 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4687 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4688 inline clones when edge disappears.
4690 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4693 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4694 Split 64-bit moves into 2 patterns. Do not allow the use of
4695 direct move for TDmode in little endian, since the decimal value
4696 has little endian bytes within a word, but the 64-bit pieces are
4697 ordered in a big endian fashion, and normal subreg's of TDmode are
4699 (mov<mode>_64bit_dm): Likewise.
4700 (movtd_64bit_nodm): Likewise.
4702 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4704 PR tree-optimization/60174
4705 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4706 statement of an SSA_NAME that occurs in an abnormal PHI node.
4708 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4711 * final.c (SEEN_BB): Remove.
4712 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4713 (final_scan_insn): Don't force_source_line on second
4714 NOTE_INSN_BASIC_BLOCK.
4716 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4719 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4720 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4721 (type_natural_mode): Warn ABI change when %zmm register is not
4722 available for AVX512F vector value passing.
4724 2014-02-18 Kai Tietz <ktietz@redhat.com>
4727 * config/i386/i386.c (ix86_expand_prologue): Use value in
4728 rax register as displacement when restoring %r10 or %rax.
4729 Fix wrong offset when restoring both registers.
4731 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4733 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4734 assertion with conditional return.
4736 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4737 Uros Bizjak <ubizjak@gmail.com>
4740 * config/i386/driver-i386.c (host_detect_local_cpu): If
4741 YMM state is not saved by the OS, also clear has_f16c. Move
4742 CPUID 0x80000001 handling before YMM state saving checking.
4744 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4746 PR rtl-optimization/58960
4747 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4748 factored out from ...
4749 (sched_init): ... here.
4750 (free_global_sched_pressure_data): New, factored out from ...
4751 (sched_finish): ... here.
4752 * sched-int.h (free_global_sched_pressure_data): Declare.
4753 * sched-rgn.c (nr_regions_initial): New static global.
4754 (haifa_find_rgns): Initialize it.
4755 (schedule_region): Disable sched-pressure for the newly
4758 2014-02-17 Richard Biener <rguenther@suse.de>
4760 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4761 release SSA defs of pattern stmts.
4763 2014-02-17 Richard Biener <rguenther@suse.de>
4765 * tree-inline.c (expand_call_inline): Release the virtual
4766 operand defined by the call we are about to inline.
4768 2014-02-17 Richard Biener <rguenther@suse.de>
4770 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4772 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4773 Ilya Tocar <ilya.tocar@intel.com>
4775 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4776 arguments order in builtin.
4777 (_mm512_permutexvar_epi64): Ditto.
4778 (_mm512_mask_permutexvar_epi64): Ditto
4779 (_mm512_maskz_permutexvar_epi32): Ditto
4780 (_mm512_permutexvar_epi32): Ditto
4781 (_mm512_mask_permutexvar_epi32): Ditto
4783 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4785 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4786 (p8_vmrgow): Likewise.
4788 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4790 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4793 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4796 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4797 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4798 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4799 using direct move instructions on ISA 2.07. Also adjust
4800 instruction length for 64-bit.
4801 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4802 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4804 2014-02-15 Alan Modra <amodra@gmail.com>
4808 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4809 find_replacement on parts of insn rtl that might be reloaded.
4811 2014-02-15 Richard Biener <rguenther@suse.de>
4813 PR tree-optimization/60183
4814 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
4815 (tree_ssa_phiprop): Calculate and free post-dominators.
4817 2014-02-14 Jeff Law <law@redhat.com>
4819 PR rtl-optimization/60131
4820 * ree.c (get_extended_src_reg): New function.
4821 (combine_reaching_defs): Use it rather than assuming location of REG.
4822 (find_and_remove_re): Verify first operand of extension is
4823 a REG before adding the insns to the copy list.
4825 2014-02-14 Roland McGrath <mcgrathr@google.com>
4827 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
4828 * configure: Regenerated.
4829 * config.in: Regenerated.
4830 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
4831 instead of ASM_SHORT.
4833 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
4834 Richard Earnshaw <rearnsha@arm.com>
4836 PR rtl-optimization/59535
4837 * lra-constraints.c (process_alt_operands): Encourage alternative
4838 when unassigned pseudo class is superset of the alternative class.
4839 (inherit_reload_reg): Don't inherit when optimizing for code size.
4840 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
4841 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
4842 modes not less than 4 for Thumb1.
4844 2014-02-14 Kyle McMartin <kyle@redhat.com>
4847 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
4849 2014-02-14 Richard Biener <rguenther@suse.de>
4851 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
4852 (get_frame_arg): Drop the assert with langhook types_compatible_p.
4853 Do not strip INDIRECT_REFs.
4855 2014-02-14 Richard Biener <rguenther@suse.de>
4858 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
4859 DECL_FUNCTION_SPECIFIC_TARGET.
4860 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
4861 * tree-streamer-out.c (pack_ts_target_option): Remove.
4862 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
4863 (write_ts_function_decl_tree_pointers): Do not stream
4864 DECL_FUNCTION_SPECIFIC_TARGET.
4865 * tree-streamer-in.c (unpack_ts_target_option): Remove.
4866 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
4867 (lto_input_ts_function_decl_tree_pointers): Do not stream
4868 DECL_FUNCTION_SPECIFIC_TARGET.
4870 2014-02-14 Jakub Jelinek <jakub@redhat.com>
4872 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
4873 (get_initial_def_for_induction, vectorizable_induction): Ignore
4874 debug stmts when looking for exit_phi.
4875 (vectorizable_live_operation): Fix up condition.
4877 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4879 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
4880 nreverse() because it changes the content of original tree list.
4882 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4884 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
4885 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
4887 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4889 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
4890 GNU coding standards.
4892 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4895 * dwarf2out.c (gen_subprogram_die): Don't call
4896 add_calling_convention_attribute if subr_die is old_die.
4898 2014-02-13 Sharad Singhai <singhai@google.com>
4900 * doc/optinfo.texi: Fix order of nodes.
4902 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
4904 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
4905 operands[2], not operands[3].
4907 2014-02-13 Richard Biener <rguenther@suse.de>
4910 * doc/install.texi (ISL): Update recommended version to 0.12.2,
4911 mention the possibility of an in-tree build.
4912 (CLooG): Update recommended version to 0.18.1, mention the
4913 possibility of an in-tree build and clarify that the ISL
4914 bundled with CLooG does not work.
4916 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4919 * expr.c (compress_float_constant): If x is a hard register,
4920 extend into a pseudo and then move to x.
4922 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
4924 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
4925 caused by bad second argument to warning_at() with -mhotpatch and
4926 nested functions (e.g. with gfortran).
4928 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
4930 * opts.c (option_name): Remove "enabled by default" rider.
4932 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
4934 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
4936 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
4937 Uros Bizjak <ubizjak@gmail.com>
4940 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
4941 * configure: Regenerated.
4943 2014-02-12 Richard Biener <rguenther@suse.de>
4945 * vec.c (vec_prefix::calculate_allocation): Move as
4946 inline variant to vec.h.
4947 (vec_prefix::calculate_allocation_1): New out-of-line version.
4948 * vec.h (vec_prefix::calculate_allocation_1): Declare.
4949 (vec_prefix::m_has_auto_buf): Rename to ...
4950 (vec_prefix::m_using_auto_storage): ... this.
4951 (vec_prefix::calculate_allocation): Inline the easy cases
4952 and dispatch to calculate_allocation_1 which doesn't need the
4954 (va_heap::reserve): Use gcc_checking_assert.
4955 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
4956 m_using_auto_storage.
4957 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
4959 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
4960 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
4961 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
4963 2014-02-12 Richard Biener <rguenther@suse.de>
4965 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
4966 when we found a dependence.
4968 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
4970 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
4972 (maybe_fold_stmt): ... into this new function.
4973 * omp-low.c (lower_omp): Update comment.
4975 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
4978 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
4981 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
4983 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
4984 identifiers in comments.
4985 (cortexa53_extra_costs): Likewise.
4986 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
4987 (cortexa7_extra_costs): Likewise.
4988 (cortexa12_extra_costs): Likewise.
4989 (cortexa15_extra_costs): Likewise.
4990 (v7m_extra_costs): Likewise.
4992 2014-02-12 Richard Biener <rguenther@suse.de>
4995 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
4996 of posix_memalign being successful.
4997 (lower_stmt): Restrict lowering of posix_memalign to when
4998 -ftree-bit-ccp is enabled.
5000 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5002 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5004 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5006 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5008 PR rtl-optimization/60116
5009 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5010 other_insn once the combination has been validated.
5012 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5015 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5017 * ipa-devirt.c: Include demangle.h
5018 (odr_violation_reported): New static variable.
5019 (add_type_duplicate): Update odr_violations.
5020 (maybe_record_node): Add completep parameter; update it.
5021 (record_target_from_binfo): Add COMPLETEP parameter;
5022 update it as needed.
5023 (possible_polymorphic_call_targets_1): Likewise.
5024 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5025 rename FINAL to COMPLETE.
5026 (record_targets_from_bases): Sanity check we found the binfo;
5027 fix COMPLETEP updating.
5028 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5029 parameter, fix computing of COMPLETEP.
5030 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5031 at LTO time do demangling.
5032 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5033 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5035 (gimple_get_virt_method_for_binfo): Likewise.
5036 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5037 gimple_get_virt_method_for_vtable): Update prototypes.
5039 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5042 * genautomata.c (add_presence_absence): Fix typo with
5043 {final_}presence_list.
5045 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5048 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5049 for VSX/Altivec vectors that land in GPR registers.
5051 2014-02-11 Richard Henderson <rth@redhat.com>
5052 Jakub Jelinek <jakub@redhat.com>
5055 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5056 around drhs if type conversion to lacc->type is not useless.
5058 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5060 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5062 (cortex-a57.cortex-a53): Likewise.
5063 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5065 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5067 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5070 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5072 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5073 add_options_for_arm_vfp3.
5075 2014-02-11 Jeff Law <law@redhat.com>
5078 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5079 object with an undesirable mode.
5081 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5084 * config/i386/sol2-9.h: New file.
5085 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5086 *-*-solaris2.9*): Use it.
5088 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5090 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5091 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5093 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5095 * config/microblaze/microblaze.c: Extend mcpu version format
5097 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5099 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5101 2014-02-10 Richard Henderson <rth@redhat.com>
5104 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5105 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5106 ms-abi vs -mno-accumulate-outgoing-args.
5107 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5108 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5111 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5114 * cfgexpand.c (expand_asm_operands): Attach source location to
5115 ASM_INPUT rtx objects.
5116 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5118 2014-02-10 Nick Clifton <nickc@redhat.com>
5120 * config/mn10300/mn10300.c (popcount): New function.
5121 (mn10300_expand_prologue): Include saved registers in stack usage
5124 2014-02-10 Jeff Law <law@redhat.com>
5127 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5128 when changing the SET_DEST of a prior insn to avoid an input reload.
5130 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5132 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5133 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5134 -mcall-openbsd, or -mcall-linux.
5135 (CC1_ENDIAN_BIG_SPEC): Remove.
5136 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5137 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5138 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5139 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5140 and %cc1_endian_default.
5141 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5143 2014-02-10 Richard Biener <rguenther@suse.de>
5145 PR tree-optimization/60115
5146 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5147 MEM_REF handling. Properly verify that the accesses are not
5148 out of the objects bound.
5150 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5152 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5155 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5157 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5158 proper constants and fix formatting.
5159 (possible_polymorphic_call_targets): Fix formatting.
5161 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5162 Ilya Tocar <ilya.tocar@intel.com>
5164 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5165 (_mm512_loadu_epi32): Renamed into...
5166 (_mm512_loadu_si512): This.
5167 (_mm512_storeu_epi32): Renamed into...
5168 (_mm512_storeu_si512): This.
5169 (_mm512_maskz_ceil_ps): Removed.
5170 (_mm512_maskz_ceil_pd): Ditto.
