1 2014-05-05 Richard Biener <rguenther@suse.de>
3 * passes.c (execute_function_todo): Move TODO_verify_flow under
4 the TODO_verify_ul umbrella.
6 2014-05-05 Richard Biener <rguenther@suse.de>
9 * fold-const.c (fold_binary_loc): Consistently avoid
10 canonicalizing X & CST away from a CST that is the mask
13 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
16 int argument to enum machine_mode.
17 (picochip_class_max_nregs): Ditto.
18 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
19 (picochip_class_max_nregs): Ditto.
21 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23 * target.def: Add new target hook.
24 * doc/tm.texi: Regenerate.
25 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
26 * targhooks.c (default_keep_leaf_when_profiled): New function.
28 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
29 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
31 2014-05-05 Bin Cheng <bin.cheng@arm.com>
33 PR tree-optimization/60363
34 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
35 (copy_phi_args): New parameters. Call get_value_locus_in_path.
36 (update_destination_phis): New parameter.
37 (create_edge_and_update_destination_phis): Ditto.
38 (ssa_fix_duplicate_block_edges): Pass new arguments.
39 (thread_single_edge): Ditto.
41 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
43 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
44 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
45 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
46 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
47 Use RS6000_BTM_HARD_FLOAT.
48 (BU_MISC_2): Likewise.
49 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
50 RS6000_BTM_HARD_FLOAT.
51 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
53 (rs6000_invalid_builtin): Add hard floating builtin support.
54 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
56 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
58 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
60 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
61 Add missing function* argument.
63 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
65 * lra-constraints.c (valid_address_p): Move earlier in file.
66 Add a constraint argument to the address_info version.
67 (satisfies_memory_constraint_p): New function.
68 (satisfies_address_constraint_p): Likewise.
69 (process_alt_operands, curr_insn_transform): Use them.
70 (process_address): Pass the constraint to valid_address_p when
71 checking address operands.
73 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
75 * config/mips/mips.c (mips_isa_rev): New variable.
76 (mips_set_architecture): Set it.
77 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
79 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
80 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
81 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
82 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
83 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
84 conditions in terms of mips_isa_rev.
85 (mips_isa_rev): Declare.
87 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
89 * config/sh/sh-mem.cc: Use tabs instead of spaces.
90 (prob_unlikely, prob_likely): Make variables const.
92 2014-05-03 Denis Chertykov <chertykov@gmail.com>
94 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
96 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
98 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
100 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
102 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
103 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
104 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
106 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
107 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
109 Replace usage of ROUND_REG with sh_round_reg.
110 Use CEIL instead of ROUND_ADVANCE.
112 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
115 * config/sh/sh.c: Include stdlib headers before everything else.
117 2014-05-02 Jakub Jelinek <jakub@redhat.com>
119 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
120 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
121 (gimplify_adjust_omp_clauses): Simd region is never
122 directly nested in combined parallel. Instead, for linear
123 with copyin/copyout, if in combined for simd loop, make decl
124 firstprivate/lastprivate on OMP_FOR.
125 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
126 expand_omp_for_static_chunk): When setting endvar, also set
127 fd->loop.v to the same value.
129 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
131 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
133 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
135 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
138 2014-05-02 Marek Polacek <polacek@redhat.com>
140 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
142 2014-02-26 Kito Cheng <kito@0xlab.org>
144 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
145 to a C expression marco.
146 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
147 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
148 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
149 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
150 HONOR_REG_ALLOC_ORDER.
151 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
153 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
155 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
157 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
159 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
161 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
163 * tree-if-conv.c (is_cond_scalar_reduction): New function.
164 (convert_scalar_cond_reduction): Likewise.
165 (predicate_scalar_phi): Add recognition and transformation
166 of simple conditioanl reduction to be vectorizable.
168 2014-05-01 Marek Polacek <polacek@redhat.com>
171 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
173 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
175 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
176 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
177 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
178 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
179 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
180 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
181 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
182 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
184 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
186 * config/arc/arc.opt (mlra): Move comment above option name
187 to avoid mis-parsing as language options.
189 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
191 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
192 * config/sol2.h: ... here.
193 * config/sol2-10.h: Remove.
195 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
196 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
197 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
198 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
199 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
200 * config/sol2.h: ... here.
201 (SECTION_NAME_FORMAT): Don't redefine.
202 (STARTFILE_ARCH32_SPEC): Rename to ...
203 (STARTFILE_ARCH_SPEC): ... this.
204 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
205 * config/sparc/sol2.h: ... here.
206 (SECTION_NAME_FORMAT): Don't undef.
207 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
208 (SUBTARGET_EXTRA_SPECS): Remove.
209 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
211 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
212 (MD_STARTFILE_PREFIX): Remove.
213 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
214 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
215 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
216 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
217 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
218 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
219 * config/i386/sol2.h: ... here.
220 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
221 * config/i386/sol2-bi.h: Remove.
222 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
223 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
225 * config/i386/t-sol2-64: Rename to ...
226 * config/i386/t-sol2: ... this.
227 * config/sparc/t-sol2-64: Rename to ...
228 * config/sparc/t-sol2: ... this.
230 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
231 sol2_tm_file_head, sol2_tm_file_tail.
232 Include ${cpu_type}/sol2.h before sol2.h.
234 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
235 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
236 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
237 Reflect i386/t-sol2-64 renaming.
238 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
239 Reflect sparc/t-sol2-64 renaming.
241 2014-04-30 Richard Biener <rguenther@suse.de>
243 * passes.c (execute_function_todo): Move TODO_verify_stmts
244 and TODO_verify_ssa under the TODO_verify_il umbrella.
245 * tree-ssa.h (verify_ssa): Adjust prototype.
246 * tree-ssa.c (verify_ssa): Add parameter to tell whether
247 we should verify SSA operands.
248 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
249 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
250 whether we should verify whether not throwing stmts have EH info.
251 * graphite-scop-detection.c (create_sese_edges): Adjust.
252 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
253 * tree-eh.c (lower_try_finally_switch): Do not add the
254 default case label twice.
256 2014-04-30 Marek Polacek <polacek@redhat.com>
258 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
259 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
260 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
261 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
263 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
265 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
266 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
267 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
268 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
269 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
270 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
271 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
272 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
274 2014-04-29 David Malcolm <dmalcolm@redhat.com>
276 * tree-cfg.c (dump_function_to_file): Dump the return type of
277 functions, in a line to itself before the function body, mimicking
278 the layout of a C function.
280 2014-04-29 Jakub Jelinek <jakub@redhat.com>
282 PR tree-optimization/60971
283 * tree-tailcall.c (process_assignment): Reject conversions which
286 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
288 * calls.c (initialize_argument_information): Always treat
289 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
290 (expand_call): Likewise.
291 (emit_library_call_calue_1): Likewise.
292 * expr.c (PUSH_ARGS_REVERSED): Do not define.
293 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
296 2014-04-29 Nick Clifton <nickc@redhat.com>
298 * config/msp430/msp430.md (umulsidi): Fix typo.
299 (mulhisi3): Enable even inside interrupt handlers.
300 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
301 bigger return address pushed in large mode.
303 2014-04-29 Nick Clifton <nickc@redhat.com>
305 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
306 (arc_init_reg_tables): Use a machine_mode enum to iterate over
308 * config/m32r/m32r.c (init_reg_tables): Likewise.
309 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
310 enum to hold the modes.
312 2014-04-29 Richard Biener <rguenther@suse.de>
314 * dominance.c (free_dominance_info): Add overload with
316 (dom_info_state): Likewise.
317 (dom_info_available_p): Likewise.
318 * basic-block.h (free_dominance_info, dom_info_state,
319 dom_info_available_p): Declare overloads.
320 * passes.c (execute_function_todo): Verify that verifiers
321 don't change dominator info state. Drop dominator info
322 for IPA pass invocations.
323 * cgraph.c (release_function_body): Restore asserts that
324 dominator information is released.
326 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
328 * doc/invoke.texi: Fix typo.
329 * tree-vrp.c: Fix typos.
330 * gimple.c (infer_nonnull_range): Reorder operands of an &&
333 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
335 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
337 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
339 * config/aarch64/aarch64-builtins.c
340 (aarch64_types_storestruct_lane_qualifiers): New.
341 (TYPES_STORESTRUCT_LANE): Likewise.
342 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
343 (st3_lane): Likewise.
344 (st4_lane): Likewise.
345 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
346 (vec_store_lanesci_lane<mode>): Likewise.
347 (vec_store_lanesxi_lane<mode>): Likewise.
348 (aarch64_st2_lane<VQ:mode>): Likewise.
349 (aarch64_st3_lane<VQ:mode>): Likewise.
350 (aarch64_st4_lane<VQ:mode>): Likewise.
351 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
352 * config/aarch64/arm_neon.h
353 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
354 use new macro arguments.
355 (__ST3_LANE_FUNC): Likewise.
356 (__ST4_LANE_FUNC): Likewise.
357 * config/aarch64/iterators.md (V_TWO_ELEM): New.
358 (V_THREE_ELEM): Likewise.
359 (V_FOUR_ELEM): Likewise.
361 2014-04-28 David Malcolm <dmalcolm@redhat.com>
363 * doc/gimple.texi: Replace the description of the now-defunct
364 union gimple_statement_d with a diagram showing the
365 gimple_statement_base class hierarchy and its relationships to
366 the GSS_ and GIMPLE_ enums.
368 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
370 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
371 * config/aarch64/aarch64.c
372 (aarch64_cannot_change_mode_class): Weaken conditions.
373 (aarch64_modes_tieable_p): New.
374 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
376 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
378 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
379 (loadsync_<mode>): Change mode.
380 (load_quadpti, store_quadpti): New.
381 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
382 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
384 2014-04-28 Martin Jambor <mjambor@suse.cz>
386 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
387 same alias type as the original statement.
388 (subreplacement_assignment_data): New type.
389 (handle_unscalarized_data_in_subtree): New type of parameter,
390 generate new memory accesses with same alias type as the original
392 (load_assign_lhs_subreplacements): Likewise.
393 (sra_modify_constructor_assign): Generate new memory accesses with
394 same alias type as the original statement.
396 2014-04-28 Richard Biener <rguenther@suse.de>
398 * tree-pass.h (TODO_verify_il): Define.
399 (TODO_verify_all): Complete properly.
400 * passes.c (execute_function_todo): Move existing loop-closed
401 SSA verification under TODO_verify_il.
402 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
403 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
404 Fix tree sharing issue.
406 2014-04-28 Richard Biener <rguenther@suse.de>
409 * builtins.def (DEF_C11_BUILTIN): Add.
410 (BUILT_IN_ALIGNED_ALLOC): Likewise.
411 * coretypes.h (enum function_class): Add function_c11_misc.
412 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
413 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
414 (call_may_clobber_ref_p_1): Likewise.
415 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
416 (mark_all_reaching_defs_necessary_1): Likewise.
417 (propagate_necessity): Likewise.
418 (eliminate_unnecessary_stmts): Likewise.
419 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
421 2014-04-28 Richard Biener <rguenther@suse.de>
423 * tree-vrp.c (vrp_var_may_overflow): Remove.
424 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
425 with overflow immediately bump to one before that value and
426 let iteration figure out overflow status.
428 2014-04-28 Richard Biener <rguenther@suse.de>
430 * configure.ac: Do valgrind header checks unconditionally.
431 Add --enable-valgrind-annotations.
432 * system.h: Guard valgrind header inclusion with
433 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
434 * alloc-pool.c (pool_alloc, pool_free): Use
435 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
436 to guard possibly dead code.
437 * config.in: Regenerated.
438 * configure: Likewise.
440 2014-04-28 Jeff Law <law@redhat.com>
442 PR tree-optimization/60902
443 * tree-ssa-threadedge.c
444 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
445 over real defs when invalidating outputs from statements that do not
446 produce useful outputs for threading.
448 2014-04-28 Richard Biener <rguenther@suse.de>
450 PR tree-optimization/60979
451 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
452 SCOPs that end in a block with a successor with abnormal
455 2014-04-28 Richard Biener <rguenther@suse.de>
457 * tree-pass.h (execute_pass_list): Adjust prototype.
458 * passes.c (pass_manager::execute_early_local_passes):
460 (do_per_function): Change callback signature, push all actual
461 work to the callbals.
462 (do_per_function_toporder): Likewise.
463 (execute_function_dump): Adjust.
464 (execute_function_todo): Likewise.
465 (clear_last_verified): Likewise.
466 (verify_curr_properties): Likewise.
467 (update_properties_after_pass): Likewise.
468 (execute_pass_list_1): Split out from ...
469 (execute_pass_list): ... here. Adjust.
470 (execute_ipa_pass_list): Likewise.
471 * cgraphunit.c (cgraph_add_new_function): Adjust.
472 (analyze_function): Likewise.
473 (expand_function): Likewise.
474 * cgraph.c (release_function_body): Free dominance info
475 here instead of asserting it was magically freed elsewhere.
477 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
479 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
480 * configure: Regenerate.
481 * config/sparc/sparc.opt (muser-mode): New option.
482 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
484 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
485 * doc/invoke.texi (SPARC options): Document -muser-mode.
487 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
489 * cselib.c (find_slot_memmode): Delete.
490 (cselib_hasher): Change compare_type to a struct.
491 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
493 (preserve_constants_and_equivs): Adjust for new compare_type.
494 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
495 (wrap_constant): Delete.
496 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
498 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
500 * doc/install.texi (Building with profile feedback): Remove
503 2014-04-26 Tom de Vries <tom@codesourcery.com>
505 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
508 2014-04-25 Cary Coutant <ccoutant@google.com>
511 * dwarf2out.c (should_move_die_to_comdat): A type definition
512 can contain a subprogram definition, but don't move it to a
514 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
515 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
517 (clone_tree_hash): Rename to...
518 (clone_tree_partial): ...this; change callers. Copy
519 DW_TAG_subprogram DIEs as declarations.
520 (copy_decls_walk): Don't copy children of a declaration into a
523 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
526 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
529 2014-04-25 Jiong Wang <jiong.wang@arm.com>
531 * config/arm/predicates.md (call_insn_operand): Add long_call check.
532 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
534 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
537 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
539 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
542 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
544 PR tree-optimization/60930
545 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
546 creating a multiply candidate by folding two constant
547 multiplicands when the result overflows.
549 2014-04-25 Jakub Jelinek <jakub@redhat.com>
551 PR tree-optimization/60960
552 * tree-vect-generic.c (expand_vector_operation): Only call
553 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
555 2014-04-25 Tom de Vries <tom@codesourcery.com>
557 * expr.c (clobber_reg_mode): New function.
558 * expr.h (clobber_reg): New function.
560 2014-04-25 Tom de Vries <tom@codesourcery.com>
562 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
565 2014-04-25 Radovan Obradovic <robradovic@mips.com>
566 Tom de Vries <tom@codesourcery.com>
568 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
570 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
571 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
572 new argument to find_all_hard_reg_sets call.
574 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
577 Use HOST_WIDE_INT_C for mask literal.
578 (aarch_rev16_shleft_mask_imm_p): Likewise.
580 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
583 * config/sparc/sparc.md (ashlsi3_extend): Delete.
585 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
587 PR preprocessor/56540
588 * config/i386/i386-c.c (ix86_target_macros): Define
589 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
591 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
593 * configure.ac (tga_func): Remove.
594 (LIB_TLS_SPEC): Remove.
595 * configure: Regenerate.
596 * config.in: Regenerate.
597 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
599 2014-04-25 Richard Biener <rguenther@suse.de>
602 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
603 call stmt use/clobber sets during stmt walk instead of
604 walking the possibly incomplete set of caller edges.
606 2014-04-25 Richard Biener <rguenther@suse.de>
609 * passes.c (apply_ipa_transforms): Inline into only caller ...
610 (execute_one_pass): ... here. Properly bring in function
611 bodies for nodes we want to apply IPA transforms to.
613 2014-04-24 Cong Hou <congh@google.com>
615 PR tree-optimization/60896
616 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
617 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
618 (vect_mark_pattern_stmts): Set the def type of all statements in
619 PATTERN_DEF_SEQ as vect_internal_def.
621 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
623 * doc/extend.texi (PowerPC Built-in Functions): Document new
624 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
626 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
628 * config/rs6000/predicates.md (const_0_to_3_operand): New
629 predicate to match 0..3 integer constants.
631 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
632 to support adding miscellaneous builtin functions.
633 (BU_DFP_MISC_2): Likewise.
634 (BU_P7_MISC_1): Likewise.
635 (BU_P7_MISC_2): Likewise.
636 (BU_P8V_MISC_3): Likewise.
637 (BU_MISC_1): Likewise.
638 (BU_MISC_2): Likewise.
639 (DIVWE): Add extended divide builtin functions.
647 (DXEX): Add decimal floating-point builtin functions.
659 (CDTBCD): Add new BCD builtin functions.
663 (BCDADD_LT): Likewise.
664 (BCDADD_EQ): Likewise.
665 (BCDADD_GT): Likewise.
666 (BCDADD_OV): Likewise.
668 (BCDSUB_LT): Likewise.
669 (BCDSUB_EQ): Likewise.
670 (BCDSUB_GT): Likewise.
671 (BCDSUB_OV): Likewise.
672 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
673 (UNPACK_TD): Likewise.
675 (UNPACK_TF): Likewise.
676 (UNPACK_TF_0): Likewise.
677 (UNPACK_TF_1): Likewise.
678 (PACK_V1TI): Likewise.
679 (UNPACK_V1TI): Likewise.
681 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
682 support for decimal floating point builtin functions.
683 (rs6000_expand_ternop_builtin): Add checks for the new builtin
684 functions that take constant arguments.
685 (rs6000_invalid_builtin): Add decimal floating point builtin
687 (rs6000_init_builtins): Setup long double, _Decimal64, and
688 _Decimal128 types for new builtin functions.
689 (builtin_function_type): Set the unsigned flags appropriately for
690 the new builtin functions.
691 (rs6000_opt_masks): Add support for decimal floating point builtin
694 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
695 floating point builtin functions.
696 (RS6000_BTM_COMMON): Likewise.
697 (RS6000_BTI_long_double): Likewise.
698 (RS6000_BTI_dfloat64): Likewise.
699 (RS6000_BTI_dfloat128): Likewise.
700 (long_double_type_internal_node): Likewise.
701 (dfloat64_type_internal_node): Likewise.
702 (dfloat128_type_internal_node): Likewise.
704 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
705 2.07 bcd arithmetic instructions.
706 (UNSPEC_BCDSUB): Likewise.
707 (UNSPEC_BCD_OVERFLOW): Likewise.
708 (UNSPEC_BCD_ADD_SUB): Likewise.
709 (bcd_add_sub): Likewise.
710 (BCD_TEST): Likewise.
711 (bcd<bcd_add_sub>): Likewise.
712 (bcd<bcd_add_sub>_test): Likewise.
713 (bcd<bcd_add_sub>_test2): Likewise.
714 (bcd<bcd_add_sub>_<code>): Likewise.
715 (peephole2 for combined bcd ops): Likewise.
717 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
718 decimal floating point builtin functions.
719 (UNSPEC_DENBCD): Likewise.
720 (UNSPEC_DXEX): Likewise.
721 (UNSPEC_DIEX): Likewise.
722 (UNSPEC_DSCLI): Likewise.
723 (UNSPEC_DSCRI): Likewise.
724 (D64_D128): Likewise.
725 (dfp_suffix): Likewise.
726 (dfp_ddedpd_<mode>): Likewise.
727 (dfp_denbcd_<mode>): Likewise.
728 (dfp_dxex_<mode>): Likewise.
729 (dfp_diex_<mode>): Likewise.
730 (dfp_dscli_<mode>): Likewise.
731 (dfp_dscri_<mode>): Likewise.
733 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
735 (UNSPEC_CDTBCD): Likewise.
736 (UNSPEC_CBCDTD): Likewise.
737 (UNSPEC_DIVE): Add support for new extended divide builtin
739 (UNSPEC_DIVEO): Likewise.
740 (UNSPEC_DIVEU): Likewise.
741 (UNSPEC_DIVEUO): Likewise.
742 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
743 pack/unpack 128-bit types.
744 (UNSPEC_PACK_128BIT): Likewise.
745 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
746 (udiv<mode>3): Use idiv_ldiv mode attribute.
747 (div<mode>3): Likewise.
748 (addg6s): Add new BCD builtin functions.
751 (UNSPEC_DIV_EXTEND): Add support for new extended divide
753 (div_extend): Likewise.
754 (div<div_extend>_<mode>"): Likewise.
755 (FP128_64): Add support for new builtin functions to pack/unpack
757 (unpack<mode>): Likewise.
758 (unpacktf_0): Likewise.
759 (unpacktf_1): Likewise.
760 (unpack<mode>_dm): Likewise.
761 (unpack<mode>_nodm): Likewise.
762 (pack<mode>): Likewise.
763 (unpackv1ti): Likewise.
764 (packv1ti): Likewise.
766 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
768 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
771 2014-04-24 Jakub Jelinek <jakub@redhat.com>
773 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
774 * gimplify.c (omp_is_private): Change last argument's type to int.
775 Only diagnose lastprivate if the simd argument is 1, only diagnose
776 linear if the simd argument is 2.
777 (gimplify_omp_for): Adjust omp_is_private callers. When adding
778 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
779 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
780 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
781 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
782 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
783 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
784 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
785 * tree-nested.c (convert_nonlocal_omp_clauses,
786 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
788 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
791 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
794 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
796 * flag-types.h (enum ivar_visibility): Add.
798 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
800 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
803 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
805 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
807 2014-04-24 Radovan Obradovic <robradovic@mips.com>
808 Tom de Vries <tom@codesourcery.com>
810 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
811 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
813 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
814 * emit-rtl.c (try_split): Same.
816 2014-04-24 Radovan Obradovic <robradovic@mips.com>
817 Tom de Vries <tom@codesourcery.com>
819 * common.opt (fuse-caller-save): New option.
821 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
823 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
826 2014-04-24 Richard Biener <rguenther@suse.de>
828 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
829 during TER and instead use the sepops interface for expanding
830 non-GIMPLE_SINGLE_RHS.
832 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
834 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
835 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
837 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
839 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
840 assembler 64-bit option.
841 * configure: Regenerate.
843 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
845 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
846 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
847 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
848 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
849 (TARGET_CRYPTO): Take TARGET_SIMD into account.
851 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
853 * config/aarch64/aarch64-builtins.c
854 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
855 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
856 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
857 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
859 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
860 (Vrevsuff): New mode attribute.
862 2014-04-24 Terry Guo <terry.guo@arm.com>
864 * config/arm/arm.h (machine_function): Define variable
865 after_arm_reorg here.
866 * config/arm/arm.c (after_arm_reorg): Remove the definition.
867 (arm_split_constant): Update the way to access variable
870 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
872 2014-04-23 Tom de Vries <tom@codesourcery.com>
874 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
876 2014-04-23 David Malcolm <dmalcolm@redhat.com>
878 * is-a.h: Update comments to reflect the following changes to the
879 "pointerness" of the API, making the template parameter match the
880 return type, allowing use of is-a.h with typedefs of pointers.
881 (is_a_helper::cast): Return a T rather then a pointer to a T, so
882 that the return type matches the parameter to the is_a_helper.
884 (dyn_cast): Likewise.
886 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
887 pointer from the is-a.h API.
889 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
890 (is_a_helper <cgraph_node *>::test): ...this, matching change to
892 (is_a_helper <varpool_node>::test): Likewise, convert to...
893 (is_a_helper <varpool_node *>::test): ...this.
895 (varpool_first_variable): Update for removal of implicit pointer
897 (varpool_next_variable): Likewise.
898 (varpool_first_static_initializer): Likewise.
899 (varpool_next_static_initializer): Likewise.
900 (varpool_first_defined_variable): Likewise.
901 (varpool_next_defined_variable): Likewise.
902 (cgraph_first_defined_function): Likewise.
903 (cgraph_next_defined_function): Likewise.
904 (cgraph_first_function): Likewise.
905 (cgraph_next_function): Likewise.
906 (cgraph_first_function_with_gimple_body): Likewise.
907 (cgraph_next_function_with_gimple_body): Likewise.
908 (cgraph_alias_target): Likewise.
909 (varpool_alias_target): Likewise.
910 (cgraph_function_or_thunk_node): Likewise.
911 (varpool_variable_node): Likewise.
912 (symtab_real_symbol_p): Likewise.
913 * cgraphunit.c (referred_to_p): Likewise.
914 (analyze_functions): Likewise.
915 (handle_alias_pairs): Likewise.
916 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
917 * gimple-ssa.h (gimple_vuse_op): Likewise.
918 (gimple_vdef_op): Likewise.
919 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
920 * gimple.c (gimple_build_asm_1): Likewise.
921 (gimple_build_try): Likewise.
922 (gimple_build_resx): Likewise.
923 (gimple_build_eh_dispatch): Likewise.
924 (gimple_build_omp_for): Likewise.
925 (gimple_omp_for_set_clauses): Likewise.
927 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
928 (is_a_helper <gimple_statement_asm *>::test): ...this.
929 (is_a_helper <gimple_statement_bind>::test): Convert to...
930 (is_a_helper <gimple_statement_bind *>::test): ...this.
931 (is_a_helper <gimple_statement_call>::test): Convert to...
932 (is_a_helper <gimple_statement_call *>::test): ...this.
933 (is_a_helper <gimple_statement_catch>::test): Convert to...
934 (is_a_helper <gimple_statement_catch *>::test): ...this.
935 (is_a_helper <gimple_statement_resx>::test): Convert to...
936 (is_a_helper <gimple_statement_resx *>::test): ...this.
937 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
938 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
939 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
940 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
941 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
942 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
943 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
944 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
945 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
946 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
947 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
948 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
949 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
950 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
951 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
952 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
953 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
954 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
955 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
956 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
957 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
958 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
959 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
960 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
961 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
962 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
963 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
964 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
965 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
966 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
967 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
968 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
969 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
970 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
971 (is_a_helper <gimple_statement_phi>::test): Convert to...
972 (is_a_helper <gimple_statement_phi *>::test): ...this.
973 (is_a_helper <gimple_statement_transaction>::test): Convert to...
974 (is_a_helper <gimple_statement_transaction *>::test): ...this.
975 (is_a_helper <gimple_statement_try>::test): Convert to...
976 (is_a_helper <gimple_statement_try *>::test): ...this.
977 (is_a_helper <gimple_statement_wce>::test): Convert to...
978 (is_a_helper <gimple_statement_wce *>::test): ...this.
979 (is_a_helper <const gimple_statement_asm>::test): Convert to...
980 (is_a_helper <const gimple_statement_asm *>::test): ...this.
981 (is_a_helper <const gimple_statement_bind>::test): Convert to...
982 (is_a_helper <const gimple_statement_bind *>::test): ...this.
983 (is_a_helper <const gimple_statement_call>::test): Convert to...
984 (is_a_helper <const gimple_statement_call *>::test): ...this.
985 (is_a_helper <const gimple_statement_catch>::test): Convert to...
986 (is_a_helper <const gimple_statement_catch *>::test): ...this.
987 (is_a_helper <const gimple_statement_resx>::test): Convert to...
988 (is_a_helper <const gimple_statement_resx *>::test): ...this.
989 (is_a_helper <const gimple_statement_eh_dispatch>::test):
991 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
993 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
995 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
996 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
998 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1000 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1002 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1004 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1006 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1007 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1009 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1010 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1012 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1013 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1014 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1015 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1017 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1018 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1020 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1021 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1023 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1024 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1026 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1027 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1029 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1030 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1032 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1033 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1035 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1036 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1037 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1038 (is_a_helper <const gimple_statement_transaction>::test): Convert
1040 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1041 (is_a_helper <const gimple_statement_with_ops>::test): Convert
1043 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
1044 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
1045 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
1046 (is_a_helper <const gimple_statement_with_memory_ops>::test):
1048 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
1050 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
1052 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
1054 (gimple_use_ops): Update for removal of implicit pointer from the
1056 (gimple_set_use_ops): Likewise.
1057 (gimple_vuse): Likewise.
1058 (gimple_vdef): Likewise.
1059 (gimple_vuse_ptr): Likewise.
1060 (gimple_vdef_ptr): Likewise.
1061 (gimple_set_vuse): Likewise.
1062 (gimple_set_vdef): Likewise.
1063 (gimple_omp_return_set_lhs): Likewise.
1064 (gimple_omp_return_lhs): Likewise.
1065 (gimple_omp_return_lhs_ptr): Likewise.
1066 (gimple_call_fntype): Likewise.
1067 (gimple_call_set_fntype): Likewise.
1068 (gimple_call_set_internal_fn): Likewise.
1069 (gimple_call_use_set): Likewise.
1070 (gimple_call_clobber_set): Likewise.
1071 (gimple_bind_vars): Likewise.
1072 (gimple_bind_set_vars): Likewise.
1073 (gimple_bind_body_ptr): Likewise.
1074 (gimple_bind_set_body): Likewise.
1075 (gimple_bind_add_stmt): Likewise.
1076 (gimple_bind_block): Likewise.
1077 (gimple_bind_set_block): Likewise.
1078 (gimple_asm_ninputs): Likewise.
1079 (gimple_asm_noutputs): Likewise.
1080 (gimple_asm_nclobbers): Likewise.
1081 (gimple_asm_nlabels): Likewise.
1082 (gimple_asm_input_op): Likewise.
1083 (gimple_asm_input_op_ptr): Likewise.
1084 (gimple_asm_output_op): Likewise.
1085 (gimple_asm_output_op_ptr): Likewise.
1086 (gimple_asm_set_output_op): Likewise.
1087 (gimple_asm_clobber_op): Likewise.
1088 (gimple_asm_set_clobber_op): Likewise.
1089 (gimple_asm_label_op): Likewise.
1090 (gimple_asm_set_label_op): Likewise.
1091 (gimple_asm_string): Likewise.
1092 (gimple_catch_types): Likewise.
1093 (gimple_catch_types_ptr): Likewise.
1094 (gimple_catch_handler_ptr): Likewise.
1095 (gimple_catch_set_types): Likewise.
1096 (gimple_catch_set_handler): Likewise.
1097 (gimple_eh_filter_types): Likewise.
1098 (gimple_eh_filter_types_ptr): Likewise.
1099 (gimple_eh_filter_failure_ptr): Likewise.
1100 (gimple_eh_filter_set_types): Likewise.
1101 (gimple_eh_filter_set_failure): Likewise.
1102 (gimple_eh_must_not_throw_fndecl): Likewise.
1103 (gimple_eh_must_not_throw_set_fndecl): Likewise.