5171 (_mm512_maskz_floor_ps): Ditto.
5172 (_mm512_maskz_floor_pd): Ditto.
5173 (_mm512_floor_round_ps): Ditto.
5174 (_mm512_floor_round_pd): Ditto.
5175 (_mm512_ceil_round_ps): Ditto.
5176 (_mm512_ceil_round_pd): Ditto.
5177 (_mm512_mask_floor_round_ps): Ditto.
5178 (_mm512_mask_floor_round_pd): Ditto.
5179 (_mm512_mask_ceil_round_ps): Ditto.
5180 (_mm512_mask_ceil_round_pd): Ditto.
5181 (_mm512_maskz_floor_round_ps): Ditto.
5182 (_mm512_maskz_floor_round_pd): Ditto.
5183 (_mm512_maskz_ceil_round_ps): Ditto.
5184 (_mm512_maskz_ceil_round_pd): Ditto.
5185 (_mm512_expand_pd): Ditto.
5186 (_mm512_expand_ps): Ditto.
5187 * config/i386/i386.c (ix86_builtins): Remove
5188 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5189 (bdesc_args): Ditto.
5190 * config/i386/predicates.md (const1256_operand): New.
5191 (const_1_to_2_operand): Ditto.
5192 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5193 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5194 (*avx512pf_gatherpf<mode>sf): Ditto.
5195 (avx512pf_gatherpf<mode>df): Ditto.
5196 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5197 (*avx512pf_gatherpf<mode>df): Ditto.
5198 (avx512pf_scatterpf<mode>sf): Ditto.
5199 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5200 (*avx512pf_scatterpf<mode>sf): Ditto.
5201 (avx512pf_scatterpf<mode>df): Ditto.
5202 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5203 (*avx512pf_scatterpf<mode>df): Ditto.
5204 (avx512f_expand<mode>): Removed.
5205 (<shift_insn><mode>3<mask_name>): Change predicate type.
5207 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5209 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5210 not at the end of datarefs vector use ordered_remove to avoid
5211 reordering datarefs vector.
5214 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5215 mark local addressable non-static vars as GOVD_PRIVATE
5216 instead of GOVD_LOCAL.
5217 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5218 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5222 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5223 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5224 assume_aligned or alloc_align attributes.
5225 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5226 arguments. Handle also assume_aligned and alloc_align attributes.
5227 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5228 calls to functions with assume_aligned or alloc_align attributes.
5229 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5231 2014-02-08 Terry Guo <terry.guo@arm.com>
5233 * doc/invoke.texi: Document ARM -march=armv7e-m.
5235 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5237 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5238 flag on __cilkrts_rethrow builtin.
5241 * ipa-cp.c (determine_versionability): Fail at -O0
5242 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5243 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5246 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5249 * tree-inline.c (copy_forbidden): Fail for
5250 __attribute__((optimize (0))) functions.
5252 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5254 * varpool.c: Include pointer-set.h.
5255 (varpool_remove_unreferenced_decls): Variables in other partitions
5256 will not be output; be however careful to not lose information
5259 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5261 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5262 lookup in the vtable constructor.
5264 2014-02-07 Jeff Law <law@redhat.com>
5267 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5268 define_insn_and_split.
5270 * ipa-inline.c (inline_small_functions): Fix typos.
5272 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5274 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5275 (s390_can_use_return_insn): Declare.
5276 * config/s390/s390.h (EPILOGUE_USES): Define.
5277 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5279 (s390_chunkify_start): Handle return JUMP_LABELs.
5280 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5281 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5282 (s390_can_use_return_insn): New functions.
5283 (s390_fix_long_loop_prediction): Handle conditional returns.
5284 (TARGET_SET_UP_BY_PROLOGUE): Define.
5285 * config/s390/s390.md (ANY_RETURN): New code iterator.
5286 (*creturn, *csimple_return, return, simple_return): New patterns.
5288 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5290 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5291 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5292 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5293 REG_CFA_RESTORE list when deciding not to restore a register.
5295 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5297 * config/s390/s390.c: Include tree-pass.h and context.h.
5298 (s390_early_mach): New function, split out from...
5299 (s390_emit_prologue): ...here.
5300 (pass_data_s390_early_mach): New pass structure.
5301 (pass_s390_early_mach): New class.
5302 (s390_option_override): Create and register early_mach pass.
5303 Move to end of file.
5305 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5307 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5308 to match for the exit block.
5310 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5312 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5313 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5314 Reject misaligned operands.
5316 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5318 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5320 2014-02-07 Richard Biener <rguenther@suse.de>
5323 * gimple-low.c (lower_builtin_posix_memalign): New function.
5324 (lower_stmt): Call it to lower posix_memalign in a way
5325 to make alignment info accessible.
5327 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5330 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5331 __builtin_setjmp_receiver.
5333 2014-02-07 Richard Biener <rguenther@suse.de>
5336 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5337 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5338 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5339 Handle BUILT_IN_POSIX_MEMALIGN.
5340 (find_func_clobbers): Likewise.
5341 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5342 (call_may_clobber_ref_p_1): Likewise.
5344 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5347 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5350 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5353 * lto-cgraph.c (lto_output_node): Use
5354 symtab_get_symbol_partitioning_class.
5355 (lto_output_varpool_node): likewise.
5356 (symtab_get_symbol_partitioning_class): Move here from
5358 * cgraph.h (symbol_partitioning_class): Likewise.
5359 (symtab_get_symbol_partitioning_class): Declare.
5361 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5363 * ggc.h (ggc_internal_cleared_alloc): New macro.
5364 * vec.h (vec_safe_copy): Handle memory stats.
5365 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5366 * target-globals.c (save_target_globals): Likewise.
5368 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5371 * expr.c (emit_move_resolve_push): Export; be bit more selective
5372 on when to clear alias set.
5373 * expr.h (emit_move_resolve_push): Declare.
5374 * function.h (struct function): Add tail_call_marked.
5375 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5376 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5377 * config/i386/i386.md (TImode move expander): De not call
5379 (FP push expanders): Preserve memory attributes.
5380 * config/i386/sse.md (push<mode>1): Remove.
5381 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5382 (ix86_expand_push): Remove.
5383 * config/i386/mmx.md (push<mode>1): Remove.
5385 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5387 PR rtl-optimization/60030
5388 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5389 lopart with paradoxical subreg before shifting it up by hprec.
5391 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5393 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5394 Remove extra newline at end of file.
5395 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5396 (arm_issue_rate): Handle cortexa57.
5397 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5398 (cortex-a57.cortex-a53): Likewise.
5400 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5403 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5404 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5406 (arm_expand_prologue): Adjust all callers.
5407 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5408 info, registers also at the lowest numbered registers side. Use
5409 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5413 * var-tracking.c (adjust_mems): Before adding a SET to
5414 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5416 2014-02-06 Alan Modra <amodra@gmail.com>
5419 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5420 change SDmode to DDmode when lra_in_progress.
5422 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5425 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5426 free_data_ref on the dr first, and before goto again also set dr
5427 to the next dr. For simd_lane_access, free old datarefs[i] before
5428 overwriting it. For get_vectype_for_scalar_type failure, don't
5429 free_data_ref if simd_lane_access.
5431 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5434 * tree.h (opts_for_fn): New inline function.
5435 (opt_for_fn): Define.
5436 * config/i386/i386.c (ix86_function_regparm): Use
5437 opt_for_fn (decl, optimize) instead of optimize.
5439 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5441 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5442 for SYMBOL_REF in large memory model.
5444 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5446 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5448 (cortex-a57): Likewise.
5449 (cortex-a57.cortex-a53): Likewise.
5451 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5452 Kugan Vivekanandarajah <kuganv@linaro.org>
5454 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5456 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5458 2014-02-06 Richard Biener <rguenther@suse.de>
5460 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5461 set_loop_copy and initialize_original_copy_tables.
5463 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5465 * config/aarch64/aarch64-simd.md
5466 (aarch64_ashr_simddi): Change QI to SI.
5468 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5469 Jakub Jelinek <jakub@redhat.com>
5472 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5475 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5477 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5478 CODE_FOR_altivec_vpku[hw]um to
5479 CODE_FOR_altivec_vpku[hw]um_direct.
5480 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5481 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5482 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5483 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5485 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5487 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5488 generation for -maltivec=be.
5489 (altivec_vsumsws): Simplify redundant test.
5491 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5493 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5494 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5495 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5496 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5497 gen_altivec_vpkuwum.
5498 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5500 (altivec_vpks<VI_char>ss): Likewise.
5501 (altivec_vpks<VI_char>us): Likewise.
5502 (altivec_vpku<VI_char>us): Likewise.
5503 (altivec_vpku<VI_char>um): Likewise.
5504 (altivec_vpku<VI_char>um_direct): New (copy of
5505 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5507 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5508 target is little endian and -maltivec=be is not specified.
5509 (*altivec_vupkhs<VU_char>_direct): New (copy of
5510 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5511 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5512 target is little endian and -maltivec=be is not specified.
5513 (*altivec_vupkls<VU_char>_direct): New (copy of
5514 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5515 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5516 little endian and -maltivec=be is not specified.
5517 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5518 little endian and -maltivec=be is not specified.
5520 2014-02-05 Richard Henderson <rth@redhat.com>
5523 * combine-stack-adj.c: Revert r206943.
5524 * sched-int.h (struct deps_desc): Add last_args_size.
5525 * sched-deps.c (init_deps): Initialize it.
5526 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5527 contain REG_ARGS_SIZE notes.
5529 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5531 * lto-cgraph.c (asm_nodes_output): Make global.
5532 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5533 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5534 (driver_handle_option): Handle OPT_fwpa.
5536 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5539 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5540 a comment typo and formatting issue. If odr_hash hasn't been
5541 created, return vNULL and set *completep to false.
5544 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5545 bb with no successors.
5547 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5550 * doc/invoke.texi (-march): Clarify documentation for ARM.
5554 2014-02-05 Richard Biener <rguenther@suse.de>
5556 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5557 when not vectorizing because of too many alias checks.
5558 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5559 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5561 2014-02-05 Nick Clifton <nickc@redhat.com>
5563 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5564 accept extended registers in any mode when compiling for the MN10300.
5566 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5568 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5569 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5570 sanitization attributes.
5571 (can_inline_edge_p): Likewise.
5572 (sanitize_attrs_match_for_inline_p): New function.
5574 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5576 * ipa-prop.c (detect_type_change): Shor circuit testing of
5577 type changes on THIS pointer.
5579 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5582 * config/pa/pa.c (legitimize_tls_address): Return original address
5583 if not passed a SYMBOL_REF rtx.
5584 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5586 (pa_emit_move_sequence): Simplify TLS source operands.
5587 (pa_legitimate_constant_p): Reject all TLS constants.
5588 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5589 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5591 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5593 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5594 groups when we know they are controlled by LTO.
5595 * varasm.c (default_binds_local_p_1): If object is in other partition,
5596 it will be resolved locally.
5598 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5600 * config/host-linux.c (linux_gt_pch_use_address): Don't
5601 use SSIZE_MAX because it is not always defined.
5603 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5606 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5607 threshold for pseudo splitting.
5608 (update_ebb_live_info): Process call argument hard registers and
5609 hard registers from insn definition too.
5610 (max_small_class_regs_num): New constant.
5611 (inherit_in_ebb): Update live hard regs through EBBs. Update
5612 reloads_num only for small register classes. Don't split for
5615 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5618 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5621 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5623 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5626 2014-02-04 Marek Polacek <polacek@redhat.com>
5628 * gdbinit.in (pel): Define.
5630 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5632 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5635 2014-02-04 Richard Biener <rguenther@suse.de>
5638 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5639 in function context local.
5640 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5641 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5642 similar to LTO_imported_decl_ref.
5644 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5646 PR tree-optimization/60002
5647 * cgraphclones.c (build_function_decl_skip_args): Clear
5650 PR tree-optimization/60023
5651 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5652 false to gsi_replace.
5653 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5654 has been in some EH region and vec_stmt could throw, add
5655 vec_stmt into the same EH region.
5656 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5657 has no lhs, ignore it.
5658 * internal-fn.c (expand_MASK_LOAD): Likewise.
5661 * tree-inline.c (copy_forbidden): Fail for
5662 __attribute__((optimize (0))) functions.
5665 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5666 is set, copy one less argument.
5667 (expand_simd_clones): Don't subtract clone_info->inbranch
5668 from simd_clone_struct_alloc argument.
5670 PR rtl-optimization/57915
5671 * recog.c (simplify_while_replacing): If all unary/binary/relational
5672 operation arguments are constant, attempt to simplify those.
5675 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5676 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5678 2014-02-04 Richard Biener <rguenther@suse.de>
5680 PR tree-optimization/60012
5681 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5682 TBAA disambiguation to all DDRs.
5684 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5687 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5688 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5690 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5693 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5695 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5697 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5698 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5700 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5703 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5704 to figure out targets of polymorphic calls with known decl.
5705 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5706 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5707 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5708 (get_polymorphic_call_info): ... here.
5709 (get_polymorphic_call_info_from_invariant): New function.
5711 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5713 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5714 lookup via vtable pointer; check for type consistency
5715 and turn inconsitent facts into UNREACHABLE.
5716 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5717 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5718 type inconsistent querries; return UNREACHABLE instead.
5720 2014-02-03 Richard Henderson <rth@twiddle.net>
5723 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5724 already processed this node.
5725 (normalize_one_pred_1): Pass along mark_set.
5726 (normalize_one_pred): Create and destroy a pointer_set_t.
5727 (normalize_one_pred_chain): Likewise.
5729 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5731 PR gcov-profile/58602
5732 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5734 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5737 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5738 -fno-devirtualize; try to devirtualize by the knowledge of
5739 virtual table pointer given by aggregate propagation.
5740 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5741 (ipa_print_node_jump_functions): Dump also offset that
5742 is relevant for polymorphic calls.
5743 (determine_known_aggregate_parts): Add arg_type parameter; use it
5744 instead of determining the type from pointer type.
5745 (ipa_compute_jump_functions_for_edge): Update call of
5746 determine_known_aggregate_parts.
5747 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5748 (gimple_get_virt_method_for_binfo): ... here; simplify using
5749 vtable_pointer_value_to_vtable.
5750 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5751 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5752 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5753 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5755 (vtable_pointer_value_to_binfo): ... here.
5756 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5758 2014-02-03 Teresa Johnson <tejohnson@google.com>
5760 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5761 redef of outer loop index variable.
5763 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5767 * doc/extend.texi (Function Attributes): Typo.
5769 2014-02-03 Cong Hou <congh@google.com>
5771 PR tree-optimization/60000
5772 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5773 if the vectorized statement is a store. A store statement can only
5774 appear at the end of pattern statements.
5776 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5778 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5779 (ix86_option_override_internal): Default long double to 64-bit for
5780 32-bit Bionic and to 128-bit for 64-bit Bionic.
5782 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5783 TARGET_LONG_DOUBLE_128 is true.
5784 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5786 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5787 (mlong-double-64): Negate -mlong-double-128.
5788 (mlong-double-128): New option.
5790 * config/i386/i386-c.c (ix86_target_macros): Define
5791 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5793 * doc/invoke.texi: Document -mlong-double-128.
5795 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5797 PR rtl-optimization/60024
5798 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5800 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5802 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5804 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5806 PR rtl-optimization/57662
5807 * sel-sched.c (code_motion_path_driver): Do not mark already not
5808 existing blocks in the visiting bitmap.
5810 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5812 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
5813 on the insn being emitted.
5815 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
5816 Will Deacon <will.deacon@arm.com>
5818 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
5820 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5822 * config/arm/arm-tables.opt: Regenerate.
5824 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5826 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
5827 for vector types other than V16QImode.
5828 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
5829 define_expand, and call altivec_expand_vec_perm_le when producing
5830 code with little endian element order.
5831 (*altivec_vperm_<mode>_internal): New insn having previous
5832 behavior of altivec_vperm_<mode>.
5833 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
5834 altivec_expand_vec_perm_le when producing code with little endian
5836 (*altivec_vperm_<mode>_uns_internal): New insn having previous
5837 behavior of altivec_vperm_<mode>_uns.
5839 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5841 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
5842 (altivec_vsumsws): Add handling for -maltivec=be with a little
5844 (altivec_vsumsws_direct): New.
5845 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
5846 gen_altivec_vsumsws.
5848 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5850 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
5851 vtable_pointer_value_to_binfo): New functions.
5852 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
5853 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
5855 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
5857 * config/nios2/nios2.md (load_got_register): Initialize GOT
5858 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
5859 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
5861 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5863 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
5864 preserverd by passthrough, do not propagate the type.
5866 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5868 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
5869 (mips_atomic_assign_expand_fenv): New function.
5870 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
5872 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5874 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
5875 (__builtin_mips_set_fcsr): Likewise.
5876 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
5877 MIPS_USI_FTYPE_VOID.
5878 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
5879 (mips16_expand_set_fcsr): Likewise.
5880 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
5881 (mips16_set_fcsr_stub): Likewise.
5882 (mips16_get_fcsr_one_only_stub): New class.
5883 (mips16_set_fcsr_one_only_stub): Likewise.
5884 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
5885 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
5886 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
5887 (hard_float): New availability predicate.
5888 (mips_builtins): Add get_fcsr and set_fcsr.
5889 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
5890 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
5891 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
5892 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
5893 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
5896 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5898 * config/mips/mips.c (mips_one_only_stub): New class.
5899 (mips_need_mips16_rdhwr_p): Replace with...
5900 (mips16_rdhwr_stub): ...this new variable.
5901 (mips16_stub_call_address): New function.
5902 (mips16_rdhwr_one_only_stub): New class.
5903 (mips_expand_thread_pointer): Use mips16_stub_call_address.
5904 (mips_output_mips16_rdhwr): Delete.
5905 (mips_finish_stub): New function.
5906 (mips_code_end): Use it to handle rdhwr stubs.
5908 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
5911 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
5912 when calculating size of integer atomic types.
5914 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
5916 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
5918 2014-02-01 Jakub Jelinek <jakub@redhat.com>
5920 PR tree-optimization/60003
5921 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
5922 * profile.c (branch_prob): Use gimple_call_builtin_p
5923 to check for BUILT_IN_SETJMP_RECEIVER.
5924 * tree-inline.c (copy_bb): Call notice_special_calls.
5926 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
5929 * lra-constraints.c (process_alt_operands): Update reload_sum only
5932 2014-01-31 Richard Henderson <rth@redhat.com>
5935 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
5936 until after else_eh is processed.
5938 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
5940 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
5941 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
5942 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
5943 in smmintrin.h, remove them.
5944 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
5945 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
5946 * config/i386/i386.md (ROUND_SAE): Fix value.
5947 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
5948 (const48_operand): New.
5949 * config/i386/subst.md (round), (round_expand): Use
5950 const_4_or_8_to_11_operand.
5951 (round_saeonly), (round_saeonly_expand): Use const48_operand.
5953 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
5955 * config/i386/constraints.md (Yk): Swap meaning with k.
5956 * config/i386/i386.md (movhi_internal): Change Yk to k.
5957 (movqi_internal): Ditto.
5958 (*k<logic><mode>): Ditto.
5961 (kandn<mode>): Ditto.
5962 (*<code>hi_1): Ditto.
5963 (*<code>qi_1): Ditto.
5964 (kxnor<mode>): Ditto.
5965 (kortestzhi): Ditto.
5966 (kortestchi): Ditto.
5968 (*one_cmplhi2_1): Ditto.
5969 (*one_cmplqi2_1): Ditto.
5970 * config/i386/sse.md (): Change k to Yk.
5971 (avx512f_load<mode>_mask): Ditto.
5972 (avx512f_blendm<mode>): Ditto.
5973 (avx512f_store<mode>_mask): Ditto.
5974 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
5975 (avx512f_storedqu<mode>_mask): Ditto.
5976 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
5978 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
5979 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
5980 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
5981 (avx512f_maskcmp<mode>3): Ditto.
5982 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
5983 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
5984 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
5985 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
5986 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
5987 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
5988 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
5989 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
5990 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
5991 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
5992 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
5993 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
5994 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
5995 (vec_extract_lo_<mode>_maskm): Ditto.
5996 (vec_extract_hi_<mode>_maskm): Ditto.
5997 (avx512f_vternlog<mode>_mask): Ditto.
5998 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
5999 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6000 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6001 (avx512f_<code>v8div16qi2_mask): Ditto.
6002 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6003 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6004 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6005 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6006 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6007 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6008 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6009 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6010 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6011 (avx512cd_maskb_vec_dupv8di): Ditto.
6012 (avx512cd_maskw_vec_dupv16si): Ditto.
6013 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6014 (avx512f_vpermi2var<mode>3_mask): Ditto.
6015 (avx512f_vpermi2var<mode>3_mask): Ditto.
6016 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6017 (*avx512f_gathersi<mode>): Ditto.
6018 (*avx512f_gathersi<mode>_2): Ditto.
6019 (*avx512f_gatherdi<mode>): Ditto.
6020 (*avx512f_gatherdi<mode>_2): Ditto.
6021 (*avx512f_scattersi<mode>): Ditto.
6022 (*avx512f_scatterdi<mode>): Ditto.
6023 (avx512f_compress<mode>_mask): Ditto.
6024 (avx512f_compressstore<mode>_mask): Ditto.
6025 (avx512f_expand<mode>_mask): Ditto.
6026 * config/i386/subst.md (mask): Change k to Yk.
6027 (mask_scalar_merge): Ditto.
6030 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6032 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6034 2014-01-31 Richard Biener <rguenther@suse.de>
6037 * builtins.c (fold_builtin_memory_op): Make sure to not
6038 use a floating-point mode or a boolean or enumeral type for
6041 2014-01-30 DJ Delorie <dj@redhat.com>
6043 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6044 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6045 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6046 whenever main() has an epilogue.
6048 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6050 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6051 unused variable "field".
6052 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6053 (vsx_mergeh_<mode>): Likewise.
6054 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6055 (altivec_vmrghh): Likewise.
6056 (altivec_vmrghw): Likewise.
6057 (altivec_vmrglb): Likewise.
6058 (altivec_vmrglh): Likewise.
6059 (altivec_vmrglw): Likewise.
6060 (altivec_vspltb): Add missing uses.
6061 (altivec_vsplth): Likewise.
6062 (altivec_vspltw): Likewise.
6063 (altivec_vspltsf): Likewise.
6065 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6068 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6069 frame related instructions.
6071 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6073 PR rtl-optimization/59959
6074 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6075 any reload of register whose subreg is invalid.
6077 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6079 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6080 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6081 Add missing return type - void.
6083 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6085 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6086 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6087 remove element index adjustment for endian (now handled in vsx.md
6089 (altivec_expand_vec_perm_const): Use
6090 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6091 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6092 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6093 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6094 define_expand and a new define_insn *altivec_vspltb_internal;
6095 adjust for -maltivec=be on a little endian target.
6096 (altivec_vspltb_direct): New.
6097 (altivec_vsplth): Divide into a define_expand and a new
6098 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6099 little endian target.
6100 (altivec_vsplth_direct): New.
6101 (altivec_vspltw): Divide into a define_expand and a new
6102 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6103 little endian target.
6104 (altivec_vspltw_direct): New.
6105 (altivec_vspltsf): Divide into a define_expand and a new
6106 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6107 a little endian target.
6109 2014-01-30 Richard Biener <rguenther@suse.de>
6111 PR tree-optimization/59993
6112 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6113 can propagate form the earlier stmt and avoid the transform
6114 when the intermediate result is needed.
6116 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6118 * README.Portability: Fix typo.
6120 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6122 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6123 comparison_operator with ordered_comparison_operator.