1104 (gimple_eh_else_n_body_ptr): Likewise.
1105 (gimple_eh_else_e_body_ptr): Likewise.
1106 (gimple_eh_else_set_n_body): Likewise.
1107 (gimple_eh_else_set_e_body): Likewise.
1108 (gimple_try_eval_ptr): Likewise.
1109 (gimple_try_cleanup_ptr): Likewise.
1110 (gimple_try_set_eval): Likewise.
1111 (gimple_try_set_cleanup): Likewise.
1112 (gimple_wce_cleanup_ptr): Likewise.
1113 (gimple_wce_set_cleanup): Likewise.
1114 (gimple_phi_capacity): Likewise.
1115 (gimple_phi_num_args): Likewise.
1116 (gimple_phi_result): Likewise.
1117 (gimple_phi_result_ptr): Likewise.
1118 (gimple_phi_set_result): Likewise.
1119 (gimple_phi_arg): Likewise.
1120 (gimple_phi_set_arg): Likewise.
1121 (gimple_resx_region): Likewise.
1122 (gimple_resx_set_region): Likewise.
1123 (gimple_eh_dispatch_region): Likewise.
1124 (gimple_eh_dispatch_set_region): Likewise.
1125 (gimple_omp_critical_name): Likewise.
1126 (gimple_omp_critical_name_ptr): Likewise.
1127 (gimple_omp_critical_set_name): Likewise.
1128 (gimple_omp_for_clauses): Likewise.
1129 (gimple_omp_for_clauses_ptr): Likewise.
1130 (gimple_omp_for_set_clauses): Likewise.
1131 (gimple_omp_for_collapse): Likewise.
1132 (gimple_omp_for_index): Likewise.
1133 (gimple_omp_for_index_ptr): Likewise.
1134 (gimple_omp_for_set_index): Likewise.
1135 (gimple_omp_for_initial): Likewise.
1136 (gimple_omp_for_initial_ptr): Likewise.
1137 (gimple_omp_for_set_initial): Likewise.
1138 (gimple_omp_for_final): Likewise.
1139 (gimple_omp_for_final_ptr): Likewise.
1140 (gimple_omp_for_set_final): Likewise.
1141 (gimple_omp_for_incr): Likewise.
1142 (gimple_omp_for_incr_ptr): Likewise.
1143 (gimple_omp_for_set_incr): Likewise.
1144 (gimple_omp_for_pre_body_ptr): Likewise.
1145 (gimple_omp_for_set_pre_body): Likewise.
1146 (gimple_omp_parallel_clauses): Likewise.
1147 (gimple_omp_parallel_clauses_ptr): Likewise.
1148 (gimple_omp_parallel_set_clauses): Likewise.
1149 (gimple_omp_parallel_child_fn): Likewise.
1150 (gimple_omp_parallel_child_fn_ptr): Likewise.
1151 (gimple_omp_parallel_set_child_fn): Likewise.
1152 (gimple_omp_parallel_data_arg): Likewise.
1153 (gimple_omp_parallel_data_arg_ptr): Likewise.
1154 (gimple_omp_parallel_set_data_arg): Likewise.
1155 (gimple_omp_task_clauses): Likewise.
1156 (gimple_omp_task_clauses_ptr): Likewise.
1157 (gimple_omp_task_set_clauses): Likewise.
1158 (gimple_omp_task_child_fn): Likewise.
1159 (gimple_omp_task_child_fn_ptr): Likewise.
1160 (gimple_omp_task_set_child_fn): Likewise.
1161 (gimple_omp_task_data_arg): Likewise.
1162 (gimple_omp_task_data_arg_ptr): Likewise.
1163 (gimple_omp_task_set_data_arg): Likewise.
1164 (gimple_omp_taskreg_clauses): Likewise.
1165 (gimple_omp_taskreg_clauses_ptr): Likewise.
1166 (gimple_omp_taskreg_set_clauses): Likewise.
1167 (gimple_omp_taskreg_child_fn): Likewise.
1168 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1169 (gimple_omp_taskreg_set_child_fn): Likewise.
1170 (gimple_omp_taskreg_data_arg): Likewise.
1171 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1172 (gimple_omp_taskreg_set_data_arg): Likewise.
1173 (gimple_omp_task_copy_fn): Likewise.
1174 (gimple_omp_task_copy_fn_ptr): Likewise.
1175 (gimple_omp_task_set_copy_fn): Likewise.
1176 (gimple_omp_task_arg_size): Likewise.
1177 (gimple_omp_task_arg_size_ptr): Likewise.
1178 (gimple_omp_task_set_arg_size): Likewise.
1179 (gimple_omp_task_arg_align): Likewise.
1180 (gimple_omp_task_arg_align_ptr): Likewise.
1181 (gimple_omp_task_set_arg_align): Likewise.
1182 (gimple_omp_single_clauses): Likewise.
1183 (gimple_omp_single_clauses_ptr): Likewise.
1184 (gimple_omp_single_set_clauses): Likewise.
1185 (gimple_omp_target_clauses): Likewise.
1186 (gimple_omp_target_clauses_ptr): Likewise.
1187 (gimple_omp_target_set_clauses): Likewise.
1188 (gimple_omp_target_child_fn): Likewise.
1189 (gimple_omp_target_child_fn_ptr): Likewise.
1190 (gimple_omp_target_set_child_fn): Likewise.
1191 (gimple_omp_target_data_arg): Likewise.
1192 (gimple_omp_target_data_arg_ptr): Likewise.
1193 (gimple_omp_target_set_data_arg): Likewise.
1194 (gimple_omp_teams_clauses): Likewise.
1195 (gimple_omp_teams_clauses_ptr): Likewise.
1196 (gimple_omp_teams_set_clauses): Likewise.
1197 (gimple_omp_sections_clauses): Likewise.
1198 (gimple_omp_sections_clauses_ptr): Likewise.
1199 (gimple_omp_sections_set_clauses): Likewise.
1200 (gimple_omp_sections_control): Likewise.
1201 (gimple_omp_sections_control_ptr): Likewise.
1202 (gimple_omp_sections_set_control): Likewise.
1203 (gimple_omp_for_set_cond): Likewise.
1204 (gimple_omp_for_cond): Likewise.
1205 (gimple_omp_atomic_store_set_val): Likewise.
1206 (gimple_omp_atomic_store_val): Likewise.
1207 (gimple_omp_atomic_store_val_ptr): Likewise.
1208 (gimple_omp_atomic_load_set_lhs): Likewise.
1209 (gimple_omp_atomic_load_lhs): Likewise.
1210 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1211 (gimple_omp_atomic_load_set_rhs): Likewise.
1212 (gimple_omp_atomic_load_rhs): Likewise.
1213 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1214 (gimple_omp_continue_control_def): Likewise.
1215 (gimple_omp_continue_control_def_ptr): Likewise.
1216 (gimple_omp_continue_set_control_def): Likewise.
1217 (gimple_omp_continue_control_use): Likewise.
1218 (gimple_omp_continue_control_use_ptr): Likewise.
1219 (gimple_omp_continue_set_control_use): Likewise.
1220 (gimple_transaction_body_ptr): Likewise.
1221 (gimple_transaction_label): Likewise.
1222 (gimple_transaction_label_ptr): Likewise.
1223 (gimple_transaction_set_body): Likewise.
1224 (gimple_transaction_set_label): Likewise.
1226 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1227 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1228 * ipa-ref.c (ipa_record_reference): Likewise.
1229 * ipa-reference.c (analyze_function): Likewise.
1230 (ipa_reference_write_optimization_summary): Likewise.
1231 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1232 (address_taken_from_non_vtable_p): Likewise.
1233 (comdat_can_be_unshared_p_1): Likewise.
1234 * lto-cgraph.c (lto_output_ref): Likewise.
1235 (add_references): Likewise.
1236 (compute_ltrans_boundary): Likewise.
1237 (output_symtab): Likewise.
1238 (input_ref): Likewise.
1239 (input_cgraph_1): Likewise.
1240 (output_cgraph_opt_summary): Likewise.
1241 * lto-streamer-out.c (lto_output): Likewise.
1242 (output_symbol_p): Likewise.
1243 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1244 (lsei_start_function_in_partition): Likewise.
1245 (lsei_next_variable_in_partition): Likewise.
1246 (lsei_start_variable_in_partition): Likewise.
1247 * symtab.c (insert_to_assembler_name_hash): Likewise.
1248 (unlink_from_assembler_name_hash): Likewise.
1249 (symtab_unregister_node): Likewise.
1250 (symtab_remove_node): Likewise.
1251 (dump_symtab_node): Likewise.
1252 (verify_symtab_base): Likewise.
1253 (verify_symtab_node): Likewise.
1254 (symtab_make_decl_local): Likewise.
1255 (symtab_alias_ultimate_target): Likewise.
1256 (symtab_resolve_alias): Likewise.
1257 (symtab_get_symbol_partitioning_class): Likewise.
1258 * tree-phinodes.c (allocate_phi_node): Likewise.
1259 (reserve_phi_args_for_new_edge): Likewise.
1260 (remove_phi_args): Likewise.
1261 * varpool.c (varpool_node_for_asm): Likewise.
1262 (varpool_remove_unreferenced_decls): Likewise.
1264 2014-04-23 Jeff Law <law@redhat.com>
1266 PR tree-optimization/60902
1267 * tree-ssa-threadedge.c
1268 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1269 invalidate outputs from statements that do not produce useful
1270 outputs for threading.
1272 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1274 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1275 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1276 machine descriptions for Stack Smashing Protector.
1278 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1280 * aarch64.md (<optab>_rol<mode>3): New pattern.
1281 (<optab>_rolsi3_uxtw): Likewise.
1282 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1284 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1286 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1287 (arm_cortex_a12_tune): Likewise.
1289 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1291 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1293 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1295 * config/arm/arm.md (arm_rev16si2): New pattern.
1296 (arm_rev16si2_alt): Likewise.
1297 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1299 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1301 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1302 (rev16<mode>2_alt): Likewise.
1303 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1304 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1305 (aarch_rev16_shleft_mask_imm_p): Likewise.
1306 (aarch_rev16_p_1): Likewise.
1307 (aarch_rev16_p): Likewise.
1308 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1309 (aarch_rev16_shright_mask_imm_p): Likewise.
1310 (aarch_rev16_shleft_mask_imm_p): Likewise.
1312 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1314 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1315 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1317 (cortex_a53_extra_costs): Likewise.
1318 (cortex_a57_extra_costs): Likewise.
1319 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1320 (cortexa7_extra_costs): Likewise.
1321 (cortexa8_extra_costs): Likewise.
1322 (cortexa12_extra_costs): Likewise.
1323 (cortexa15_extra_costs): Likewise.
1324 (v7m_extra_costs): Likewise.
1325 (arm_new_rtx_costs): Handle BSWAP.
1327 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1329 * config/arm/arm.c (cortexa8_extra_costs): New table.
1330 (arm_cortex_a8_tune): New tuning struct.
1331 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1333 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1335 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1337 2014-04-23 Richard Biener <rguenther@suse.de>
1339 * Makefile.in (OBJS): Remove loop-unswitch.o.
1340 * loop-unswitch.c: Delete.
1341 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1342 * passes.def (pass_rtl_unswitch): Likewise.
1343 * loop-init.c (gate_rtl_unswitch): Likewise.
1344 (rtl_unswitch): Likewise.
1345 (pass_data_rtl_unswitch): Likewise.
1346 (pass_rtl_unswitch): Likewise.
1347 (make_pass_rtl_unswitch): Likewise.
1348 * rtl.h (reversed_condition): Likewise.
1349 (compare_and_jump_seq): Likewise.
1350 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1352 * loop-unroll.c (compare_and_jump_seq): Likewise.
1354 2014-04-23 Richard Biener <rguenther@suse.de>
1356 PR tree-optimization/60903
1357 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1358 commented code block.
1359 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1360 loop flags to newly created BBs and edges.
1362 2014-04-23 Nick Clifton <nickc@redhat.com>
1364 * config/msp430/msp430.c (msp430_handle_option): Move function
1366 (msp430_option_override): Simplify mcu and mcpu option handling.
1367 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1368 support for -mhwmult command line option.
1369 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1370 -mhwmult command line option.
1371 (msp430_hwmult_enabled): Delete.
1372 (msp43o_output_labelref): Add support for -mhwmult command line
1374 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1375 (umulsidi3): Likewise.
1376 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1377 (mcpu, mlarge, msmall): Likewise.
1378 (mhwmult): New option.
1379 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1381 (msp430_is_f5_mcu): Remove prototype.
1382 (msp430_use_f5_series_hwmult): Add prototype.
1383 * config/msp430/msp430-opts.h: New file.
1384 * common/config/msp430: New directory.
1385 * common/config/msp430/msp430-common.c: New file.
1386 * config.gcc (msp430): Remove target_has_targetm_common.
1387 * doc/invoke.texi: Document -mhwmult command line option.
1389 2014-04-23 Nick Clifton <nickc@redhat.com>
1391 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1392 default-manifest.o if it can be found in the search path.
1393 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1395 2014-04-23 Terry Guo <terry.guo@arm.com>
1397 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1399 2014-04-23 Richard Biener <rguenther@suse.de>
1402 * tree-inline.c (declare_return_variable): Use mark_addressable.
1404 2014-04-23 Richard Biener <rguenther@suse.de>
1407 * loop-init.c (loop_optimizer_init): Make sure to apply
1408 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1410 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1413 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1415 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1416 if flag_sanitize_undefined_trap_on_error.
1417 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1418 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1419 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1420 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1421 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1422 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1423 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1424 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1425 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1426 * ubsan.c (ubsan_instrument_unreachable): Return
1427 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1428 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1429 if flag_sanitize_undefined_trap_on_error and
1430 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1431 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1432 instrument_bool_enum_load): Emit __builtin_trap () if
1433 flag_sanitize_undefined_trap_on_error and
1434 __builtin_handle_*_abort () if !flag_sanitize_recover.
1435 * doc/invoke.texi (-fsanitize-recover,
1436 -fsanitize-undefined-trap-on-error): Document.
1438 2014-04-22 Christian Bruel <christian.bruel@st.com>
1440 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1441 Force immediates to SImode.
1443 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1445 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1447 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1448 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1449 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1450 (nios2_fpu_insn): Add entry for round.
1451 (N2FPU_NO_ERRNO_P): Define.
1452 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1454 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1456 2014-04-22 Richard Henderson <rth@redhat.com>
1458 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1459 (add<GPI>3_compare0): Remove leading * from name.
1460 (add<GPI>3_carryin): Likewise.
1461 (sub<GPI>3_compare0): Likewise.
1462 (sub<GPI>3_carryin): Likewise.
1463 (<su_optab>mulditi3): New expander.
1464 (multi3): New expander.
1465 (madd<GPI>): Remove leading * from name.
1467 2014-04-22 Martin Jambor <mjambor@suse.cz>
1469 * cgraphclones.c (cgraph_function_versioning): Copy
1470 ipa_transforms_to_apply instead of asserting it is empty.
1472 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1475 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1476 on count_exp to get mode.
1478 2014-04-22 Andrew Pinski <apinski@cavium.com>
1480 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1481 Handle TLS for ILP32.
1482 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1483 (tlsie_small_<mode>): this and handle PTR.
1484 (tlsie_small_sidi): New pattern.
1485 (tlsle_small): Change to an expand to handle ILP32.