6125 2014-01-30 Nick Clifton <nickc@redhat.com>
6127 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6128 Rename to mn10300_store_multiple_regs.
6129 * config/mn10300/mn10300.c: Likewise.
6130 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6131 store_multiple_regs.
6132 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6133 Call mn10300_store_multiple_regs.
6135 2014-01-30 Nick Clifton <nickc@redhat.com>
6136 DJ Delorie <dj@redhat.com>
6138 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6139 %fp 2 to keep registers after it properly word-aligned.
6140 (rl78_alloc_physical_registers_umul): Handle the case where both
6141 input operands are the same.
6143 2014-01-30 Richard Biener <rguenther@suse.de>
6145 PR tree-optimization/59903
6146 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6149 2014-01-30 Jason Merrill <jason@redhat.com>
6152 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6155 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6157 2014-01-30 Richard Biener <rguenther@suse.de>
6159 PR tree-optimization/59951
6160 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6162 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6165 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6166 SFmode to DFmode case.
6168 2014-01-29 DJ Delorie <dj@redhat.com>
6170 * config/msp430/msp430.opt (-minrt): New.
6171 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6173 (ENDFILE_SPEC): Likewise.
6175 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6177 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6178 (estimate_function_body_sizes): Use it.
6180 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6183 * dwarf2out.c (is_cxx_auto): New.
6184 (is_base_type): Use it.
6185 (gen_type_die_with_usage): Likewise.
6187 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6189 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6190 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6191 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6192 -maltivec=be with LE targets.
6193 (vsx_mergeh_<mode>): Likewise.
6194 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6195 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6196 (altivec_vmrghb): Replace with define_expand and new
6197 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6198 (altivec_vmrghb_direct): New define_insn.
6199 (altivec_vmrghh): Replace with define_expand and new
6200 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6201 (altivec_vmrghh_direct): New define_insn.
6202 (altivec_vmrghw): Replace with define_expand and new
6203 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6204 (altivec_vmrghw_direct): New define_insn.
6205 (*altivec_vmrghsf): Adjust for endianness.
6206 (altivec_vmrglb): Replace with define_expand and new
6207 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6208 (altivec_vmrglb_direct): New define_insn.
6209 (altivec_vmrglh): Replace with define_expand and new
6210 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6211 (altivec_vmrglh_direct): New define_insn.
6212 (altivec_vmrglw): Replace with define_expand and new
6213 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6214 (altivec_vmrglw_direct): New define_insn.
6215 (*altivec_vmrglsf): Adjust for endianness.
6216 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6217 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6218 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6219 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6220 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6221 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6222 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6223 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6225 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6227 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6228 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6231 2014-01-29 Richard Biener <rguenther@suse.de>
6233 PR tree-optimization/58742
6234 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6235 associate_pointerplus_align.
6236 (associate_pointerplus_diff): New function.
6237 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6238 and associate_pointerplus_diff.
6240 2014-01-29 Richard Biener <rguenther@suse.de>
6242 * lto-streamer.h (LTO_major_version): Bump to 3.
6243 (LTO_minor_version): Reset to 0.
6245 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6247 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6248 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6249 (arm_file_start): Generate correct asm header for armv7ve.
6250 * config/arm/bpabi.h: Add multilib support for armv7ve.
6251 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6252 and cortex-a15 to armv7ve.
6253 * config/arm/t-aprofile: Add multilib support for armv7ve.
6254 * doc/invoke.texi: Document -march=armv7ve.
6256 2014-01-29 Richard Biener <rguenther@suse.de>
6258 PR tree-optimization/58742
6259 * tree-ssa-forwprop.c (associate_plusminus): Return true
6260 if we changed sth, defer EH cleanup to ...
6261 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6262 (simplify_mult): New function.
6264 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6267 PR tree-optimization/59920
6268 * tree.c (build_common_builtin_nodes): Remove
6269 __builtin_setjmp_dispatcher initialization.
6270 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6271 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6272 instead of gsi_after_labels + manually skipping debug stmts.
6273 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6274 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6275 * tree-inline.c (copy_edges_for_bb): Remove
6276 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6277 argument. Ignore computed_goto_p stmts. Don't call
6278 make_abnormal_goto_edges. If a call might need abnormal edges
6279 for non-local gotos, see if it already has an edge to
6280 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6281 with true argument, don't do anything then, otherwise add
6282 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6283 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6285 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6286 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6287 (lower_stmt): Don't set data->calls_builtin_setjmp.
6288 (lower_builtin_setjmp): Adjust comment.
6289 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6290 * tree-cfg.c (found_computed_goto): Remove.
6291 (factor_computed_gotos): Remove.
6292 (make_goto_expr_edges): Return bool, true for computed gotos.
6293 Don't call make_abnormal_goto_edges.
6294 (build_gimple_cfg): Don't set found_computed_goto, don't call
6295 factor_computed_gotos.
6296 (computed_goto_p): No longer static.
6297 (make_blocks): Don't set found_computed_goto.
6298 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6299 (make_edges): If make_goto_expr_edges returns true, push bb
6300 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6301 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6302 vector. Record mapping between bbs and OpenMP regions if there
6303 are any, adjust make_gimple_omp_edges caller. Call
6304 handle_abnormal_edges.
6305 (make_abnormal_goto_edges): Remove.
6306 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6307 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6308 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6309 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6310 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6311 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6312 filling *region also set *region_idx to (*region)->entry->index.
6315 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6316 For REGs set ORIGINAL_REGNO.
6318 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6320 * doc/md.texi: Mention that a target shouldn't implement
6321 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6324 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6326 PR tree-optimization/59594
6327 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6328 a copy of the datarefs vector rather than the vector itself.
6330 2014-01-28 Jason Merrill <jason@redhat.com>
6333 * dwarf2out.c (auto_die): New static.
6334 (gen_type_die_with_usage): Handle C++1y 'auto'.
6335 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6338 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6341 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6342 (SPEC_X32): Likewise.
6343 (SPEC_64): Likewise.
6344 * config/i386/i386.c (ix86_option_override_internal): Turn off
6345 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6347 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6348 * config/i386/i386.h (TARGET_16BIT): New macro.
6349 (TARGET_16BIT_P): Likewise.
6350 * config/i386/i386.opt: Add m16.
6351 * doc/invoke.texi: Document -m16.
6353 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6355 PR preprocessor/59935
6356 * input.c (location_get_source_line): Bail out on when line number
6357 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6359 2014-01-28 Richard Biener <rguenther@suse.de>
6361 PR tree-optimization/58742
6362 * tree-ssa-forwprop.c (associate_plusminus): Handle
6363 pointer subtraction of the form (T)(P + A) - (T)P.
6365 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6367 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6370 2014-01-28 Richard Biener <rguenther@suse.de>
6373 2014-01-28 Richard Biener <rguenther@suse.de>
6375 PR rtl-optimization/45364
6376 PR rtl-optimization/59890
6377 * var-tracking.c (local_get_addr_clear_given_value): Handle
6378 already cleared slot.
6379 (val_reset): Handle not allocated local_get_addr_cache.
6380 (vt_find_locations): Use post-order on the inverted CFG.
6382 2014-01-28 Richard Biener <rguenther@suse.de>
6384 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6386 2014-01-28 Richard Biener <rguenther@suse.de>
6388 PR rtl-optimization/45364
6389 PR rtl-optimization/59890
6390 * var-tracking.c (local_get_addr_clear_given_value): Handle
6391 already cleared slot.
6392 (val_reset): Handle not allocated local_get_addr_cache.
6393 (vt_find_locations): Use post-order on the inverted CFG.
6395 2014-01-28 Alan Modra <amodra@gmail.com>
6397 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6398 * configure.ac <recursive call for build != host>: Define
6399 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6400 and LD_FOR_BUILD too.
6401 * configure: Regenerate.
6403 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6405 * config/i386/i386.c (get_builtin_code_for_version): Separate
6406 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6407 Broadwell from Haswell.
6409 2014-01-27 Steve Ellcey <sellcey@mips.com>
6411 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6412 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6413 * config/mips/mips.c (mips_option_override): Change setting
6415 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6416 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6417 Change from Mask to Var.
6419 2014-01-27 Jeff Law <law@redhat.com>
6421 * ipa-inline.c (inline_small_functions): Fix typo.
6423 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6425 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6426 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6427 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6428 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6429 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6430 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6431 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6432 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6433 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6434 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6435 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6436 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6437 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6438 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6439 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6440 (_mm512_storeu_epi64): Ditto.
6441 (_mm512_cmpge_epi32_mask): Ditto.
6442 (_mm512_cmpge_epu32_mask): Ditto.
6443 (_mm512_cmpge_epi64_mask): Ditto.
6444 (_mm512_cmpge_epu64_mask): Ditto.
6445 (_mm512_cmple_epi32_mask): Ditto.
6446 (_mm512_cmple_epu32_mask): Ditto.
6447 (_mm512_cmple_epi64_mask): Ditto.
6448 (_mm512_cmple_epu64_mask): Ditto.
6449 (_mm512_cmplt_epi32_mask): Ditto.
6450 (_mm512_cmplt_epu32_mask): Ditto.
6451 (_mm512_cmplt_epi64_mask): Ditto.
6452 (_mm512_cmplt_epu64_mask): Ditto.
6453 (_mm512_cmpneq_epi32_mask): Ditto.
6454 (_mm512_cmpneq_epu32_mask): Ditto.
6455 (_mm512_cmpneq_epi64_mask): Ditto.
6456 (_mm512_cmpneq_epu64_mask): Ditto.
6457 (_mm512_expand_pd): Ditto.
6458 (_mm512_expand_ps): Ditto.
6459 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6460 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6461 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6462 * config/i386/i386.c (ix86_builtins): Add
6463 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6464 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6465 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6466 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6467 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6468 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6469 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6470 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6471 IX86_BUILTIN_PMOVUSQW512_MEM.
6472 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6473 __builtin_ia32_pmovsqd512mem_mask,
6474 __builtin_ia32_pmovqd512mem_mask,
6475 __builtin_ia32_pmovusqw512mem_mask,
6476 __builtin_ia32_pmovsqw512mem_mask,
6477 __builtin_ia32_pmovqw512mem_mask,
6478 __builtin_ia32_pmovusdw512mem_mask,
6479 __builtin_ia32_pmovsdw512mem_mask,
6480 __builtin_ia32_pmovdw512mem_mask,
6481 __builtin_ia32_pmovqb512mem_mask,
6482 __builtin_ia32_pmovusqb512mem_mask,
6483 __builtin_ia32_pmovsqb512mem_mask,
6484 __builtin_ia32_pmovusdb512mem_mask,
6485 __builtin_ia32_pmovsdb512mem_mask,
6486 __builtin_ia32_pmovdb512mem_mask.
6487 (bdesc_args): Add __builtin_ia32_expanddf512,
6488 __builtin_ia32_expandsf512.
6489 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6490 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6491 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6492 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6493 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6494 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6495 (avx512f_<code>v8div16qi2_mask_store): This.
6496 (avx512f_expand<mode>): New.
6498 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6500 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6502 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6503 (_mm512_prefetch_i32scatter_pd): Ditto.
6504 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6505 (_mm512_prefetch_i64scatter_pd): Ditto.
6506 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6507 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6508 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6509 (_mm512_prefetch_i32scatter_ps): Ditto.
6510 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6511 (_mm512_prefetch_i64scatter_ps): Ditto.
6512 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6513 * config/i386/i386-builtin-types.def: Define
6514 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6515 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6516 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6517 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6518 IX86_BUILTIN_SCATTERPFQPD.
6519 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6520 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6521 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6522 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6523 __builtin_ia32_scatterpfqps.
6524 (ix86_expand_builtin): Expand new built-ins.
6525 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6526 fix memory access data type.
6527 (*avx512pf_gatherpf<mode>_mask): Ditto.
6528 (*avx512pf_gatherpf<mode>): Ditto.
6529 (avx512pf_scatterpf<mode>): Ditto.
6530 (*avx512pf_scatterpf<mode>_mask): Ditto.
6531 (*avx512pf_scatterpf<mode>): Ditto.
6532 (GATHER_SCATTER_SF_MEM_MODE): New.
6533 (avx512pf_gatherpf<mode>df): Ditto.
6534 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6535 (*avx512pf_scatterpf<mode>df): Ditto.
6537 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6540 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6541 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6544 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6546 * common/config/arm/arm-common.c
6547 (arm_rewrite_mcpu): Handle multiple names.
6549 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6551 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6553 * gimple-builder.h (create_gimple_tmp): Delete.
6555 2014-01-27 Christian Bruel <christian.bruel@st.com>
6557 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6560 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6562 * config/pa/pa.md (call): Generate indirect long calls to non-local
6563 functions when outputing 32-bit code.
6564 (call_value): Likewise except for special call to buggy powf function.
6566 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6567 portable runtime and PIC indirect calls.
6568 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6569 and PIC call sequences. Use ldo instead of blr to set return register
6570 in PIC call sequence.
6572 2014-01-25 Walter Lee <walt@tilera.com>
6574 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6575 avoid clobbering a live register.
6577 2014-01-25 Walter Lee <walt@tilera.com>
6579 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6580 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6581 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6582 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6584 2014-01-25 Walter Lee <walt@tilera.com>
6586 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6587 arguments on even registers.
6588 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6590 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6591 (BIGGEST_ALIGNMENT): Ditto.
6592 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6594 2014-01-25 Walter Lee <walt@tilera.com>
6596 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6597 insns before bundling.
6598 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6600 2014-01-25 Walter Lee <walt@tilera.com>
6602 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6603 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6604 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6606 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6608 * config/mips/constraints.md (kl): Delete.
6609 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6610 define expands, using...
6611 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6612 instructions for MIPS16.
6613 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6614 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6616 2014-01-25 Walter Lee <walt@tilera.com>
6618 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6622 2014-01-25 Walter Lee <walt@tilera.com>
6624 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6625 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6627 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6629 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6632 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6634 * print-rtl.c (in_call_function_usage): New var.
6635 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6636 EXPR_LIST mode as mode and not as reg note name.
6639 * cfgloopmanip.c (copy_loop_info): If
6640 loop->warned_aggressive_loop_optimizations, make sure
6641 the flag is set in target loop too.
6643 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6645 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6647 * builtins.def: Likewise.
6648 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6649 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6650 * ira.c (ira_setup_eliminable_regset): Likewise.
6651 * omp-low.c (gate_expand_omp): Likewise.
6652 (execute_lower_omp): Likewise.
6653 (diagnose_sb_0): Likewise.
6654 (gate_diagnose_omp_blocks): Likewise.
6655 (simd_clone_clauses_extract): Likewise.
6658 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6660 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6661 correction for little endian...
6662 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6665 2014-01-24 Jeff Law <law@redhat.com>
6667 PR tree-optimization/59919
6668 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6669 for non-returning calls.
6671 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6673 * common/config/aarch64/aarch64-common.c
6674 (aarch64_rewrite_mcpu): Handle multiple names.
6675 * config/aarch64/aarch64.h
6676 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6678 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6680 * input.c (add_file_to_cache_tab): Handle the case where fopen
6683 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6686 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6687 from push operand if code of push isn't PRE_DEC.
6689 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6692 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6693 -mquad-memory-atomic. Update -mquad-memory documentation to say
6694 it is only used for non-atomic loads/stores.
6696 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6697 -mquad-memory or -mquad-memory-atomic switches.
6699 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6700 -mquad-memory-atomic to ISA 2.07 support.
6702 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6703 to separate support of normal quad word memory operations (ldq, stq)
6704 from the atomic quad word memory operations.
6706 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6707 support to separate non-atomic quad word operations from atomic
6708 quad word operations. Disable non-atomic quad word operations in
6709 little endian mode so that we don't have to swap words after the
6710 load and before the store.
6711 (quad_load_store_p): Add comment about atomic quad word support.
6712 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6713 options printed with -mdebug=reg.
6715 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6716 -mquad-memory-atomic as the test for whether we have quad word
6717 atomic instructions.
6718 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6719 or -mp8-vector are used, allow byte/half-word atomic operations.
6721 * config/rs6000/sync.md (load_lockedti): Insure that the address
6722 is a proper indexed or indirect address for the lqarx instruction.
6723 On little endian systems, swap the hi/lo registers after the lqarx
6725 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6726 insure the address is valid for the lqarx instruction.
6727 (store_conditionalti): Insure that the address is a proper indexed
6728 or indirect address for the stqcrx. instruction. On little endian
6729 systems, swap the hi/lo registers before doing the stqcrx.
6731 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6732 insure the address is valid for the stqcrx. instruction.
6734 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6735 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6736 type of quad memory support is available.
6738 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6741 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6742 there is a danger of looping.
6744 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6746 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6747 force flag_ira_loop_pressure if set via command line.
6749 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6751 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6752 (ashr_simd): New builtin handling DI mode.
6753 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6754 (aarch64_sshr_simddi): New match pattern.
6755 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6756 (vshrd_n_s64): Likewise.
6757 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6759 2014-01-23 Nick Clifton <nickc@redhat.com>
6761 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6762 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6763 favour of mcu specific scripts.
6764 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6767 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6768 Alex Velenko <Alex.Velenko@arm.com>
6770 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6771 (vaddv_s16): Likewise.
6772 (vaddv_s32): Likewise.
6773 (vaddv_u8): Likewise.
6774 (vaddv_u16): Likewise.
6775 (vaddv_u32): Likewise.
6776 (vaddvq_s8): Likewise.
6777 (vaddvq_s16): Likewise.
6778 (vaddvq_s32): Likewise.
6779 (vaddvq_s64): Likewise.
6780 (vaddvq_u8): Likewise.
6781 (vaddvq_u16): Likewise.
6782 (vaddvq_u32): Likewise.
6783 (vaddvq_u64): Likewise.
6784 (vaddv_f32): Likewise.
6785 (vaddvq_f32): Likewise.
6786 (vaddvq_f64): Likewise.
6787 (vmaxv_f32): Likewise.
6788 (vmaxv_s8): Likewise.
6789 (vmaxv_s16): Likewise.
6790 (vmaxv_s32): Likewise.
6791 (vmaxv_u8): Likewise.
6792 (vmaxv_u16): Likewise.
6793 (vmaxv_u32): Likewise.
6794 (vmaxvq_f32): Likewise.
6795 (vmaxvq_f64): Likewise.
6796 (vmaxvq_s8): Likewise.
6797 (vmaxvq_s16): Likewise.
6798 (vmaxvq_s32): Likewise.
6799 (vmaxvq_u8): Likewise.
6800 (vmaxvq_u16): Likewise.
6801 (vmaxvq_u32): Likewise.
6802 (vmaxnmv_f32): Likewise.
6803 (vmaxnmvq_f32): Likewise.
6804 (vmaxnmvq_f64): Likewise.
6805 (vminv_f32): Likewise.
6806 (vminv_s8): Likewise.
6807 (vminv_s16): Likewise.
6808 (vminv_s32): Likewise.
6809 (vminv_u8): Likewise.
6810 (vminv_u16): Likewise.
6811 (vminv_u32): Likewise.
6812 (vminvq_f32): Likewise.
6813 (vminvq_f64): Likewise.
6814 (vminvq_s8): Likewise.
6815 (vminvq_s16): Likewise.
6816 (vminvq_s32): Likewise.
6817 (vminvq_u8): Likewise.
6818 (vminvq_u16): Likewise.
6819 (vminvq_u32): Likewise.
6820 (vminnmv_f32): Likewise.
6821 (vminnmvq_f32): Likewise.
6822 (vminnmvq_f64): Likewise.
6824 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6826 * config/aarch64/aarch64-simd.md
6827 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
6828 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
6829 (*aarch64_mul3_elt<mode>): Likewise.
6830 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
6831 (*aarch64_mul3_elt_to_64v2df): Likewise.
6832 (*aarch64_mla_elt<mode>): Likewise.
6833 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6834 (*aarch64_mls_elt<mode>): Likewise.
6835 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6836 (*aarch64_fma4_elt<mode>): Likewise.
6837 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6838 (*aarch64_fma4_elt_to_64v2df): Likewise.
6839 (*aarch64_fnma4_elt<mode>): Likewise.
6840 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6841 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6842 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6843 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6844 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6845 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6846 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6847 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6848 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6850 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
6852 * config/aarch64/aarch64-simd.md
6853 (aarch64_be_checked_get_lane<mode>): New define_expand.
6854 * config/aarch64/aarch64-simd-builtins.def
6855 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
6856 New builtin definition.
6857 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
6858 Use new safe be builtin.
6860 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6862 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
6864 (aarch64_be_st1<mode>): Likewise.
6865 (aarch_ld1<VALL:mode>): Define_expand modified.
6866 (aarch_st1<VALL:mode>): Likewise.
6867 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
6868 (UNSPEC_ST1): Likewise.
6870 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
6872 * config/microblaze/microblaze.md: Add trap insn and attribute
6874 2014-01-23 Dodji Seketeli <dodji@redhat.com>
6876 PR preprocessor/58580
6877 * input.h (location_get_source_line): Take an additional line_size
6879 (void diagnostics_file_cache_fini): Declare new function.
6880 * input.c (struct fcache): New type.
6881 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
6882 New static constants.
6883 (diagnostic_file_cache_init, total_lines_num)
6884 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
6885 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
6886 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
6887 (get_next_line, read_next_line, goto_next_line, read_line_num):
6888 New static function definitions.
6889 (diagnostic_file_cache_fini): New function.
6890 (location_get_source_line): Take an additional output line_len
6891 parameter. Re-write using lookup_or_add_file_to_cache_tab and
6893 * diagnostic.c (diagnostic_finish): Call
6894 diagnostic_file_cache_fini.
6895 (adjust_line): Take an additional input parameter for the length
6896 of the line, rather than calculating it with strlen.
6897 (diagnostic_show_locus): Adjust the use of
6898 location_get_source_line and adjust_line with respect to their new
6899 signature. While displaying a line now, do not stop at the first
6900 null byte. Rather, display the zero byte as a space and keep
6901 going until we reach the size of the line.
6902 * Makefile.in: Add vec.o to OBJS-libcommon
6904 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6905 Ilya Tocar <ilya.tocar@intel.com>
6907 * config/i386/avx512fintrin.h (_mm512_kmov): New.
6908 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
6909 (__builtin_ia32_kmov16): Ditto.
6910 * config/i386/i386.md (UNSPEC_KMOV): New.
6913 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6915 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
6916 (_mm512_storeu_si512): Ditto.
6918 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
6921 * rtl.h (get_referenced_operands): Declare.
6922 * recog.c (get_referenced_operands): New function.
6923 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
6924 operands have been referenced when recording LO_SUM references.
6926 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
6928 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
6930 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
6932 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
6933 Enable for generic and recent AMD targets.
6935 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
6937 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
6938 ARG_SIZE note when adjustment was eliminated.
6940 2014-01-22 Jeff Law <law@redhat.com>
6942 PR tree-optimization/59597
6943 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
6944 in file. Accept new argument REGISTERING and use it to modify
6945 dump output appropriately.
6946 (register_jump_thread): Corresponding changes.
6947 (mark_threaded_blocks): Reinstate code to cancel unprofitable
6948 thread paths involving joiner blocks. Add code to dump cancelled
6949 jump threading paths.
6951 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
6953 PR rtl-optimization/59477
6954 * lra-constraints.c (inherit_in_ebb): Process call for living hard
6955 regs. Update reloads_num and potential_reload_hard_regs for all insns.
6957 2014-01-22 Tom Tromey <tromey@redhat.com>
6959 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
6961 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
6963 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
6965 PR rtl-optimization/59896
6966 * lra-constraints.c (process_alt_operands): Check unused note for
6967 matched operands of insn with no output reloads.