1486 (tlsle_small_<mode>): New pattern.
1487 (tlsdesc_small): Rename to ...
1488 (tlsdesc_small_<mode>): this and handle PTR.
1490 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1492 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1494 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1496 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1497 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1498 (aarch64_types_signed_poly_qualifiers): Likewise.
1499 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1500 (aarch64_types_poly_signed_qualifiers): Likewise.
1501 (TYPES_REINTERP_SS): Type macro added.
1502 (TYPES_REINTERP_SU): Likewise.
1503 (TYPES_REINTERP_SP): Likewise.
1504 (TYPES_REINTERP_US): Likewise.
1505 (TYPES_REINTERP_PS): Likewise.
1506 (aarch64_fold_builtin): New expression folding added.
1507 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1508 Declarations removed.
1509 (REINTERP_SS): Declarations added.
1510 (REINTERP_US): Likewise.
1511 (REINTERP_PS): Likewise.
1512 (REINTERP_SU): Likewise.
1513 (REINTERP_SP): Likewise.
1514 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1515 (vreinterpretq_p8_f64): Likewise.
1516 (vreinterpret_p16_f64): Likewise.
1517 (vreinterpretq_p16_f64): Likewise.
1518 (vreinterpret_f32_f64): Likewise.
1519 (vreinterpretq_f32_f64): Likewise.
1520 (vreinterpret_f64_f32): Likewise.
1521 (vreinterpret_f64_p8): Likewise.
1522 (vreinterpret_f64_p16): Likewise.
1523 (vreinterpret_f64_s8): Likewise.
1524 (vreinterpret_f64_s16): Likewise.
1525 (vreinterpret_f64_s32): Likewise.
1526 (vreinterpret_f64_s64): Likewise.
1527 (vreinterpret_f64_u8): Likewise.
1528 (vreinterpret_f64_u16): Likewise.
1529 (vreinterpret_f64_u32): Likewise.
1530 (vreinterpret_f64_u64): Likewise.
1531 (vreinterpretq_f64_f32): Likewise.
1532 (vreinterpretq_f64_p8): Likewise.
1533 (vreinterpretq_f64_p16): Likewise.
1534 (vreinterpretq_f64_s8): Likewise.
1535 (vreinterpretq_f64_s16): Likewise.
1536 (vreinterpretq_f64_s32): Likewise.
1537 (vreinterpretq_f64_s64): Likewise.
1538 (vreinterpretq_f64_u8): Likewise.
1539 (vreinterpretq_f64_u16): Likewise.
1540 (vreinterpretq_f64_u32): Likewise.
1541 (vreinterpretq_f64_u64): Likewise.
1542 (vreinterpret_s64_f64): Likewise.
1543 (vreinterpretq_s64_f64): Likewise.
1544 (vreinterpret_u64_f64): Likewise.
1545 (vreinterpretq_u64_f64): Likewise.
1546 (vreinterpret_s8_f64): Likewise.
1547 (vreinterpretq_s8_f64): Likewise.
1548 (vreinterpret_s16_f64): Likewise.
1549 (vreinterpretq_s16_f64): Likewise.
1550 (vreinterpret_s32_f64): Likewise.
1551 (vreinterpretq_s32_f64): Likewise.
1552 (vreinterpret_u8_f64): Likewise.
1553 (vreinterpretq_u8_f64): Likewise.
1554 (vreinterpret_u16_f64): Likewise.
1555 (vreinterpretq_u16_f64): Likewise.
1556 (vreinterpret_u32_f64): Likewise.
1557 (vreinterpretq_u32_f64): Likewise.
1559 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1561 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1562 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1563 (vreinterpret_p8_s8): Likewise.
1564 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1565 (vreinterpret_p8_s16): Likewise.
1566 (vreinterpret_p8_s32): Likewise.
1567 (vreinterpret_p8_s64): Likewise.
1568 (vreinterpret_p8_f32): Likewise.
1569 (vreinterpret_p8_u8): Likewise.
1570 (vreinterpret_p8_u16): Likewise.
1571 (vreinterpret_p8_u32): Likewise.
1572 (vreinterpret_p8_u64): Likewise.
1573 (vreinterpret_p8_p16): Likewise.
1574 (vreinterpretq_p8_s8): Likewise.
1575 (vreinterpretq_p8_s16): Likewise.
1576 (vreinterpretq_p8_s32): Likewise.
1577 (vreinterpretq_p8_s64): Likewise.
1578 (vreinterpretq_p8_f32): Likewise.
1579 (vreinterpretq_p8_u8): Likewise.
1580 (vreinterpretq_p8_u16): Likewise.
1581 (vreinterpretq_p8_u32): Likewise.
1582 (vreinterpretq_p8_u64): Likewise.
1583 (vreinterpretq_p8_p16): Likewise.
1584 (vreinterpret_p16_s8): Likewise.
1585 (vreinterpret_p16_s16): Likewise.
1586 (vreinterpret_p16_s32): Likewise.
1587 (vreinterpret_p16_s64): Likewise.
1588 (vreinterpret_p16_f32): Likewise.
1589 (vreinterpret_p16_u8): Likewise.
1590 (vreinterpret_p16_u16): Likewise.
1591 (vreinterpret_p16_u32): Likewise.
1592 (vreinterpret_p16_u64): Likewise.
1593 (vreinterpret_p16_p8): Likewise.
1594 (vreinterpretq_p16_s8): Likewise.
1595 (vreinterpretq_p16_s16): Likewise.
1596 (vreinterpretq_p16_s32): Likewise.
1597 (vreinterpretq_p16_s64): Likewise.
1598 (vreinterpretq_p16_f32): Likewise.
1599 (vreinterpretq_p16_u8): Likewise.
1600 (vreinterpretq_p16_u16): Likewise.
1601 (vreinterpretq_p16_u32): Likewise.
1602 (vreinterpretq_p16_u64): Likewise.
1603 (vreinterpretq_p16_p8): Likewise.
1604 (vreinterpret_f32_s8): Likewise.
1605 (vreinterpret_f32_s16): Likewise.
1606 (vreinterpret_f32_s32): Likewise.
1607 (vreinterpret_f32_s64): Likewise.
1608 (vreinterpret_f32_u8): Likewise.
1609 (vreinterpret_f32_u16): Likewise.
1610 (vreinterpret_f32_u32): Likewise.
1611 (vreinterpret_f32_u64): Likewise.
1612 (vreinterpret_f32_p8): Likewise.
1613 (vreinterpret_f32_p16): Likewise.
1614 (vreinterpretq_f32_s8): Likewise.
1615 (vreinterpretq_f32_s16): Likewise.
1616 (vreinterpretq_f32_s32): Likewise.
1617 (vreinterpretq_f32_s64): Likewise.
1618 (vreinterpretq_f32_u8): Likewise.
1619 (vreinterpretq_f32_u16): Likewise.
1620 (vreinterpretq_f32_u32): Likewise.
1621 (vreinterpretq_f32_u64): Likewise.
1622 (vreinterpretq_f32_p8): Likewise.
1623 (vreinterpretq_f32_p16): Likewise.
1624 (vreinterpret_s64_s8): Likewise.
1625 (vreinterpret_s64_s16): Likewise.
1626 (vreinterpret_s64_s32): Likewise.
1627 (vreinterpret_s64_f32): Likewise.
1628 (vreinterpret_s64_u8): Likewise.
1629 (vreinterpret_s64_u16): Likewise.
1630 (vreinterpret_s64_u32): Likewise.
1631 (vreinterpret_s64_u64): Likewise.
1632 (vreinterpret_s64_p8): Likewise.
1633 (vreinterpret_s64_p16): Likewise.
1634 (vreinterpretq_s64_s8): Likewise.
1635 (vreinterpretq_s64_s16): Likewise.
1636 (vreinterpretq_s64_s32): Likewise.
1637 (vreinterpretq_s64_f32): Likewise.
1638 (vreinterpretq_s64_u8): Likewise.
1639 (vreinterpretq_s64_u16): Likewise.
1640 (vreinterpretq_s64_u32): Likewise.
1641 (vreinterpretq_s64_u64): Likewise.
1642 (vreinterpretq_s64_p8): Likewise.
1643 (vreinterpretq_s64_p16): Likewise.
1644 (vreinterpret_u64_s8): Likewise.
1645 (vreinterpret_u64_s16): Likewise.
1646 (vreinterpret_u64_s32): Likewise.
1647 (vreinterpret_u64_s64): Likewise.
1648 (vreinterpret_u64_f32): Likewise.
1649 (vreinterpret_u64_u8): Likewise.
1650 (vreinterpret_u64_u16): Likewise.
1651 (vreinterpret_u64_u32): Likewise.
1652 (vreinterpret_u64_p8): Likewise.
1653 (vreinterpret_u64_p16): Likewise.
1654 (vreinterpretq_u64_s8): Likewise.
1655 (vreinterpretq_u64_s16): Likewise.
1656 (vreinterpretq_u64_s32): Likewise.
1657 (vreinterpretq_u64_s64): Likewise.
1658 (vreinterpretq_u64_f32): Likewise.
1659 (vreinterpretq_u64_u8): Likewise.
1660 (vreinterpretq_u64_u16): Likewise.
1661 (vreinterpretq_u64_u32): Likewise.
1662 (vreinterpretq_u64_p8): Likewise.
1663 (vreinterpretq_u64_p16): Likewise.
1664 (vreinterpret_s8_s16): Likewise.
1665 (vreinterpret_s8_s32): Likewise.
1666 (vreinterpret_s8_s64): Likewise.
1667 (vreinterpret_s8_f32): Likewise.
1668 (vreinterpret_s8_u8): Likewise.
1669 (vreinterpret_s8_u16): Likewise.
1670 (vreinterpret_s8_u32): Likewise.
1671 (vreinterpret_s8_u64): Likewise.
1672 (vreinterpret_s8_p8): Likewise.
1673 (vreinterpret_s8_p16): Likewise.
1674 (vreinterpretq_s8_s16): Likewise.
1675 (vreinterpretq_s8_s32): Likewise.
1676 (vreinterpretq_s8_s64): Likewise.
1677 (vreinterpretq_s8_f32): Likewise.
1678 (vreinterpretq_s8_u8): Likewise.
1679 (vreinterpretq_s8_u16): Likewise.
1680 (vreinterpretq_s8_u32): Likewise.
1681 (vreinterpretq_s8_u64): Likewise.
1682 (vreinterpretq_s8_p8): Likewise.
1683 (vreinterpretq_s8_p16): Likewise.
1684 (vreinterpret_s16_s8): Likewise.
1685 (vreinterpret_s16_s32): Likewise.
1686 (vreinterpret_s16_s64): Likewise.
1687 (vreinterpret_s16_f32): Likewise.
1688 (vreinterpret_s16_u8): Likewise.
1689 (vreinterpret_s16_u16): Likewise.
1690 (vreinterpret_s16_u32): Likewise.
1691 (vreinterpret_s16_u64): Likewise.
1692 (vreinterpret_s16_p8): Likewise.
1693 (vreinterpret_s16_p16): Likewise.
1694 (vreinterpretq_s16_s8): Likewise.
1695 (vreinterpretq_s16_s32): Likewise.
1696 (vreinterpretq_s16_s64): Likewise.
1697 (vreinterpretq_s16_f32): Likewise.
1698 (vreinterpretq_s16_u8): Likewise.
1699 (vreinterpretq_s16_u16): Likewise.
1700 (vreinterpretq_s16_u32): Likewise.
1701 (vreinterpretq_s16_u64): Likewise.
1702 (vreinterpretq_s16_p8): Likewise.
1703 (vreinterpretq_s16_p16): Likewise.
1704 (vreinterpret_s32_s8): Likewise.
1705 (vreinterpret_s32_s16): Likewise.
1706 (vreinterpret_s32_s64): Likewise.
1707 (vreinterpret_s32_f32): Likewise.
1708 (vreinterpret_s32_u8): Likewise.
1709 (vreinterpret_s32_u16): Likewise.
1710 (vreinterpret_s32_u32): Likewise.
1711 (vreinterpret_s32_u64): Likewise.
1712 (vreinterpret_s32_p8): Likewise.
1713 (vreinterpret_s32_p16): Likewise.
1714 (vreinterpretq_s32_s8): Likewise.
1715 (vreinterpretq_s32_s16): Likewise.
1716 (vreinterpretq_s32_s64): Likewise.
1717 (vreinterpretq_s32_f32): Likewise.
1718 (vreinterpretq_s32_u8): Likewise.
1719 (vreinterpretq_s32_u16): Likewise.
1720 (vreinterpretq_s32_u32): Likewise.
1721 (vreinterpretq_s32_u64): Likewise.
1722 (vreinterpretq_s32_p8): Likewise.
1723 (vreinterpretq_s32_p16): Likewise.
1724 (vreinterpret_u8_s8): Likewise.
1725 (vreinterpret_u8_s16): Likewise.
1726 (vreinterpret_u8_s32): Likewise.
1727 (vreinterpret_u8_s64): Likewise.
1728 (vreinterpret_u8_f32): Likewise.
1729 (vreinterpret_u8_u16): Likewise.
1730 (vreinterpret_u8_u32): Likewise.
1731 (vreinterpret_u8_u64): Likewise.
1732 (vreinterpret_u8_p8): Likewise.
1733 (vreinterpret_u8_p16): Likewise.
1734 (vreinterpretq_u8_s8): Likewise.
1735 (vreinterpretq_u8_s16): Likewise.
1736 (vreinterpretq_u8_s32): Likewise.
1737 (vreinterpretq_u8_s64): Likewise.
1738 (vreinterpretq_u8_f32): Likewise.
1739 (vreinterpretq_u8_u16): Likewise.
1740 (vreinterpretq_u8_u32): Likewise.
1741 (vreinterpretq_u8_u64): Likewise.
1742 (vreinterpretq_u8_p8): Likewise.
1743 (vreinterpretq_u8_p16): Likewise.
1744 (vreinterpret_u16_s8): Likewise.
1745 (vreinterpret_u16_s16): Likewise.
1746 (vreinterpret_u16_s32): Likewise.
1747 (vreinterpret_u16_s64): Likewise.
1748 (vreinterpret_u16_f32): Likewise.
1749 (vreinterpret_u16_u8): Likewise.
1750 (vreinterpret_u16_u32): Likewise.
1751 (vreinterpret_u16_u64): Likewise.
1752 (vreinterpret_u16_p8): Likewise.
1753 (vreinterpret_u16_p16): Likewise.
1754 (vreinterpretq_u16_s8): Likewise.
1755 (vreinterpretq_u16_s16): Likewise.
1756 (vreinterpretq_u16_s32): Likewise.
1757 (vreinterpretq_u16_s64): Likewise.
1758 (vreinterpretq_u16_f32): Likewise.
1759 (vreinterpretq_u16_u8): Likewise.
1760 (vreinterpretq_u16_u32): Likewise.
1761 (vreinterpretq_u16_u64): Likewise.
1762 (vreinterpretq_u16_p8): Likewise.
1763 (vreinterpretq_u16_p16): Likewise.
1764 (vreinterpret_u32_s8): Likewise.
1765 (vreinterpret_u32_s16): Likewise.
1766 (vreinterpret_u32_s32): Likewise.
1767 (vreinterpret_u32_s64): Likewise.
1768 (vreinterpret_u32_f32): Likewise.
1769 (vreinterpret_u32_u8): Likewise.
1770 (vreinterpret_u32_u16): Likewise.