6969 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
6971 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
6972 (mips_move_from_gpr_cost): Likewise.
6974 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
6976 PR rtl-optimization/59858
6977 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
6978 ira_class_hard_regs_num.
6979 (process_alt_operands): Increase reject for dying matched operand.
6981 2014-01-21 Jakub Jelinek <jakub@redhat.com>
6984 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
6985 smaller than size, perform several stores or loads and stores
6986 at dst + count - size to store or copy all of size bytes, rather
6987 than just last modesize bytes.
6989 2014-01-20 DJ Delorie <dj@redhat.com>
6991 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
6992 that CLOBBERs are REGs before propogating their values.
6994 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
6997 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
6998 (cgraph_inline_failed_type): New function.
6999 * cgraph.h (DEFCIFCODE): Add type.
7000 (cgraph_inline_failed_type_t): New enum.
7001 (cgraph_inline_failed_type): New prototype.
7002 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7003 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7004 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7005 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7006 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7007 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7008 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7009 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7010 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7011 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7012 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7013 OPTIMIZATION_MISMATCH.
7014 * tree-inline.c (expand_call_inline): Emit errors during
7015 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7017 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7020 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7021 mode attribute in insn output.
7023 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7025 * output.h (output_constant): Delete.
7026 * varasm.c (output_constant): Make private.
7028 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7030 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7032 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7035 * tree.h (fold_builtin_strcat): New prototype.
7036 * builtins.c (fold_builtin_strcat): No longer static. Add len
7037 argument, if non-NULL, don't call c_strlen. Optimize
7038 directly into __builtin_memcpy instead of __builtin_strcpy.
7039 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7040 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7042 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7044 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7045 for SImode_address_operand operands, having only a REG argument.
7047 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7049 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7050 loader name using mbig-endian.
7051 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7053 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7055 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7059 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7061 * config/aarch64/aarch64-protos.h
7062 (aarch64_cannot_change_mode_class_ptr): Declare.
7063 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7064 aarch64_cannot_change_mode_class_ptr): New.
7065 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7066 backend hook aarch64_cannot_change_mode_class.
7068 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7070 * common/config/aarch64/aarch64-common.c
7071 (aarch64_handle_option): Don't handle any option order logic here.
7072 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7073 selected_cpu, warn on architecture version mismatch.
7074 (aarch64_override_options): Fix parsing order for option strings.
7076 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7077 Iain Sandoe <iain@codesourcery.com>
7080 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7081 warning. Amend comment to reflect current functionality.
7083 2014-01-20 Richard Biener <rguenther@suse.de>
7086 * builtins.c (fold_builtin_strcat): Remove case better handled
7087 by tree-ssa-strlen.c.
7089 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7091 * config/aarch64/aarch64.opt
7092 (mcpu, march, mtune): Make case-insensitive.
7094 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7097 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7098 if operands[1] is a REG or ZERO_EXTEND of a REG.
7100 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7102 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7104 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7106 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7107 long non-pic millicode calls.
7109 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7111 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7113 2014-01-19 Kito Cheng <kito@0xlab.org>
7115 * builtins.c (expand_movstr): Check movstr expand done or fail.
7117 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7118 H.J. Lu <hongjiu.lu@intel.com>
7121 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7122 to DImode for zero-extended addresses.
7124 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7126 PR rtl-optimization/57763
7127 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7128 on the new indirect jump_insn and increment LABEL_NUSES (label).
7130 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7134 * config.gcc (x86_archs): New variable.
7135 (x86_64_archs): Likewise.
7136 (x86_cpus): Likewise.
7137 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7138 --with-arch/--with-cpu= options.
7139 Support --with-arch=/--with-cpu={nehalem,westmere,
7140 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7142 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7144 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7145 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7147 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7149 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7151 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7154 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7155 clear cpp_get_options (parse_in)->warn_unused_macros for
7156 ix86_target_macros_internal with cpp_define.
7158 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7160 * jump.c (delete_related_insns): Keep (use (insn))s.
7161 * reorg.c (redundant_insn): Check for barriers too.
7163 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7165 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7167 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7169 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7170 call to $$dyncall when TARGET_LONG_CALLS is true.
7172 2014-01-17 Jeff Law <law@redhat.com>
7174 * ree.c (combine_set_extension): Temporarily disable test for
7175 changing number of hard registers.
7177 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7180 * ipa-inline-analysis.c (inline_free_summary):
7181 Do not free summary of aliases.
7183 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7186 * gimplify.c (gimplify_expr): Use create_tmp_var
7187 instead of create_tmp_var_raw. If cond doesn't have
7188 integral type, don't add the IFN_ANNOTATE builtin at all.
7190 2014-01-17 Martin Jambor <mjambor@suse.cz>
7193 * ipa-cp.c (prev_edge_clone): New variable.
7194 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7195 Also resize prev_edge_clone vector.
7196 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7197 (ipcp_edge_removal_hook): New function.
7198 (ipcp_driver): Register ipcp_edge_removal_hook.
7200 2014-01-17 Andrew Pinski <apinski@cavium.com>
7201 Steve Ellcey <sellcey@mips.com>
7204 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7207 2014-01-17 Jeff Law <law@redhat.com>
7210 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7211 so that pass_ccp runs first.
7213 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7215 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7216 (ix86_adjust_cost): Use !TARGET_XXX.
7217 (do_reorder_for_imul): Likewise.
7218 (swap_top_of_ready_list): Likewise.
7219 (ix86_sched_reorder): Likewise.
7221 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7223 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7224 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7225 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7226 (intel_memset): New. Duplicate slm_memset.
7227 (intel_cost): New. Duplicate slm_cost.
7228 (m_INTEL): New macro.
7229 (processor_target_table): Add "intel".
7230 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7231 with PROCESSOR_INTEL for "intel".
7232 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7233 PROCESSOR_SILVERMONT.
7234 (ix86_issue_rate): Likewise.
7235 (ix86_adjust_cost): Likewise.
7236 (ia32_multipass_dfa_lookahead): Likewise.
7237 (swap_top_of_ready_list): Likewise.
7238 (ix86_sched_reorder): Likewise.
7239 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7240 instead of TARGET_OPT_AGU.
7241 * config/i386/i386.h (TARGET_INTEL): New.
7242 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7243 (processor_type): Add PROCESSOR_INTEL.
7244 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7245 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7247 2014-01-17 Marek Polacek <polacek@redhat.com>
7250 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7253 2014-01-17 Richard Biener <rguenther@suse.de>
7255 PR tree-optimization/46590
7256 * opts.c (default_options_table): Add entries for
7257 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7258 all enabled at -O1 but not for -Og.
7259 * common.opt (fbranch-count-reg): Remove Init(1).
7260 (fmove-loop-invariants): Likewise.
7261 (ftree-pta): Likewise.
7263 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7265 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7266 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7267 decls to at least the GCC 4.8 used alignments.
7270 * tree-nested.c (convert_nonlocal_reference_stmt,
7271 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7272 of GIMPLE_BIND stmts, adjust associated decls.
7274 2014-01-17 Richard Biener <rguenther@suse.de>
7276 PR tree-optimization/46590
7277 * vec.h (vec<>::bseach): New member function implementing
7278 binary search according to C89 bsearch.
7279 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7280 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7281 bitmap pointer again. Make accesses_in_loop a flat array.
7282 (mem_ref_obstack): New global.
7283 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7284 (mark_ref_stored): Likewise.
7285 (ref_indep_loop_p_2): Likewise.
7286 (set_ref_stored_in_loop): New helper function.
7287 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7288 (memref_free): Adjust.
7289 (record_mem_ref_loc): Simplify.
7290 (gather_mem_refs_stmt): Adjust.
7291 (sort_locs_in_loop_postorder_cmp): New function.
7292 (analyze_memory_references): Sort accesses_in_loop after
7293 loop postorder number.
7294 (find_ref_loc_in_loop_cmp): New function.
7295 (for_all_locs_in_loop): Find relevant cluster of locs in
7296 accesses_in_loop and iterate without recursion.
7297 (execute_sm): Avoid uninit warning.
7298 (struct ref_always_accessed): Simplify.
7299 (ref_always_accessed::operator ()): Likewise.
7300 (ref_always_accessed_p): Likewise.
7301 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7302 loop postorder numbers here.
7303 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7306 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7309 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7310 on decls for which assemble_alias has been called.
7312 2014-01-17 Nick Clifton <nickc@redhat.com>
7314 * config/msp430/msp430.opt: (mcpu): New option.
7315 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7316 (msp430_option_override): Parse target_cpu. If the MCU name
7317 matches a generic string, clear target_mcu.
7318 (msp430_attr): Allow numeric interrupt values up to 63.
7319 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7320 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7322 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7324 * config/msp430/msp430.md (popm): Use %J rather than %I.
7325 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7326 (addhi_cy_i): Use immediate_operand for operand 2.
7327 * doc/invoke.texi: Document -mcpu option.
7329 2014-01-17 Richard Biener <rguenther@suse.de>
7331 PR rtl-optimization/38518
7332 * df.h (df_analyze_loop): Declare.
7333 * df-core.c: Include cfgloop.h.
7334 (df_analyze_1): Split out main part of df_analyze.
7335 (df_analyze): Adjust.
7336 (loop_inverted_post_order_compute): New function.
7337 (loop_post_order_compute): Likewise.
7338 (df_analyze_loop): New function avoiding whole-function
7340 * loop-invariant.c (find_defs): Use df_analyze_loop.
7341 (find_invariants): Adjust.
7342 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7344 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7346 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7347 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7349 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7351 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7352 traversal when removing references.
7354 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7357 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7359 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7362 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7363 pushing a reload for an autoinc when we had previously reloaded an
7364 inner part of the address.
7366 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7368 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7370 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7371 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7372 when not giving up or versioning for alias only because of
7374 (vect_analyze_data_ref_dependences): Set to true.
7375 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7377 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7378 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7382 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7385 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7386 operand 0 predicate for gathers, use a new pseudo as subtarget.
7388 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7391 * lra-constraints.c (process_alt_operands): Add printing debug info.
7392 Check absence of input/output reloads for matched operands too.
7394 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7396 PR rtl-optimization/59835
7397 * ira.c (ira_init_register_move_cost): Increase cost for
7400 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7402 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7404 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7407 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7408 non-register objects. Use gen_(high/low)part more consistently.
7411 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7414 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7415 endian support, remove tests for WORDS_BIG_ENDIAN.
7416 (p8_mfvsrd_3_<mode>): Likewise.
7417 (reload_gpr_from_vsx<mode>): Likewise.
7418 (reload_gpr_from_vsxsf): Likewise.
7419 (p8_mfvsrd_4_disf): Likewise.
7421 2014-01-16 Richard Biener <rguenther@suse.de>
7423 PR rtl-optimization/46590
7424 * lcm.c (compute_antinout_edge): Use postorder iteration.
7425 (compute_laterin): Use inverted postorder iteration.
7427 2014-01-16 Nick Clifton <nickc@redhat.com>
7430 * varasm.c (output_constant): Return the number of bytes actually
7432 (output_constructor_array_range): Update the field size with the
7433 number of bytes emitted by output_constant.
7434 (output_constructor_regular_field): Likewise. Also do not
7435 complain if the total number of bytes emitted is now greater
7436 than the expected fieldpos.
7437 * output.h (output_constant): Update prototype and descriptive comment.
7439 2014-01-16 Marek Polacek <polacek@redhat.com>
7442 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7443 it is error_mark_node.
7445 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7447 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7448 VALID_AVX256_REG_OR_OI_MODE.
7450 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7452 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7453 current procedure should be profiled.
7455 2014-01-15 Andrew Pinski <apinski@cavium.com>
7457 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7458 of moving from/to the STACK_REG register class.
7460 2014-01-15 Richard Henderson <rth@redhat.com>
7463 * reginfo.c (global_regs_decl): Globalize.
7464 * rtl.h (global_regs_decl): Declare.
7465 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7466 reserved via global_regs.
7468 2014-01-15 Teresa Johnson <tejohnson@google.com>
7470 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7472 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7474 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7475 and vmulosh rather than call gen_vec_widen_smult_*.
7476 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7477 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7478 (vec_widen_smult_even_v16qi): Likewise.
7479 (vec_widen_umult_even_v8hi): Likewise.
7480 (vec_widen_smult_even_v8hi): Likewise.
7481 (vec_widen_umult_odd_v16qi): Likewise.
7482 (vec_widen_smult_odd_v16qi): Likewise.
7483 (vec_widen_umult_odd_v8hi): Likewise.
7484 (vec_widen_smult_odd_v8hi): Likewise.
7485 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7486 vmuloub rather than call gen_vec_widen_umult_*.
7487 (vec_widen_umult_lo_v16qi): Likewise.
7488 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7489 vmulosb rather than call gen_vec_widen_smult_*.
7490 (vec_widen_smult_lo_v16qi): Likewise.
7491 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7492 rather than call gen_vec_widen_umult_*.
7493 (vec_widen_umult_lo_v8hi): Likewise.
7494 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7495 rather than call gen_vec_widen_smult_*.
7496 (vec_widen_smult_lo_v8hi): Likewise.
7498 2014-01-15 Jeff Law <law@redhat.com>
7500 PR tree-optimization/59747
7501 * ree.c (find_and_remove_re): Properly handle case where a second
7502 eliminated extension requires widening a copy created for elimination
7503 of a prior extension.
7504 (combine_set_extension): Ensure that the number of hard regs needed
7505 for a destination register does not change when we widen it.
7507 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7509 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7510 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7511 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7512 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7513 (avr-*-rtems*): Likewise.
7514 (bfin*-rtems*): Likewise.
7515 (moxie-*-rtems*): Likewise.
7516 (h8300-*-rtems*): Likewise.
7517 (i[34567]86-*-rtems*): Likewise.
7518 (lm32-*-rtems*): Likewise.
7519 (m32r-*-rtems*): Likewise.
7520 (m68k-*-rtems*): Likewise.
7521 (microblaze*-*-rtems*): Likewise.
7522 (mips*-*-rtems*): Likewise.
7523 (powerpc-*-rtems*): Likewise.
7524 (sh-*-rtems*): Likewise.
7525 (sparc-*-rtems*): Likewise.
7526 (sparc64-*-rtems*): Likewise.
7527 (v850-*-rtems*): Likewise.
7528 (m32c-*-rtems*): Likewise.
7530 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7532 PR rtl-optimization/59511
7533 * ira.c (ira_init_register_move_cost): Use memory costs for some
7534 cases of register move cost calculations.
7535 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7536 instead of BB frequency.
7537 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7538 * lra-assigns.c (find_hard_regno_for): Ditto.
7540 2014-01-15 Richard Biener <rguenther@suse.de>
7542 PR tree-optimization/59822
7543 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7544 (vectorizable_load): Use it to hoist defs of uses of invariant
7545 loads out of the loop.
7547 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7548 Kugan Vivekanandarajah <kuganv@linaro.org>
7551 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7554 2014-01-15 Richard Biener <rguenther@suse.de>
7556 PR rtl-optimization/59802
7557 * lcm.c (compute_available): Use inverted postorder to seed
7558 the initial worklist.
7560 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7563 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7564 ADDR_REGS for invalid symrefs in non-PIC code.
7566 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7569 * builtins.c (determine_block_size): Initialize *probable_max_size
7570 even if len_rtx is CONST_INT.
7572 2014-01-14 Andrew Pinski <apinski@cavium.com>
7574 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7575 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7576 (cortexa53_tunings): Likewise.
7577 (aarch64_sched_issue_rate): New function.
7578 (TARGET_SCHED_ISSUE_RATE): Define.
7580 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7582 * ira-costs.c (find_costs_and_classes): Add missed
7583 ira_init_register_move_cost_if_necessary.
7585 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7588 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7590 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7593 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7594 to indicate if type is used for function return value. Warn ABI
7595 change if the vector mode isn't available for function return value.
7596 (ix86_function_arg_advance): Pass false to type_natural_mode.
7597 (ix86_function_arg): Likewise.
7598 (ix86_gimplify_va_arg): Likewise.
7599 (function_arg_32): Don't warn ABI change.
7600 (ix86_function_value): Pass true to type_natural_mode.
7601 (ix86_return_in_memory): Likewise.
7602 (ix86_struct_value_rtx): Removed.
7603 (TARGET_STRUCT_VALUE_RTX): Likewise.
7605 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7607 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7608 converting a conditional jump into a conditional return.
7610 2014-01-14 Richard Biener <rguenther@suse.de>
7612 PR tree-optimization/58921
7613 PR tree-optimization/59006
7614 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7615 hoisting invariant stmts.
7616 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7617 invariant loads on the preheader edge if possible.
7619 2014-01-14 Joey Ye <joey.ye@arm.com>
7621 * doc/plugin.texi (Building GCC plugins): Update to C++.
7623 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7625 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7626 (_mm_rcp28_round_ss): Ditto.
7627 (_mm_rsqrt28_round_sd): Ditto.
7628 (_mm_rsqrt28_round_ss): Ditto.
7629 (_mm_rcp28_sd): Ditto.
7630 (_mm_rcp28_ss): Ditto.
7631 (_mm_rsqrt28_sd): Ditto.
7632 (_mm_rsqrt28_ss): Ditto.
7633 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7634 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7635 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7636 (IX86_BUILTIN_RCP28SD): Ditto.
7637 (IX86_BUILTIN_RCP28SS): Ditto.
7638 (IX86_BUILTIN_RSQRT28SD): Ditto.
7639 (IX86_BUILTIN_RSQRT28SS): Ditto.
7640 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7641 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7642 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7643 (ix86_expand_special_args_builtin): Expand new FTYPE.
7644 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7645 (srcp14<mode>): Make insn unary.
7646 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7647 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7648 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7649 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7650 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7651 Fix rounding: make it SAE only.
7652 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7654 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7656 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7657 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7658 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7659 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7660 (round_saeonly_mask_scalar_operand4): Ditto.
7661 (round_saeonly_mask_scalar_op3): Ditto.
7662 (round_saeonly_mask_scalar_op4): Ditto.
7664 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7666 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7667 Implement -maltivec=be for vec_insert and vec_extract.
7669 2014-01-10 DJ Delorie <dj@redhat.com>
7671 * config/msp430/msp430.md (call_internal): Don't allow memory
7672 references with SP as the base register.
7673 (call_value_internal): Likewise.
7674 * config/msp430/constraints.md (Yc): New. For memory references
7675 that don't use SP as a base register.
7677 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7678 "an integer without a # prefix"
7679 * config/msp430/msp430.md (epilogue_helper): Use it.
7681 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7684 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7685 AVX512F gather builtins.
7686 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7687 on gather decls with INTEGER_TYPE masktype.
7688 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7689 directly into the builtin rather than hoisting it before loop.
7691 PR tree-optimization/59387
7692 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7693 (scev_const_prop): If folded_casts and type has undefined overflow,
7694 use force_gimple_operand instead of force_gimple_operand_gsi and
7695 for each added stmt if it is assign with
7696 arith_code_with_undefined_signed_overflow, call
7697 rewrite_to_defined_overflow.
7698 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7699 gimple-fold.h instead.
7700 (arith_code_with_undefined_signed_overflow,
7701 rewrite_to_defined_overflow): Moved to ...
7702 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7703 rewrite_to_defined_overflow): ... here. No longer static.
7704 Include gimplify-me.h.
7705 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7706 rewrite_to_defined_overflow): New prototypes.
7708 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7710 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7712 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7714 * builtins.c (get_object_alignment_2): Minor tweak.
7715 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7717 2014-01-13 Christian Bruel <christian.bruel@st.com>
7719 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7720 optimized non constant lengths.
7722 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7725 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7726 load as __atomic_load_N if possible.
7728 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7730 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7732 (rs6000_expand_builtin): Adjust call.
7734 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7737 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7738 * config/rs6000/rs6000.c: Include target-globals.h.
7739 (rs6000_set_current_function): Instead of doing target_reinit
7740 unconditionally, use save_target_globals_default_opts and
7741 restore_target_globals.
7743 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7745 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7746 (rs6000_expand_builtin): Handle mffs and mtfsf.
7747 (rs6000_init_builtins): Define mffs and mtfsf.
7748 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7749 (rs6000_mffs): New pattern.
7750 (rs6000_mtfsf): New pattern.
7752 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7754 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7755 Start narrowing with START. Apply candidate-use pair
7756 and check overall cost in narrowing.
7757 (iv_ca_prune): Pass new argument.
7759 2014-01-10 Jeff Law <law@redhat.com>
7762 * ree.c (combine_reaching_defs): Ensure the defining statement
7763 occurs before the extension when optimizing extensions with
7764 different source and destination hard registers.
7766 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7769 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7770 vtables into the type inheritance graph.
7772 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7774 PR rtl-optimization/59754
7775 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7776 modes in the REGNO != REGNO case.
7778 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7780 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7782 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7784 PR tree-optimization/59745
7785 * tree-predcom.c (tree_predictive_commoning_loop): Call
7786 free_affine_expand_cache if giving up because components is NULL.
7788 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7789 GC in payload of target_globals struct instead of allocating them on
7790 the heap and the larger structs separately using GC.
7791 * target-globals.h (struct target_globals): Make regs, hard_regs,
7792 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7793 of GTY((skip)) and change type to void *.
7794 (reset_target_globals): Cast loads from those fields to corresponding
7797 2014-01-10 Steve Ellcey <sellcey@mips.com>
7800 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7801 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7802 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7804 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7807 * aarch64-modes.def (CC_Zmode): New flags mode.
7808 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7809 represents an equality.
7810 (aarch64_get_condition_code): Handle CC_Zmode.
7811 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
7813 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7815 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
7816 extraction in good case.
7818 2014-01-10 Richard Biener <rguenther@suse.de>
7820 PR tree-optimization/59374
7821 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
7822 checking after SLP discovery. Mark stmts not participating
7823 in any SLP instance properly.
7825 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7827 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
7828 when handling a SET rtx.
7830 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7832 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
7833 (cortex-a57): Likewise.
7834 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
7836 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7838 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
7839 non-iwmmxt builtins.
7841 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7845 * ipa-devirt.c record_target_from_binfo): Take as argument
7846 stack of binfos and lookup matching one for virtual inheritance.
7847 (possible_polymorphic_call_targets_1): Update.
7849 2014-01-10 Huacai Chen <chenhc@lemote.com>
7851 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
7852 kernel strings for Loongson-2E/2F/3A.
7854 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7857 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
7858 is_gimple_call before calling gimple_call_internal_p.
7860 2014-01-09 Steve Ellcey <sellcey@mips.com>
7862 * Makefile.in (TREE_FLOW_H): Remove.
7863 (TREE_SSA_H): Add file names from tree-flow.h.
7864 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
7865 * tree.h: Remove tree-flow.h reference.
7866 * hash-table.h: Remove tree-flow.h reference.
7867 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
7868 reference with tree-ssa-loop.h.
7870 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7872 * doc/invoke.texi: Add -maltivec={be,le} options, and document
7873 default element-order behavior for -maltivec.
7874 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
7875 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
7876 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
7877 when targeting big endian, at least for now.
7878 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
7880 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7883 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
7884 var satisfies use_register_for_decl, just take into account type
7885 alignment, rather than decl alignment.
7887 PR tree-optimization/59622
7888 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
7889 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
7890 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
7891 Don't devirtualize for inplace at all. For targets.length () == 1,
7892 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
7894 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7896 * config/i386/i386.md (cpu): Remove the unused btver1.
7898 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7900 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
7902 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7905 * tree-core.h (struct target_globals): New forward declaration.
7906 (struct tree_target_option): Add globals field.
7907 * tree.h (TREE_TARGET_GLOBALS): Define.
7908 (prepare_target_option_nodes_for_pch): New prototype.