1771 (vreinterpret_u32_u64): Likewise.
1772 (vreinterpret_u32_p8): Likewise.
1773 (vreinterpret_u32_p16): Likewise.
1774 (vreinterpretq_u32_s8): Likewise.
1775 (vreinterpretq_u32_s16): Likewise.
1776 (vreinterpretq_u32_s32): Likewise.
1777 (vreinterpretq_u32_s64): Likewise.
1778 (vreinterpretq_u32_f32): Likewise.
1779 (vreinterpretq_u32_u8): Likewise.
1780 (vreinterpretq_u32_u16): Likewise.
1781 (vreinterpretq_u32_u64): Likewise.
1782 (vreinterpretq_u32_p8): Likewise.
1783 (vreinterpretq_u32_p16): Likewise.
1785 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1787 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1789 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1792 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1793 (vqnegd_s64): Likewise.
1794 (vqabs_s64): Likewise.
1795 (vqabsd_s64): Likewise.
1797 2014-04-22 Richard Henderson <rth@redhat.com>
1799 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1800 computation to the top of the loop.
1802 2014-04-22 Renlin <renlin.li@arm.com>
1803 Jiong Wang <jiong.wang@arm.com>
1805 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1806 * config/aarch64/aarch64.c (aarch64_layout_frame)
1807 (aarch64_initial_elimination_offset): Likewise.
1809 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1811 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1814 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1816 * machmode.h (bitwise_mode_for_mode): Declare.
1817 * stor-layout.h (bitwise_type_for_mode): Likewise.
1818 * stor-layout.c (bitwise_mode_for_mode): New function.
1819 (bitwise_type_for_mode): Likewise.
1820 * builtins.c (fold_builtin_memory_op): Use it instead of
1821 int_mode_for_mode and build_nonstandard_integer_type.
1823 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1825 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1826 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1827 (*-*-solaris2*): Simplify.
1828 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1829 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1830 *-*-solaris2.9* handling.
1832 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1834 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1835 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1837 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1838 * configure: Regenerate.
1840 * config/i386/sol2-9.h: Remove.
1842 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1843 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1844 Remove Solaris 9 references.
1846 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1848 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1849 (floatuns<GPI:mode><GPF:mode>2): Remove.
1850 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1851 and floatuns conversions.
1852 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1853 and floatuns conversions.
1854 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1855 (w1,w2): New mode attributes for inequal width conversions.
1857 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1859 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1860 the output asm format.
1862 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1864 * config/aarch64/aarch64-simd.md
1865 (aarch64_cm<optab>di): Always split.
1866 (*aarch64_cm<optab>di): New.
1867 (aarch64_cmtstdi): Always split.
1868 (*aarch64_cmtstdi): New.
1870 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1872 PR tree-optimization/60823
1873 * omp-low.c (ipa_simd_modify_function_body): Go through
1874 all SSA_NAMEs and for those refering to vector arguments
1875 which are going to be replaced adjust SSA_NAME_VAR and,
1876 if it is a default definition, change it into a non-default
1877 definition assigned at the beginning of function from new_decl.
1878 (ipa_simd_modify_stmt_ops): Rewritten.
1879 * tree-dfa.c (set_ssa_default_def): When removing default def,
1880 check for NULL loc instead of NULL *loc.
1882 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1884 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1885 restrictions on core registers for DImode values in Thumb2.
1887 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1889 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1890 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1892 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1894 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1895 (*iordi_notzesidi_di): Likewise.
1896 (*iordi_notsesidi_di): Likewise.
1898 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1900 * config/arm/arm-protos.h (tune_params): New struct members.
1901 * config/arm/arm.c: Initialise tune_params per processor.
1902 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1903 for speed, based on new tune_params.
1905 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1907 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1909 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1911 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1913 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1914 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1915 (vrnda_f64): Likewise.
1916 (vrndi_f64): Likewise.
1917 (vrndm_f64): Likewise.
1918 (vrndn_f64): Likewise.
1919 (vrndp_f64): Likewise.
1920 (vrndx_f64): Likewise.
1922 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1924 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1925 GET_MODE_SIZE argument is enum machine_mode.
1927 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1930 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1931 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1933 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1936 * asan.c (asan_emit_stack_protection): Force the base to align to
1937 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1938 appropriate bits if STRICT_ALIGNMENT.
1939 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1941 (expand_used_vars): Leave a space in the stack frame for alignment
1942 if STRICT_ALIGNMENT.
1944 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1946 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1948 (gimple_store_p): Likewise.
1949 (gimple_assign_load_p): Likewise.
1950 (gimple_assign_cast_p): Likewise.
1951 (gimple_clobber_p): Likewise.
1953 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1954 rather than a gimple.
1955 (gimple_assign_cast_p): Likewise.
1957 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1960 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1961 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1963 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1964 more debug information for E500 if -mdebug=reg.
1966 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1969 * config/i386/i386.c (ix86_expand_builtin)
1970 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1971 register for target RTX.
1972 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1974 2014-04-18 Cong Hou <congh@google.com>
1976 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1977 the widen-mult pattern by handling two operands with different sizes,
1978 and operands whose size is smaller than half of the result type.
1980 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1982 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1983 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1984 (do_estimate_edge_time): Compute it.
1985 * ipa-inline.c (want_inline_small_function_p): Bypass
1986 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1988 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1990 * ipa-inline.c (spec_rem): New static variable.
1991 (dump_overall_stats): New function.
1992 (dump_inline_stats): New function.
1994 2014-04-18 Richard Henderson <rth@redhat.com>
1996 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1997 to GET_MODE_SIZE, not a reg_class_t.
1999 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2001 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2002 (vsx_xxmrglw_<mode>): Likewise.
2004 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2007 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2008 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2009 (rs6000_init_hard_regno_mode_ok): Likewise.
2011 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2013 * ipa-inline.c (inline_small_functions): Account only non-cold
2015 * doc/invoke.texi (inline-unit-growth): Update documentation.
2017 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2019 * config/rs6000/rs6000.md (addti3, subti3): New.
2021 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2024 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2025 comment. Check optimize_insn_for_size_p instead of
2026 optimize_insn_for_speed_p.
2028 2014-04-17 Martin Jambor <mjambor@suse.cz>
2030 * gimple-iterator.c (gsi_start_edge): New function.
2031 * gimple-iterator.h (gsi_start_edge): Declare.
2032 * tree-sra.c (single_non_eh_succ): New function.
2033 (disqualify_ops_if_throwing_stmt): Renamed to
2034 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2035 having one non-EH successor BB.
2036 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2037 generate loads into replacements.
2038 (sra_modify_assign): Likewise and and also use the simple path for
2040 (sra_modify_function_body): Commit statements on edges.
2042 2014-04-17 Richard Biener <rguenther@suse.de>
2045 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
2046 comparison results and add clarifying comment.
2048 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2050 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
2051 (blank_mode): Initialize it.
2052 (emit_mode_size_inline, emit_mode_nunits_inline,
2053 emit_mode_inner_inline): New functions.
2054 (emit_insn_modes_h): Call them and surround their output with
2055 #if GCC_VERSION >= 4001 ... #endif.
2056 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
2057 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
2058 mode_* arrays if the argument is __builtin_constant_p.
2059 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
2060 is enum machine_mode.
2062 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2064 * passes.c (opt_pass::execute): Adjust.
2065 (pass_manager::execute_pass_mode_switching): Likewise.
2066 (early_local_passes::execute): Likewise.
2067 (execute_one_pass): Pass cfun to the pass's execute method.
2068 * tree-pass.h (opt_pass::execute): Add function * argument.
2069 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2070 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2071 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2072 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2073 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2074 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
2075 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
2076 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2077 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2078 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
2079 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
2080 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
2081 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
2082 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
2083 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2084 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2085 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2086 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
2087 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
2088 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2089 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2090 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2091 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2092 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2093 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2094 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2095 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2096 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2097 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2100 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2102 * passes.c (opt_pass::gate): Take function * argument.
2103 (gate_all_early_local_passes): Merge into
2104 (early_local_passes::gate): this.
2105 (gate_all_early_optimizations): Merge into
2106 (all_early_optimizations::gate): this.
2107 (gate_all_optimizations): Mege into
2108 (all_optimizations::gate): this.
2109 (gate_all_optimizations_g): Merge into
2110 (all_optimizations_g::gate): this.
2111 (gate_rest_of_compilation): Mege into
2112 (rest_of_compilation::gate): this.
2113 (gate_postreload): Merge into
2114 (postreload::gate): this.
2115 (dump_one_pass): Pass cfun to the pass's gate method.
2116 (execute_ipa_summary_passes): Likewise.
2117 (execute_one_pass): Likewise.
2118 (ipa_write_summaries_2): Likewise.
2119 (ipa_write_optimization_summaries_1): Likewise.
2120 (ipa_read_summaries_1): Likewise.
2121 (ipa_read_optimization_summaries_1): Likewise.
2122 (execute_ipa_stmt_fixups): Likewise.
2123 * tree-pass.h (opt_pass::gate): Add function * argument.
2124 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
2125 combine-stack-adj.c, combine.c, compare-elim.c,
2126 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2127 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
2128 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
2129 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
2130 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2131 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2132 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2133 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
2134 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
2135 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
2136 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2137 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2138 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
2139 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2140 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2141 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2142 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2143 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2144 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2145 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2146 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2147 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2148 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2149 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
2150 var-tracking.c, vtable-verify.c, web.c: Adjust.
2152 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2154 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2155 * configure: Regenerate.
2157 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2159 * passes.c (dump_one_pass): don't check pass->has_gate.
2160 (execute_ipa_summary_passes): Likewise.
2161 (execute_one_pass): Likewise.
2162 (ipa_write_summaries_2): Likewise.
2163 (ipa_write_optimization_summaries_1): Likewise.
2164 (ipa_read_optimization_summaries_1): Likewise.
2165 (execute_ipa_stmt_fixups): Likewise.
2166 * tree-pass.h (pass_data::has_gate): Remove.
2167 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2168 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2169 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2170 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2171 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2172 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2173 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2174 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2175 gimple-low.c, gimple-ssa-isolate-paths.c,
2176 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2177 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2178 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2179 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2180 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2181 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2182 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2183 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2184 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2185 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2186 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2187 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2188 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2189 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2190 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2191 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2192 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2193 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2194 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2195 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2196 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2197 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2200 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2202 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2204 * passes.c (pass_manager::register_dump_files_1): Merge into
2205 (pass_manager::register_dump_files): this, and remove its handling of
2206 properties since the pass always has the properties anyway.
2207 (pass_manager::pass_manager): Adjust.
2209 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2211 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2212 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2213 dealing with properties.
2214 (pass_manager::register_dump_files): Adjust.
2216 2014-03-20 Mark Wielaard <mjw@redhat.com>
2218 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2219 then represent the bound as normal constant value.
2221 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2224 Forward port from 4.8 branch
2225 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2227 * config/i386/bmiintrin.h (_blsi_u32): New.
2231 (_blsmsk_u32): Ditto.
2232 (_blsmsk_u64): Ditto.
2233 (_tzcnt_u32): Ditto.
2234 (_tzcnt_u64): Ditto.
2236 2014-04-17 Kito Cheng <kito@0xlab.org>
2238 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2240 2014-04-17 Richard Biener <rguenther@suse.de>
2243 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2244 boolean results for comparisons.
2246 2014-04-17 Richard Biener <rguenther@suse.de>
2248 PR tree-optimization/60836
2249 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2250 initial PHI args to be gimple values.
2252 2014-04-17 Richard Biener <rguenther@suse.de>
2254 PR tree-optimization/60841
2255 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2256 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2257 of stmts to SLP build.
2258 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2259 (vect_analyze_slp): Likewise.
2260 (vect_analyze_slp_instance): Likewise.
2261 (vect_build_slp_tree): Limit overall SLP tree growth.
2262 * tree-vectorizer.h (vect_analyze_data_refs,
2263 vect_analyze_slp): Adjust prototypes.
2265 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2267 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2270 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2272 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2273 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2274 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2275 for TARGET_SLOW_PSHUFB
2277 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2279 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2280 * config/i386/i386.c (intel_cost): Ditto.
2282 2014-04-17 Joey Ye <joey.ye@arm.com>
2284 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2286 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2288 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2291 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2293 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2294 (type_all_derivations_known_p): New predicate.
2295 (type_all_ctors_visible_p): New predicate.
2296 (type_possibly_instantiated_p): New predicate.
2297 (get_odr_type): Compute all_derivations_known.
2298 (dump_odr_type): Dump the flag.
2299 (maybe_record_type): Cleanup.
2300 (record_target_from_binfo): Add bases_to_consider array;
2301 record bases for types w/o instances and skip CXX destructor.
2302 (possible_polymorphic_call_targets_1): Add bases_to_consider
2303 and consider_construction parameters; check if type may have instance.
2304 (get_polymorphic_call_info): Set maybe_in_construction to true
2305 when we know nothing.
2306 (record_targets_from_bases): Skip CXX destructors; they are
2307 never called for types in construction.
2308 (possible_polymorphic_call_targets): Do not record target when
2309 type may not have instance.
2311 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2314 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2315 external aliases alive, too.
2317 2014-04-16 Andrew Pinski <apinski@cavium.com>
2319 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2322 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2324 * final.c (compute_alignments): Do not apply loop alignment to a block
2325 falling through to the exit.
2327 2014-04-16 Catherine Moore <clm@codesourcery.com>
2329 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2330 Adjust constraints for microMIPS store patterns.
2332 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2334 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2336 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2338 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2339 (append_use): Run at -O0.
2340 (append_vdef): Likewise.
2341 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2342 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2344 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2346 PR tree-optimization/60844
2347 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2348 (propagate_op_to_single_use, remove_visited_stmt_chain,
2349 linearize_expr, repropagate_negates, reassociate_bb): Use it
2350 instead of gsi_remove.
2352 2014-04-16 Martin Jambor <mjambor@suse.cz>
2354 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2355 ipa_transforms_to_apply.
2356 (cgraph_function_versioning): Assert that old_node has empty
2357 ipa_transforms_to_apply.
2358 * trans-mem.c (ipa_tm_create_version): Likewise.
2359 * tree-inline.c (tree_function_versioning): Do not duplicate
2360 ipa_transforms_to_apply.
2362 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2365 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2367 Pass necessary as flags on 64-bit Solaris/x86.
2368 Use lowercase relocs for x86_64-*-*.
2369 * configure: Regenerate.
2371 2014-04-15 Jan Hubicka <jh@suse.cz>
2373 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2374 (maybe_record_node, likely_target_p): Use it.
2376 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2379 Revert following patch
2381 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2384 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2385 software floating point or no floating point registers, do not
2386 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2387 in GPRs that occurs after we tested for GPRs that would never be
2390 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2391 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2392 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2393 specifically allow DDmode, since that does not use the SPE SIMD
2396 2014-03-21 Mark Wielaard <mjw@redhat.com>
2398 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2399 as unsigned or int depending on type and value used.
2401 2014-04-15 Richard Biener <rguenther@suse.de>
2403 PR rtl-optimization/56965
2404 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2405 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2407 * alias.c (true_dependence_1): Do not call
2408 nonoverlapping_component_refs_p.
2409 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2410 nonoverlapping_component_refs_p.
2411 (indirect_refs_may_alias_p): Likewise.
2413 2014-04-15 Teresa Johnson <tejohnson@google.com>
2415 * cfg.c (dump_bb_info): Fix flags check.