7909 * target-globals.h (struct target_globals): Define even if
7911 * tree.c (prepare_target_option_node_for_pch,
7912 prepare_target_option_nodes_for_pch): New functions.
7913 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
7914 * config/i386/i386.c: Include target-globals.h.
7915 (ix86_set_current_function): Instead of doing target_reinit
7916 unconditionally, use save_target_globals_default_opts and
7917 restore_target_globals.
7919 2014-01-09 Richard Biener <rguenther@suse.de>
7921 PR tree-optimization/59715
7922 * tree-cfg.h (split_critical_edges): Declare.
7923 * tree-cfg.c (split_critical_edges): Export.
7924 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
7926 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
7928 * cfgexpand.c (expand_stack_vars): Optionally disable
7929 asan stack protection.
7930 (expand_used_vars): Likewise.
7931 (partition_stack_vars): Likewise.
7932 * asan.c (asan_emit_stack_protection): Optionally disable
7933 after return stack usage.
7934 (instrument_derefs): Optionally disable memory access instrumentation.
7935 (instrument_builtin_call): Likewise.
7936 (instrument_strlen_call): Likewise.
7937 (asan_protect_global): Optionally disable global variables protection.
7938 * doc/invoke.texi: Added doc for new options.
7939 * params.def: Added new options.
7940 * params.h: Likewise.
7942 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7944 PR rtl-optimization/59724
7945 * ifcvt.c (cond_exec_process_if_block): Don't call
7946 flow_find_head_matching_sequence with 0 longest_match.
7947 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
7948 non-active insns if !stop_after.
7949 (try_head_merge_bb): Revert 2014-01-07 changes.
7951 2014-01-08 Jeff Law <law@redhat.com>
7953 * ree.c (get_sub_rtx): New function, extracted from...
7954 (merge_def_and_ext): Here.
7955 (combine_reaching_defs): Use get_sub_rtx.
7957 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
7959 * cgraph.h (varpool_variable_node): Do not choke on null node.
7961 2014-01-08 Catherine Moore <clm@codesourcery.com>
7963 * config/mips/mips.md (simple_return): Attempt to use JRC
7965 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
7967 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
7969 PR rtl-optimization/59137
7970 * reorg.c (steal_delay_list_from_target): Call update_block for
7972 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
7974 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7976 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
7977 two duplicate entries.
7979 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
7982 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
7984 * config/mips/mips.c (mips_truncated_op_cost): New function.
7985 (mips_rtx_costs): Adjust test for BADDU.
7986 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
7988 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
7990 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
7991 (*baddu_si): ...this new pattern.
7993 2014-01-08 Jakub Jelinek <jakub@redhat.com>
7996 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
7998 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8001 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8003 (expand_expr, expand_normal): Adjust.
8004 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8005 inner_reference_p. Use inner_reference_p to expand inner references.
8006 (store_expr): Adjust.
8007 * cfgexpand.c (expand_call_stmt): Adjust.
8009 2014-01-08 Rong Xu <xur@google.com>
8011 * gcov-io.c (gcov_var): Move from gcov-io.h.
8012 (gcov_position): Ditto.
8013 (gcov_is_error): Ditto.
8014 (gcov_rewrite): Ditto.
8015 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8016 only part to libgcc/libgcov.h.
8018 2014-01-08 Marek Polacek <polacek@redhat.com>
8021 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8023 2014-01-08 Marek Polacek <polacek@redhat.com>
8026 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8028 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8030 PR rtl-optimization/59649
8031 * stor-layout.c (get_mode_bounds): For BImode return
8032 0 and STORE_FLAG_VALUE.
8034 2014-01-08 Richard Biener <rguenther@suse.de>
8037 * gimple.h (is_gimple_builtin_call): Remove.
8038 (gimple_builtin_call_types_compatible_p): New.
8039 (gimple_call_builtin_p): New overload.
8040 * gimple.c (is_gimple_builtin_call): Remove.
8041 (validate_call): Rename to ...
8042 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8044 (validate_type): New static function.
8045 (gimple_call_builtin_p): New overload and adjust.
8046 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8047 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8048 (gimple_fold_stmt_to_constant_1): Likewise.
8049 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8051 2014-01-08 Richard Biener <rguenther@suse.de>
8054 * gimplify.c (gimplify_expr): Gimplify register-register type
8055 VIEW_CONVERT_EXPRs to separate stmts.
8057 2014-01-07 Jeff Law <law@redhat.com>
8060 * ree.c (combine_set_extension): Handle case where source
8061 and destination registers in an extension insn are different.
8062 (combine_reaching_defs): Allow source and destination registers
8063 in extension to be different under limited circumstances.
8064 (add_removable_extension): Remove restriction that the
8065 source and destination registers in the extension are the same.
8066 (find_and_remove_re): Emit a copy from the extension's
8067 destination to its source after the defining insn if
8068 the source and destination registers are different.
8071 * ifcvt.c (merge_if_block): If we are merging a block with more than
8072 one successor with a block with no successors, remove any BARRIER
8073 after the second block.
8075 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8077 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8079 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8082 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8083 for 14-bit register offsets when INT14_OK_STRICT is false.
8085 2014-01-07 Roland Stigge <stigge@antcom.de>
8086 Michael Meissner <meissner@linux.vnet.ibm.com>
8089 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8090 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8092 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8094 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8097 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8099 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8100 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8101 rtx is const0_rtx or not.
8103 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8106 * target-globals.c (save_target_globals): Remove this_fn_optab
8108 * toplev.c: Include optabs.h.
8109 (target_reinit): Temporarily restore the global options if another
8110 set of options are in force.
8112 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8114 PR rtl-optimization/58668
8115 * cfgcleanup.c (flow_find_cross_jump): Don't count
8116 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8117 to determine what is counted.
8118 (flow_find_head_matching_sequence): Use active_insn_p to determine
8120 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8122 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8123 determine what is counted.
8125 PR tree-optimization/59643
8126 * tree-predcom.c (split_data_refs_to_components): If one dr is
8127 read and one write, determine_offset fails and the write isn't
8128 in the bad component, just put the read into the bad component.
8130 2014-01-07 Mike Stump <mikestump@comcast.net>
8131 Jakub Jelinek <jakub@redhat.com>
8134 * tree-core.h (struct tree_optimization_option): Change optabs
8135 type from unsigned char * to void *.
8136 * optabs.c (init_tree_optimization_optabs): Adjust
8137 TREE_OPTIMIZATION_OPTABS initialization.
8139 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8142 * config/i386/i386.h (struct machine_function): Add
8143 no_drap_save_restore field.
8144 * config/i386/i386.c (ix86_save_reg): Use
8145 !cfun->machine->no_drap_save_restore instead of
8146 crtl->stack_realign_needed.
8147 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8148 this function clears frame_pointer_needed. Set
8149 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8150 and DRAP reg is needed.
8152 2014-01-06 Marek Polacek <polacek@redhat.com>
8155 * doc/implement-c.texi: Mention that other integer types are
8156 permitted as bit-field types in strictly conforming mode.
8158 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8160 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8163 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8165 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8167 2014-01-06 Martin Jambor <mjambor@suse.cz>
8170 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8172 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8174 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8178 * var-tracking.c (add_stores): Preserve the value of the source even if
8179 we don't record the store.
8181 2014-01-06 Terry Guo <terry.guo@arm.com>
8183 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8185 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8188 * config/darwin.c (darwin_function_section): Adjust return values to
8189 correspond to optimisation changes made in r206070.
8191 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8193 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8194 from prefetch_block tune setting.
8195 (nocona_cost): Correct size of prefetch block to 64.
8197 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8199 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8200 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8201 used to save the static chain register in the computation of the offset
8202 from which the FP registers need to be restored.
8204 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8206 PR tree-optimization/59519
8207 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8208 ICE if get_current_def (current_new_name) is already non-NULL, as long
8209 as it is a phi result of some other phi in *new_exit_bb that has
8212 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8213 or vmovdqu* for misaligned_operand.
8214 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8215 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8216 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8217 aligned_mem for AVX512F masked aligned load and store builtins and for
8220 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8222 PR tree-optimization/59651
8223 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8224 Address range for negative step should be added by TYPE_SIZE_UNIT.
8226 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8228 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8229 overlapping registers also for registers other than the stack pointer.
8231 2014-01-03 Marek Polacek <polacek@redhat.com>
8234 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8237 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8240 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8243 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8244 (push splitter): Use <P:MODE_SIZE> instead of
8245 GET_MODE_SIZE (<P:MODE>mode).
8246 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8247 (mov -1, reg peephole2): Likewise.
8248 * config/i386/sse.md (*mov<mode>_internal,
8249 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8250 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8251 *<code><mode>3, *andnot<mode>3<mask_name>,
8252 <mask_codefor><code><mode>3<mask_name>): Likewise.
8253 * config/i386/subst.md (mask_mode512bit_condition,
8254 sd_mask_mode512bit_condition): Likewise.
8256 2014-01-02 Xinliang David Li <davidxl@google.com>
8258 PR tree-optimization/59303
8259 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8260 (dump_predicates): Better output format.
8261 (pred_equal_p): New function.
8262 (is_neq_relop_p): Ditto.
8263 (is_neq_zero_form_p): Ditto.
8264 (pred_expr_equal_p): Ditto.
8265 (pred_neg_p): Ditto.
8266 (simplify_pred): Ditto.
8267 (simplify_preds_2): Ditto.
8268 (simplify_preds_3): Ditto.
8269 (simplify_preds_4): Ditto.
8270 (simplify_preds): Ditto.
8272 (push_to_worklist): Ditto.
8273 (get_pred_info_from_cmp): Ditto.
8274 (is_degenerated_phi): Ditto.
8275 (normalize_one_pred_1): Ditto.
8276 (normalize_one_pred): Ditto.
8277 (normalize_one_pred_chain): Ditto.
8278 (normalize_preds): Ditto.
8279 (normalize_cond_1): Remove function.
8280 (normalize_cond): Ditto.
8281 (is_gcond_subset_of): Ditto.
8282 (is_subset_of_any): Ditto.
8283 (is_or_set_subset_of): Ditto.
8284 (is_and_set_subset_of): Ditto.
8285 (is_norm_cond_subset_of): Ditto.
8286 (pred_chain_length_cmp): Ditto.
8287 (convert_control_dep_chain_into_preds): Type change.
8288 (find_predicates): Ditto.
8289 (find_def_preds): Ditto.
8290 (destroy_predicates_vecs): Ditto.
8291 (find_matching_predicates_in_rest_chains): Ditto.
8292 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8293 (is_pred_expr_subset): Ditto.
8294 (is_pred_chain_subset_of): Ditto.
8295 (is_included_in): Ditto.
8296 (is_superset_of): Ditto.
8298 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8300 Update copyright years.
8302 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8304 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8305 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8306 config/arc/arc.md, config/arc/arc.opt,
8307 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8308 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8309 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8310 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8311 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8312 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8313 vtable-verify.c, vtable-verify.h: Use the standard form for the
8316 2014-01-02 Tobias Burnus <burnus@net-b.de>
8318 * gcc.c (process_command): Update copyright notice dates.
8319 * gcov-dump.c: Ditto.
8321 * doc/cpp.texi: Bump @copying's copyright year.
8322 * doc/cppinternals.texi: Ditto.
8323 * doc/gcc.texi: Ditto.
8324 * doc/gccint.texi: Ditto.
8325 * doc/gcov.texi: Ditto.
8326 * doc/install.texi: Ditto.
8327 * doc/invoke.texi: Ditto.
8329 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8331 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8333 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8335 * config/i386/sse.md (*mov<mode>_internal): Guard
8336 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8338 PR rtl-optimization/59647
8339 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8340 new_rtx into UNSIGNED_FLOAT rtxes.
8342 Copyright (C) 2014 Free Software Foundation, Inc.
8344 Copying and distribution of this file, with or without modification,
8345 are permitted in any medium without royalty provided the copyright
8346 notice and this notice are preserved.