2416 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2418 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2420 PR rtl-optimization/60663
2421 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2424 2014-04-15 Richard Biener <rguenther@suse.de>
2426 * lto-streamer.h (LTO_major_version): Bump to 4.
2428 2014-04-15 Richard Biener <rguenther@suse.de>
2430 * common.opt (lto_partition_model): New enum.
2431 (flto-partition=): Merge separate options with a single with argument,
2432 add -flto-partition=one support.
2433 * flag-types.h (enum lto_partition_model): Declare.
2434 * opts.c (finish_options): Remove duplicate -flto-partition=
2436 * lto-wrapper.c (run_gcc): Adjust.
2438 2014-04-15 Richard Biener <rguenther@suse.de>
2440 * alias.c (ncr_compar): New function.
2441 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2443 2014-04-15 Richard Biener <rguenther@suse.de>
2445 * alias.c (record_component_aliases): Do not walk BINFOs.
2447 2014-04-15 Richard Biener <rguenther@suse.de>
2449 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2450 Add struct function argument and adjust.
2451 (find_func_aliases_for_call): Likewise.
2452 (find_func_aliases): Likewise.
2453 (find_func_clobbers): Likewise.
2454 (intra_create_variable_infos): Likewise.
2455 (compute_points_to_sets): Likewise.
2456 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2458 2014-04-15 Richard Biener <rguenther@suse.de>
2460 * tree.c (iterative_hash_expr): Use enum tree_code_class
2461 to store TREE_CODE_CLASS.
2462 (tree_block): Likewise.
2463 (tree_set_block): Likewise.
2464 * tree.h (fold_build_pointer_plus_loc): Use
2465 convert_to_ptrofftype_loc.
2467 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2470 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2473 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2475 * cfgloop.h (struct loop): Move force_vectorize down.
2476 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2477 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2478 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2479 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2480 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2481 * tree-core.h (enum annot_expr_kind): Add new kind values.
2482 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2483 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2485 * tree.def (ANNOTATE_EXPR): Tweak comment.
2487 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2489 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2492 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2494 * tree.h (TYPE_IDENTIFIER): Declare.
2495 * tree.c (subrange_type_for_debug_p): Use it.
2496 * godump.c (go_format_type): Likewise.
2497 * dwarf2out.c (is_cxx_auto, modified_type_die,
2498 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2499 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2501 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2504 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2506 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2508 * config/i386/i386.c (examine_argument): Return bool. Return true if
2509 parameter should be passed in memory.
2510 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2511 (construct_container): Update calls to examine_argument.
2512 (function_arg_advance_64): Ditto.
2513 (return_in_memory_32): Merge with ix86_return_in_memory.
2514 (return_in_memory_64): Ditto.
2515 (return_in_memory_ms_64): Ditto.
2517 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2519 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2520 * coverage.c (coverage_compute_profile_id): Handle externally visible
2523 2014-04-14 Martin Jambor <mjambor@suse.cz>
2525 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2526 DECL_DISREGARD_INLINE_LIMITS functions.
2528 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2531 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2533 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2536 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2537 optimize_insn_for_speed_p instead of
2538 optimize_function_for_speed_p.
2540 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2542 * doc/invoke.texi (free): Document AArch64.
2544 2014-04-14 Richard Biener <rguenther@suse.de>
2546 PR tree-optimization/60042
2547 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2548 (insert_into_preds_of_block): Do not prevent PHI insertion
2549 for REFERENCE exprs here ...
2550 (eliminate_dom_walker::before_dom_children): ... but prevent
2551 their use here under similar conditions when applied to the
2552 IL after PRE optimizations.
2554 2014-04-14 Richard Biener <rguenther@suse.de>
2556 * passes.def: Move early points-to after early SRA.
2558 2014-04-14 Richard Biener <rguenther@suse.de>
2560 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2561 check for which sign-changes we allow when forwarding
2562 a converted value into a switch.
2564 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2566 * stor-layout.c (place_field): Finalize non-constant offset for the
2569 2014-04-14 Richard Biener <rguenther@suse.de>
2571 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2573 (expand_switch_using_bit_tests_p): Likewise.
2574 (process_switch): Compute and pass on speed_p based on the
2576 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2577 optimize_bb_for_speed_p.
2579 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2581 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2582 * function.h (struct function): Rename has_force_vect_loops into
2583 has_force_vectorize_loops.
2584 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2585 (input_struct_function_base): Likewise.
2586 * lto-streamer-out.c (output_cfg): Likewise.
2587 (output_struct_function_base): Likewise.
2588 * omp-low.c (expand_omp_simd): Likewise.
2589 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2590 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2591 (version_loop_for_if_conversion): Likewise.
2592 (tree_if_conversion): Likewise.
2593 (main_tree_if_conversion): Likewise.
2594 (gate_tree_if_conversion): Likewise.
2595 * tree-inline.c (copy_loops): Likewise.
2596 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2597 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2598 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2599 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2600 * tree-vectorizer.c (vectorize_loops): Likewise.
2601 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2603 2014-04-14 Richard Biener <rguenther@suse.de>
2606 * lto-streamer-out.c (wrap_refs): New function.
2607 (lto_output): Wrap symbol references in global initializes in
2608 type-preserving MEM_REFs.
2610 2014-04-14 Christian Bruel <christian.bruel@st.com>
2612 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2614 2014-04-14 Christian Bruel <christian.bruel@st.com>
2616 * config/sh/sh.md (setmemqi): New expand pattern.
2617 * config/sh/sh.h (CLEAR_RATIO): Define.
2618 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2619 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2621 2014-04-14 Richard Biener <rguenther@suse.de>
2624 * fold-const.c (negate_expr_p): Don't negate directional rounding
2626 (fold_negate_expr): Likewise.
2628 2014-04-14 Richard Biener <rguenther@suse.de>
2630 PR tree-optimization/59817
2631 PR tree-optimization/60453
2632 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2633 recursion to catch all CHRECs in the scalar evolution and restrict
2634 the predicate for the remains appropriately.
2636 2014-04-12 Catherine Moore <clm@codesourcery.com>
2638 * config/mips/constraints.md: Add new register constraint "kb".
2639 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2640 (*movhi_internal): Likewise.
2641 (*movqi_internal): Likewise.
2642 * config/mips/mips.h (M16_STORE_REGS): New register class.
2643 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2644 (REG_CLASS_CONTENTS): Likewise.
2645 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2647 2014-04-11 Tobias Burnus <burnus@net-b.de>
2650 * doc/invoke.texi (-Wformat-signedness): Document it.
2651 (Wformat=2): Mention that this enables -Wformat-signedness.
2653 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2655 * common/config/epiphany/epiphany-common.c
2656 (epiphany_option_optimization_table): Enable section anchors by
2657 default at -O1 or higher.
2658 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2659 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2660 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2661 carries no extra cost.
2662 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2663 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2664 * config/epiphany/predicates.md (memclob_operand): New predicate.
2665 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2666 Use memclob_operand predicate and X constraint for operand 3.
2668 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2670 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2671 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2674 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2676 PR rtl-optimization/60651
2677 * mode-switching.c (optimize_mode_switching): Make sure to emit
2678 sets of a lower numbered entity before sets of a higher numbered
2679 entity to a mode of the same or lower priority.
2680 When creating a seginfo for a basic block that starts with a code
2681 label, move the insertion point past the code label.
2682 (new_seginfo): Document and enforce requirement that
2683 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2684 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2685 * doc/tm.texi: Regenerate.
2687 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2690 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2692 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2694 * BASE-VER: Set to 4.10.0.
2696 2014-04-11 Tobias Burnus <burnus@net-b.de>
2699 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2700 * doc/gcc.texi (Service): Update description in the @menu
2701 * doc/invoke.texi (Option Summary): Remove misplaced and
2704 2014-04-11 Steve Ellcey <sellcey@mips.com>
2705 Jakub Jelinek <jakub@redhat.com>
2708 * expr.c (convert_move): Use emit_store_flag_force instead of
2709 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2712 2014-04-11 Richard Biener <rguenther@suse.de>
2715 * varasm.c (assemble_alias): Avoid endless error reporting
2716 recursion by setting TREE_ASM_WRITTEN.
2718 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2720 * config/s390/s390.md: Add a splitter for NOT rtx.
2722 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2724 PR rtl-optimization/60663
2725 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2727 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2728 Jakub Jelinek <jakub@redhat.com>
2731 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2732 flag from decl_node to node.
2734 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2737 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2738 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2739 ameliorating the cases where it can be.
2741 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2744 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2746 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2747 (loadsync_<mode>): Change mode.
2748 (load_quadpti, store_quadpti): New.
2749 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2750 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2751 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2753 2014-04-09 Cong Hou <congh@google.com>
2756 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2759 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2761 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2762 instead of vnor to exploit possible fusion opportunity in the
2764 (altivec_expand_vec_perm_const_le): Likewise.
2766 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2768 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2769 (loadsync_<mode>): Change mode.
2770 (load_quadpti, store_quadpti): New.
2771 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2772 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2774 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2777 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2778 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2779 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2781 2014-04-08 Richard Biener <rguenther@suse.de>
2784 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2785 a 64bit widest int print double-int similar to on HWI64 hosts.
2787 2014-04-08 Richard Biener <rguenther@suse.de>
2789 PR tree-optimization/60785
2790 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2791 default defs properly.
2793 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2795 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2796 (Weffc++): Likewise.
2798 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2800 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2801 set completep to false rather than true.
2803 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2806 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2807 ARM_TARGET2_DWARF_FORMAT.
2809 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2812 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2813 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2816 2014-04-07 Richard Biener <rguenther@suse.de>
2818 PR tree-optimization/60766
2819 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2820 (may_eliminate_iv): Convert cand_value_at result to desired type.
2822 2014-04-07 Jason Merrill <jason@redhat.com>
2825 * common.opt (-fno-gnu-unique): Add.
2826 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2828 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2830 * haifa-sched.c: Fix outdated function reference and minor
2831 grammar errors in introductory comment.
2833 2014-04-07 Richard Biener <rguenther@suse.de>
2836 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2838 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2840 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2843 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2844 size accounting for thunks.
2845 (pa_asm_output_mi_thunk): Use final_start_function() and
2846 final_end_function() to output function start and end directives.
2848 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2850 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2851 device specific ISA/ feature information. Remove short_sp and
2852 errata_skip ds. Add avr_device_specific_features enum to have device
2854 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2855 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2856 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2857 updated device specific info.
2858 * config/avr/avr-mcus.def: Merge device specific details to
2859 dev_attribute field.
2860 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2862 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2863 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2864 assembler if RMW isa supported by current device.
2865 * config/avr/genmultilib.awk: Update as device info structure changed.
2866 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2868 2014-04-04 Cong Hou <congh@google.com>
2870 PR tree-optimization/60656
2871 * tree-vect-stmts.c (supportable_widening_operation):
2872 Fix a bug that elements in a vector with vect_used_by_reduction
2873 property are incorrectly reordered when the operation on it is not
2874 consistant with the one in reduction operation.
2876 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2878 PR rtl-optimization/60155
2879 * gcse.c (record_set_data): New function.
2880 (single_set_gcse): New function.
2881 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2882 (hoist_code): Likewise.
2883 (get_pressure_class_and_nregs): Likewise.
2885 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2887 * explow.c (probe_stack_range): Emit a final optimization blockage.
2889 2014-04-04 Anthony Green <green@moxielogic.com>
2891 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2894 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2897 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2898 flags are set only during streaming.
2899 * ipa.c (process_references, walk_polymorphic_call_targets,
2900 symtab_remove_unreachable_nodes): Drop bodies of always inline
2901 after early inlining.
2902 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2904 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2905 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2908 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2911 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2914 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2916 (cortex_a53_fdivd): Likewise.
2918 2014-04-04 Martin Jambor <mjambor@suse.cz>
2921 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2923 * cgraph.c (clone_of_p): Also return true if thunks match.
2924 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2925 cgraph_function_or_thunk_node and an obsolete comment.
2926 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2928 (build_function_decl_skip_args): Likewise.
2929 (set_new_clone_decl_and_node_flags): New function.
2930 (duplicate_thunk_for_node): Likewise.
2931 (redirect_edge_duplicating_thunks): Likewise.
2932 (cgraph_clone_node): New parameter args_to_skip, pass it to
2933 redirect_edge_duplicating_thunks which is called instead of
2934 cgraph_redirect_edge_callee.
2935 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2936 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2938 2014-04-04 Jeff Law <law@redhat.com>
2941 * config/arm/predicates.md (const_int_I_operand): New predicate.
2942 (const_int_M_operand): Similarly.
2943 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2945 (insv_t2, extv_reg, extzv_t2): Likewise.
2946 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2947 (pop_multiple_with_writeback_and_return): Likewise.
2948 (vfp_pop_multiple_with_writeback): Likewise
2950 2014-04-04 Richard Biener <rguenther@suse.de>
2953 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2954 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2956 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2957 * gimplify.c (gimple_add_tmp_var_fn): New function.
2958 * gimple-expr.h (create_tmp_reg_fn): Declare.
2959 * gimple-expr.c (create_tmp_reg_fn): New function.
2960 * gimple-low.c (record_vars_into): Don't change cfun.
2961 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2962 code generation without cfun.
2964 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2967 * Makefile.in (install-driver): Fix shell scripting.
2969 2014-04-03 Cong Hou <congh@google.com>
2971 PR tree-optimization/60505
2972 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2973 threshold of number of iterations below which no vectorization
2975 * tree-vect-loop.c (new_loop_vec_info):
2976 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2977 * tree-vect-loop.c (vect_analyze_loop_operations):
2978 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2979 * tree-vect-loop.c (vect_transform_loop):
2980 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2981 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2982 of iterations of the loop and see if we should build the epilogue.
2984 2014-04-03 Richard Biener <rguenther@suse.de>
2986 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2987 (streamer_tree_cache_create): Adjust.
2988 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2989 to allow optional nodes array.
2990 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2991 (streamer_tree_cache_append): Likewise.
2992 (streamer_tree_cache_create): Create nodes array optionally
2993 as specified by parameter.
2994 * lto-streamer-out.c (create_output_block): Avoid maintaining
2995 the node array in the writer cache.
2996 (DFS_write_tree): Remove assertion.
2997 (produce_asm_for_decls): Free the out decl state hash table early.
2998 * lto-streamer-in.c (lto_data_in_create): Adjust for
2999 streamer_tree_cache_create prototype change.
3001 2014-04-03 Richard Biener <rguenther@suse.de>
3003 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3004 set TREE_CHAIN to NULL_TREE.
3006 2014-04-03 Richard Biener <rguenther@suse.de>
3008 PR tree-optimization/60740
3009 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3010 over all GIMPLE_COND operands.
3012 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3014 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3015 (Weffc++): Remove Scott's numbering, merge lists and reference
3018 2014-04-03 Nick Clifton <nickc@redhat.com>
3020 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3023 2014-04-03 Martin Jambor <mjambor@suse.cz>
3025 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3026 mention gcc_unreachable before failing.
3027 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3030 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3033 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3034 inconsistent code and instead mark the context inconsistent.
3035 (possible_polymorphic_call_targets): For inconsistent contexts
3036 return empty complete list.
3038 2014-04-02 Anthony Green <green@moxielogic.com>
3040 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3041 (extendqisi2, extendhisi2): Define.
3042 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
3043 (WCHAR_TYPE): Change to unsigned int.
3045 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3047 PR tree-optimization/60733
3048 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
3049 insertion point for PHI candidates to be the end of the feeding
3050 block for the PHI argument.
3052 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
3054 PR rtl-optimization/60650
3055 * lra-constraints.c (process_alt_operands): Decrease reject for
3056 earlyclobber matching.
3058 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3060 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
3062 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3064 * config/spu/spu.c (pad_bb): Do not crash when the last
3065 insn is CODE_FOR_blockage.
3067 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3069 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
3070 lies outside the target mode.
3072 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3075 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3076 software floating point or no floating point registers, do not
3077 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3078 in GPRs that occurs after we tested for GPRs that would never be
3081 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3082 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3083 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3084 specifically allow DDmode, since that does not use the SPE SIMD
3087 2014-04-02 Richard Biener <rguenther@suse.de>
3090 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
3091 MODE_INTs. Properly use negv_optab.
3092 (expand_abs): Likewise.
3094 2014-04-02 Richard Biener <rguenther@suse.de>
3097 * Makefile.in (install-driver): Guard extra installs with special
3100 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
3102 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3105 2014-04-01 Richard Henderson <rth@redhat.com>
3108 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
3109 alternative enabled before register allocation.
3111 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
3113 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
3114 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
3116 (nios2_large_got_address): Remove unneeded 'sym' parameter.
3117 (nios2_got_address): Update nios2_large_got_address call site.
3118 (nios2_delegitimize_address): New function.
3119 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
3120 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
3121 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
3123 2014-04-01 Martin Husemann <martin@duskware.de>
3125 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
3128 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
3130 PR rtl-optimization/60604
3131 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
3132 check from register_operand.
3133 (register_operand): Redefine in terms of general_operand.
3134 (nonmemory_operand): Use register_operand for the non-constant cases.
3136 2014-04-01 Richard Biener <rguenther@suse.de>
3138 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
3140 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
3142 * doc/invoke.texi (mapp-regs): Clarify.
3144 2014-03-31 Ulrich Drepper <drepper@gmail.com>
3146 * config/i386/avx512fintrin.h (__v32hi): Define type.
3147 (__v64qi): Likewise.
3148 (_mm512_set1_epi8): Define.
3149 (_mm512_set1_epi16): Define.
3150 (_mm512_set4_epi32): Define.
3151 (_mm512_set4_epi64): Define.
3152 (_mm512_set4_pd): Define.
3153 (_mm512_set4_ps): Define.
3154 (_mm512_setr4_epi64): Define.
3155 (_mm512_setr4_epi32): Define.
3156 (_mm512_setr4_pd): Define.
3157 (_mm512_setr4_ps): Define.
3158 (_mm512_setzero_epi32): Define.
3160 2014-03-31 Martin Jambor <mjambor@suse.cz>
3163 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3164 callsite_arguments_match_p. Updated all callers. Also check types of
3165 corresponding formal parameters and actual arguments.
3166 (not_all_callers_have_enough_arguments_p) Renamed to
3167 some_callers_have_mismatched_arguments_p.
3169 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3171 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3173 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3176 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3179 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3181 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3182 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3184 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3185 Use FMAMODE_NOVF512 mode iterator.
3186 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3187 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3188 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3189 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3191 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3192 Use VF_128_256 mode iterator.
3193 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3196 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3198 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3199 static chain if needed.
3201 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3204 * lra-constraints.c (index_part_to_reg): New.
3205 (process_address): Use it.
3207 2014-03-27 Jeff Law <law@redhat.com>
3208 Jakub Jelinek <jakub@redhat.com>
3211 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3212 gen_rtx_{PLUS,MULT} to build up the address expression.
3214 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3215 creating non-canonical RTL.
3217 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3220 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3221 functions; reorganize to make cheap checks first.
3222 (inline_small_functions): Do not estimate growth when dumping;
3224 * ipa-inline.h (inline_summary): Add min_size.
3225 (growth_likely_positive): New function.
3226 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3227 (set_cond_stmt_execution_predicate): Cleanup.
3228 (estimate_edge_size_and_time): Compute min_size.
3229 (estimate_calls_size_and_time): Likewise.
3230 (estimate_node_size_and_time): Likewise.
3231 (inline_update_overall_summary): Update min_size.
3232 (do_estimate_edge_time): Likewise.
3233 (do_estimate_edge_size): Update.
3234 (do_estimate_edge_hints): Update.
3235 (growth_likely_positive): New function.
3237 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3240 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3241 also if addr has VOIDmode.
3243 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3245 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3246 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3248 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3249 instructions as well as AdvancedSIMD loads.
3251 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3253 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3254 Use crypto_aese type.
3255 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3256 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3257 crypto_aese, crypto_aesmc. Move to types.md.
3258 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3260 * config/arm/iterators.md (crypto_type): Likewise.
3262 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3264 * cgraph.c: Include expr.h and tree-dfa.h.
3265 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3268 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3271 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3272 regs from checking multi-reg pseudos.
3274 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3276 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3278 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3280 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3281 if it would clobber the stack pointer, even temporarily.
3283 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3285 * mode-switching.c: Make small adjustments to the top comment.
3287 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3289 * config/rs6000/constraints.md (wD constraint): New constraint to
3290 match the constant integer to get the top DImode/DFmode out of a
3291 vector in a VSX register.
3293 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3294 match the constant integer to get the top DImode/DFmode out of a
3295 vector in a VSX register.
3297 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3300 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3303 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3304 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3306 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3307 Optimize vec_extract of 64-bit values, where the value being
3308 extracted is in the top word, where we can use scalar
3309 instructions. Add direct move and store support. Combine the big
3310 endian/little endian vector select load support into a single insn.
3311 (vsx_extract_<mode>_internal1): Likewise.
3312 (vsx_extract_<mode>_internal2): Likewise.
3313 (vsx_extract_<mode>_load): Likewise.
3314 (vsx_extract_<mode>_store): Likewise.
3315 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3316 combined into vsx_extract_<mode>_load.
3317 (vsx_extract_<mode>_one_le): Likewise.
3319 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3320 define the top 64-bit vector element.
3322 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3325 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3326 Document vec_vbpermq builtin.
3329 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3330 enable use of xxsldwi and xxpermdi builtin functions.
3331 (vec_xxpermdi): Likewise.
3333 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3334 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3336 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3338 PR rtl-optimization/60650
3339 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3341 (find_spills_for): New.
3342 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3343 Spill all pseudos on the second iteration.
3345 2014-03-27 Marek Polacek <polacek@redhat.com>
3348 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3351 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3353 * config/s390/s390.c (s390_can_use_return_insn): Check for
3354 call-saved FPRs on 31 bit.
3356 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3359 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3360 if they need regimplification, just drop them instead of
3361 calling gimple_regimplify_operands on them.
3363 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3366 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3367 (aarch64_frame_pointer_required): Adjust logic.
3368 (aarch64_can_eliminate): Adjust logic.
3369 (aarch64_override_options_after_change): Adjust logic.
3371 2014-03-27 Dehao Chen <dehao@google.com>
3373 * ipa-inline.c (early_inliner): Update node's inline info.
3375 2014-03-26 Dehao Chen <dehao@google.com>
3377 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3378 compiler inserted conditional jumps for NAN float check.
3380 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3382 * ubsan.h (ubsan_create_data): Change second argument's type
3383 to const location_t *.
3384 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3386 (ubsan_create_data): Change second argument to const location_t *PLOC.
3387 Create Loc field whenever PLOC is non-NULL.
3388 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3389 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3393 * real.c (real_to_integer2): Change type of low to UHWI.
3395 2014-03-26 Tobias Burnus <burnus@net-b.de>
3397 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3398 (CILK_SELF_SPECS): New define.
3399 (driver_self_specs): Use it.
3401 2014-03-26 Richard Biener <rguenther@suse.de>
3403 * tree-pretty-print.c (percent_K_format): Implement special
3404 case for LTO and its stripped down BLOCK tree.
3406 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3409 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3411 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3412 one range is range_int_cst_p, but not both, at least optimize
3413 addition/subtraction of 0 and multiplication by 0 or 1.
3414 * gimple-fold.c (gimple_fold_call): Fold
3415 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3416 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3417 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3419 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3421 PR rtl-optimization/60452
3422 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3423 <case REG>: Return 1 for invalid offsets from the frame pointer.
3425 2014-03-26 Marek Polacek <polacek@redhat.com>
3428 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3431 2014-03-26 Marek Polacek <polacek@redhat.com>
3434 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3437 2014-03-26 Marek Polacek <polacek@redhat.com>
3440 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3441 multiplication in unsigned type.
3443 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3445 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3447 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3449 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3451 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3454 * cif-code.def (UNREACHABLE) New code.
3455 * ipa-inline.c (inline_small_functions): Skip edges to
3456 __builtlin_unreachable.
3457 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3458 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3459 predicate to __bulitin_unreachable.
3460 (set_cond_stmt_execution_predicate): Fix issue when
3461 invert_tree_comparison returns ERROR_MARK.
3462 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3463 propagate to inline clones.
3464 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3466 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3467 * cgraphclones.c (cgraph_clone_node): If call destination is already
3468 ureachable, do not redirect it back.
3469 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3472 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3474 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3475 Do not modify inline clones.
3477 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3479 * config/i386/i386.md (general_sext_operand): New mode attr.
3480 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3481 don't generate (sign_extend (const_int)).
3482 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3483 operands[2]. Use We constraint instead of <i> and
3484 <general_sext_operand> predicate instead of <general_operand>.
3485 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3486 * config/i386/constraints.md (We): New constraint.
3487 * config/i386/predicates.md (x86_64_sext_operand,
3488 sext_operand): New predicates.
3490 2014-03-25 Martin Jambor <mjambor@suse.cz>
3493 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3494 inconsistent devirtualizations to __builtin_unreachable.
3496 2014-03-25 Marek Polacek <polacek@redhat.com>
3499 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3501 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3503 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3504 order of elements for big-endian.
3506 2014-03-25 Richard Biener <rguenther@suse.de>
3509 * gimplify-me.c (gimple_regimplify_operands): Update the
3512 2014-03-25 Martin Jambor <mjambor@suse.cz>
3515 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3516 (lto_output_varpool_node): Likewise.
3517 (input_overwrite_node): Likewise.
3518 (input_varpool_node): Likewise.
3520 2014-03-25 Richard Biener <rguenther@suse.de>
3522 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3523 (run_gcc): Likewise.
3525 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3527 * combine.c (simplify_compare_const): Add MODE argument.
3528 Handle mode_width 0 as very large mode_width.
3529 (try_combine, simplify_comparison): Adjust callers.
3531 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3532 type to avoid signed integer overflow.
3533 * explow.c (plus_constant): Likewise.
3535 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3537 * doc/generic.texi: Correct typos.
3539 2014-03-24 Tobias Burnus <burnus@net-b.de>
3541 * doc/invoke.texi (-flto): Expand section about
3542 using static libraries with LTO.
3544 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3546 PR rtl-optimization/60501
3547 * optabs.def (addptr3_optab): New optab.
3548 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3549 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3550 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3552 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3554 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3556 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3558 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3561 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3562 (_mm256_undefined_ps): Define.
3563 (_mm256_undefined_pd): Define.
3564 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3565 (_mm_undefined_pd): Define.
3566 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3567 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3568 (_mm512_undefined_ps): Define.
3569 (_mm512_undefined_pd): Define.
3570 Use _mm*_undefined_*.
3571 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3573 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3575 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3576 (lshr_simd): DI mode added.
3577 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3578 (aarch64_ushr_simddi): Likewise.
3579 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3580 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3581 (vshrd_n_u64): Likewise.
3583 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3585 * Makefile.in (s-macro_list): Depend on cc1.
3587 2014-03-23 Teresa Johnson <tejohnson@google.com>
3589 * ipa-utils.c (ipa_print_order): Use specified dump file.
3591 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3593 PR rtl-optimization/60601
3594 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3596 * gcc.c (eval_spec_function): Initialize save_growing_value.
3598 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3601 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3602 code == MINUS_EXPR, never swap op0 with op1.
3604 * toplev.c (init_local_tick): Avoid signed integer multiplication
3606 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3607 shift by first operand's bitsize.
3609 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3612 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3614 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3615 TARGET_ISA_64BIT_P(x).
3617 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3619 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3620 pattern for vector nor instead of subtract from splat(-1).
3621 (altivec_expand_vec_perm_const_le): Likewise.
3623 2014-03-21 Richard Henderson <rth@twiddle.net>
3626 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3627 related insns after epilogue_completed.
3629 2014-03-21 Martin Jambor <mjambor@suse.cz>
3632 * cgraph.h (symtab_node): New flag body_removed.
3633 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3634 when removing bodies.
3635 * symtab.c (dump_symtab_base): Dump body_removed flag.
3636 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3637 had their bodies removed.
3639 2014-03-21 Martin Jambor <mjambor@suse.cz>
3642 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3645 2014-03-21 Richard Biener <rguenther@suse.de>
3647 PR tree-optimization/60577
3648 * tree-core.h (struct tree_base): Document nothrow_flag use
3649 in VAR_DECL_NONALIASED.
3650 * tree.h (VAR_DECL_NONALIASED): New.
3651 (may_be_aliased): Adjust.
3652 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3654 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3656 * expr.c (expand_expr_real_1): Remove outdated comment.
3658 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3661 * ira.c (adjust_cleared_regs): Call copy_rtx on
3662 *reg_equiv[REGNO (loc)].src_p before passing it to
3663 simplify_replace_fn_rtx.
3666 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3667 into CONST, put pic register as first operand of PLUS. Use
3668 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3670 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3672 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3674 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3676 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3677 around for store forwarding issue in the FPU on the UT699.
3678 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3679 loads and operations if -mfix-ut699 is specified.
3680 (divtf3_hq): Tweak attribute.
3681 (sqrttf2_hq): Likewise.
3683 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3685 * calls.c (store_one_arg): Remove incorrect const qualification on the
3686 type of the temporary.
3687 * cfgexpand.c (expand_return): Likewise.
3688 * expr.c (expand_constructor): Likewise.
3689 (expand_expr_real_1): Likewise.
3691 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3693 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3696 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3699 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3701 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3703 * config/arm/aarch-common-protos.h
3704 (alu_cost_table): Fix spelling of "extend".
3705 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3707 2014-03-19 Richard Biener <rguenther@suse.de>
3710 * tree-core.h (tree_type_common): Re-order pointer members
3711 to reduce recursion depth during GC walks.
3713 2014-03-19 Marek Polacek <polacek@redhat.com>
3716 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3717 before accessing it.
3719 2014-03-19 Richard Biener <rguenther@suse.de>
3722 * lto-streamer-in.c (input_function): In WPA stage do not drop
3725 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3727 PR tree-optimization/60559
3728 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3729 with build_zero_cst assignment.
3731 2014-03-18 Kai Tietz <ktietz@redhat.com>
3733 PR rtl-optimization/56356
3734 * sdbout.c (sdbout_parms): Verify that parms'
3735 incoming argument is valid.
3736 (sdbout_reg_parms): Likewise.
3738 2014-03-18 Richard Henderson <rth@redhat.com>
3741 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3742 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3743 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3745 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3747 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3748 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3749 Italicize plugin event names in description. Explain that
3750 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3751 Remind that no GCC functions should be called after PLUGIN_FINISH.
3752 Explain what pragmas with expansion are.
3754 2014-03-18 Martin Liska <mliska@suse.cz>
3756 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3757 gimple call statement is update.
3758 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3759 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3761 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3764 * ubsan.c (ubsan_instrument_unreachable): Call
3765 initialize_sanitizer_builtins.
3766 (ubsan_pass): Likewise.
3769 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3770 varpool_finalize_decl instead of rest_of_decl_compilation.
3772 2014-03-18 Richard Biener <rguenther@suse.de>
3774 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3775 by using bitmap_and_compl instead of bitmap_and_compl_into.
3776 (df_rd_transfer_function): Likewise.
3778 2014-03-18 Richard Biener <rguenther@suse.de>
3780 * doc/lto.texi (fresolution): Fix typo.
3782 2014-03-18 Richard Biener <rguenther@suse.de>
3784 * doc/invoke.texi (flto): Update for changes in 4.9.
3786 2014-03-18 Richard Biener <rguenther@suse.de>
3788 * doc/loop.texi: Remove section on the removed lambda framework.
3789 Update loop docs with recent changes in preserving loop structure.
3791 2014-03-18 Richard Biener <rguenther@suse.de>
3793 * doc/lto.texi (-fresolution): Document.
3795 2014-03-18 Richard Biener <rguenther@suse.de>
3797 * doc/contrib.texi: Adjust my name.
3799 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3802 * internal-fn.c: Include diagnostic-core.h.
3803 (expand_BUILTIN_EXPECT): New function.
3804 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3805 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3806 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3807 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3809 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3810 Revert 3 argument __builtin_expect code.
3811 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3812 * gimple-fold.c (gimple_fold_call): Likewise.
3813 * tree.h (fold_builtin_expect): New prototype.
3814 * builtins.c (build_builtin_expect_predicate): Add predictor
3815 argument, if non-NULL, create 3 argument __builtin_expect.
3816 (fold_builtin_expect): No longer static. Add ARG2 argument,
3817 pass it through to build_builtin_expect_predicate.
3818 (fold_builtin_2): Adjust caller.
3819 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3820 * internal-fn.def (BUILTIN_EXPECT): New.
3822 2014-03-18 Tobias Burnus <burnus@net-b.de>
3825 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3826 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3827 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3829 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3832 * predict.c (combine_predictions_for_bb): Fix up formatting.
3833 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3834 fill what it points to if non-NULL.
3835 (tree_predict_by_opcode): Adjust caller, use the predictor.
3836 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3838 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3840 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3841 proper constant for the store mode.
3843 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3845 * symtab.c (change_decl_assembler_name): Fix transparent alias
3848 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3850 * config/aarch64/aarch64.c: Correct the comments about the
3851 aarch64 stack layout.
3853 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3855 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3856 check for GF_OMP_FOR_KIND_FOR.
3858 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3860 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3861 ymm and zmm register names.
3863 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3866 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3867 note creation for the 2010-08-31 changes.
3869 2014-03-17 Marek Polacek <polacek@redhat.com>
3872 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3873 as -fno-tree-loop-vectorize.
3874 (expand_omp_simd): Likewise.
3876 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3878 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3879 (eligible_for_call_delay): New prototype.
3880 * config/sparc/sparc.c (tls_call_delay): Rename into...
3881 (eligible_for_call_delay): ...this. Return false if the instruction
3882 cannot be put in the delay slot of a branch.
3883 (eligible_for_restore_insn): Simplify.
3884 (eligible_for_return_delay): Return false if the instruction cannot be
3885 put in the delay slot of a branch and simplify.
3886 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3887 put in the delay slot of a branch.
3888 * config/sparc/sparc.md (fix_ut699): New attribute.
3889 (tls_call_delay): Delete.
3890 (in_call_delay): Reimplement.
3891 (eligible_for_sibcall_delay): Rename into...
3892 (in_sibcall_delay): ...this.
3893 (eligible_for_return_delay): Rename into...
3894 (in_return_delay): ...this.
3895 (in_branch_delay): Reimplement.
3896 (in_uncond_branch_delay): Delete.
3897 (in_annul_branch_delay): Delete.
3899 2014-03-14 Richard Henderson <rth@redhat.com>
3902 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3903 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3904 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3905 (floathi splitters): Remove.
3906 (float<SWI48x>xf2): New pattern.
3907 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3908 code that tried to handle DImode for 32-bit, but which was excluded
3909 by the pattern's condition. Drop allocation of stack temporary.
3910 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3911 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3912 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3913 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3914 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3915 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3916 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3917 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3918 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3919 (*float<SWI48x><X87MODEF>2_i387): Remove.
3920 (all float _with_temp splitters): Remove.
3921 (*float<SWI48x><MODEF>2_i387): New pattern.
3922 (*float<SWI48><MODEF>2_sse): New pattern.
3923 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3924 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3926 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3927 Marek Polacek <polacek@redhat.com>
3930 * common.opt (dump_base_name_prefixed): New Variable.
3931 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3932 if x_dump_base_name_prefixed is already set, set it at the end.
3934 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3936 PR rtl-optimization/60508
3937 * lra-constraints.c (get_reload_reg): Add new parameter
3939 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3940 Pass the new parameter values.
3942 2014-03-14 Richard Biener <rguenther@suse.de>
3944 * common.opt: Revert unintented changes from r205065.
3947 2014-03-14 Richard Biener <rguenther@suse.de>
3950 * cfghooks.c (split_block): Properly adjust all loops the
3951 block was a latch of.
3953 2014-03-14 Martin Jambor <mjambor@suse.cz>
3956 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3959 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3962 * config/avr/avr.c (avr_set_current_function): Pass function name
3963 through default_strip_name_encoding before sanity checking instead
3964 of skipping the first char of the assembler name.
3966 2014-03-13 Richard Henderson <rth@redhat.com>
3969 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3970 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3971 * config/i386/i386-protos.h: Likewise.
3972 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3973 in the expander instead of a splitter.
3974 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3975 any possibility of requiring a memory.
3976 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3977 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3978 (fp branch splitters): Update for ix86_split_fp_branch.
3979 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3980 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3981 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3982 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3983 (*fop_<MODEF>_3_i387): Likewise.
3984 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3985 (splitters for the fop_* register patterns): Remove.
3986 (fscalexf4_i387): Rename from *fscalexf4_i387.
3987 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3989 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3991 PR tree-optimization/59779
3992 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3993 type for bitsize and maxsize instead of HOST_WIDE_INT.
3995 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3997 PR rtl-optimization/57320
3998 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3999 the CFG after thread_prologue_and_epilogue_insns.
4001 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4003 PR rtl-optimization/57189
4004 * lra-constraints.c (process_alt_operands): Disfavor spilling
4007 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4009 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4011 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4013 PR tree-optimization/59025
4015 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4016 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4018 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4021 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4022 calls of avr_out_plus_1.
4024 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4026 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4027 BB's single pred and update the father loop's latch info later.
4029 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4031 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4035 (VEC_base): Likewise.
4036 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4037 registers, we need to swap double words in little endian mode.
4039 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4040 to be a container mode for 128-bit integer operations added in ISA
4041 2.07. Unlike TImode and PTImode, the preferred register set is
4042 the Altivec/VMX registers for the 128-bit operations.
4044 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
4046 (rs6000_split_128bit_ok_p): Likewise.
4048 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
4049 macros for creating ISA 2.07 normal and overloaded builtin
4050 functions with 3 arguments.
4051 (BU_P8V_OVERLOAD_3): Likewise.
4052 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
4053 for use as overloaded functions.
4054 (VPERM_1TI_UNS): Likewise.
4055 (VSEL_1TI): Likewise.
4056 (VSEL_1TI_UNS): Likewise.
4057 (ST_INTERNAL_1ti): Likewise.
4058 (LD_INTERNAL_1ti): Likewise.
4059 (XXSEL_1TI): Likewise.
4060 (XXSEL_1TI_UNS): Likewise.
4061 (VPERM_1TI): Likewise.
4062 (VPERM_1TI_UNS): Likewise.
4063 (XXPERMDI_1TI): Likewise.
4064 (SET_1TI): Likewise.
4065 (LXVD2X_V1TI): Likewise.
4066 (STXVD2X_V1TI): Likewise.
4067 (VEC_INIT_V1TI): Likewise.
4068 (VEC_SET_V1TI): Likewise.
4069 (VEC_EXT_V1TI): Likewise.
4070 (EQV_V1TI): Likewise.
4071 (NAND_V1TI): Likewise.
4072 (ORC_V1TI): Likewise.
4073 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
4074 added in ISA 2.07. Add both normal 'altivec' builtins, and the
4076 (VADDUQM): Likewise.
4077 (VSUBCUQ): Likewise.
4078 (VADDEUQM): Likewise.
4079 (VADDECUQ): Likewise.
4080 (VSUBEUQM): Likewise.
4081 (VSUBECUQ): Likewise.
4083 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
4084 __int128_t and __uint128_t types.
4085 (__uint128_type): Likewise.
4086 (altivec_categorize_keyword): Add support for vector __int128_t,
4087 vector __uint128_t, vector __int128, and vector unsigned __int128
4088 as a container type for TImode operations that need to be done in
4089 VSX/Altivec registers.
4090 (rs6000_macro_to_expand): Likewise.
4091 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
4092 to support 128-bit integer instructions vaddcuq, vadduqm,
4093 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
4094 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
4096 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
4097 for V1TImode, and set up preferences to use VSX/Altivec registers.
4098 Setup VSX reload handlers.
4099 (rs6000_debug_reg_global): Likewise.
4100 (rs6000_init_hard_regno_mode_ok): Likewise.
4101 (rs6000_preferred_simd_mode): Likewise.
4102 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
4103 (easy_altivec_constant): Likewise.
4104 (output_vec_const_move): Likewise.
4105 (rs6000_expand_vector_set): Convert V1TImode set and extract to
4107 (rs6000_expand_vector_extract): Likewise.
4108 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
4110 (rs6000_const_vec): Add support for V1TImode.
4111 (rs6000_emit_le_vsx_load): Swap double words when loading or
4112 storing TImode/V1TImode.
4113 (rs6000_emit_le_vsx_store): Likewise.
4114 (rs6000_emit_le_vsx_move): Likewise.
4115 (rs6000_emit_move): Add support for V1TImode.
4116 (altivec_expand_ld_builtin): Likewise.
4117 (altivec_expand_st_builtin): Likewise.
4118 (altivec_expand_vec_init_builtin): Likewise.
4119 (altivec_expand_builtin): Likewise.
4120 (rs6000_init_builtins): Add support for V1TImode type. Add
4121 support for ISA 2.07 128-bit integer builtins. Define type names
4122 for the VSX/Altivec vector types.
4123 (altivec_init_builtins): Add support for overloaded vector
4124 functions with V1TImode type.
4125 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
4126 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
4128 (rs6000_split_128bit_ok_p): Likewise.
4129 (rs6000_handle_altivec_attribute): Create V1TImode from vector
4130 __int128_t and vector __uint128_t.
4132 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
4133 and mode attributes.
4140 (VS_scalar): Likewise.
4141 (VS_double): Likewise.
4142 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
4144 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
4145 we support the ISA 2.07 128-bit integer arithmetic instructions.
4146 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
4147 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
4148 and TImode types for use with the builtin functions.
4149 (V1TI_type_node): Likewise.
4150 (unsigned_V1TI_type_node): Likewise.
4151 (intTI_type_internal_node): Likewise.
4152 (uintTI_type_internal_node): Likewise.
4154 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4155 128-bit builtin functions.
4156 (UNSPEC_VADDEUQM): Likewise.
4157 (UNSPEC_VADDECUQ): Likewise.
4158 (UNSPEC_VSUBCUQ): Likewise.
4159 (UNSPEC_VSUBEUQM): Likewise.
4160 (UNSPEC_VSUBECUQ): Likewise.
4161 (VM): Add V1TImode to vector mode iterators.
4163 (VI_unit): Likewise.
4164 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4165 (altivec_vaddcuq): Likewise.
4166 (altivec_vsubuqm): Likewise.
4167 (altivec_vsubcuq): Likewise.
4168 (altivec_vaddeuqm): Likewise.
4169 (altivec_vaddecuq): Likewise.
4170 (altivec_vsubeuqm): Likewise.
4171 (altivec_vsubecuq): Likewise.
4173 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4175 (BOOL_128): Likewise.
4176 (BOOL_REGS_OUTPUT): Likewise.
4177 (BOOL_REGS_OP1): Likewise.
4178 (BOOL_REGS_OP2): Likewise.
4179 (BOOL_REGS_UNARY): Likewise.
4180 (BOOL_REGS_AND_CR0): Likewise.
4182 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4183 128-bit integer builtin support.
4184 (vec_vadduqm): Likewise.
4185 (vec_vaddecuq): Likewise.
4186 (vec_vaddeuqm): Likewise.
4187 (vec_vsubecuq): Likewise.
4188 (vec_vsubeuqm): Likewise.
4189 (vec_vsubcuq): Likewise.
4190 (vec_vsubuqm): Likewise.
4192 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4193 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4194 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4195 128-bit integer add/subtract to ISA 2.07.
4197 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4199 * config/arc/arc.c (arc_predicate_delay_insns):
4200 Fix third argument passed to conditionalize_nonjump.
4202 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4204 * config/aarch64/aarch64-builtins.c
4205 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4206 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4207 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4208 instead of __builtin_lfloor.
4209 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4211 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4213 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4214 (tree_ssa_ifcombine_bb_1): New function.
4215 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4216 is an empty forwarder block to then_bb or vice versa and then_bb
4217 and else_bb are effectively swapped.
4219 2014-03-12 Christian Bruel <christian.bruel@st.com>
4222 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4223 REG_CFA_DEF_CFA note.
4224 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4225 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4227 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4229 PR tree-optimization/60454
4230 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4232 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4234 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4235 Do not define target_cpu_default2 to generic.
4236 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4237 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4238 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4240 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4241 Marc Glisse <marc.glisse@inria.fr>
4243 PR tree-optimization/60502
4244 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4245 instead of build_low_bits_mask.
4247 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4250 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4251 if there are multiple uses, but op doesn't live on E edge.
4252 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4253 clobber stmts before __builtin_unreachable.
4255 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4257 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4258 hard_frame_pointer_rtx.
4259 * cse.c (cse_insn): Remove volatile check.
4260 * cselib.c (cselib_process_insn): Likewise.
4261 * dse.c (scan_insn): Likewise.
4263 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4265 * config/arc/arc.c (conditionalize_nonjump): New function,
4267 (arc_ifcvt): ... this.
4268 (arc_predicate_delay_insns): Use it.
4270 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4272 * config/arc/predicates.md (extend_operand): During/after reload,
4273 allow const_int_operand.
4274 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4275 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4276 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4278 (umulsi3_highpart_i): Likewise.
4280 2014-03-11 Richard Biener <rguenther@suse.de>
4282 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4283 Add asserts to guard possible wrong-code bugs.
4285 2014-03-11 Richard Biener <rguenther@suse.de>
4287 PR tree-optimization/60429
4288 PR tree-optimization/60485
4289 * tree-ssa-structalias.c (set_union_with_increment): Properly
4290 take into account all fields that overlap the shifted vars.
4291 (do_sd_constraint): Likewise.
4292 (do_ds_constraint): Likewise.
4293 (get_constraint_for_ptr_offset): Likewise.
4295 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4297 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4298 (nios2_compute_frame_layout):
4299 Add calculation of cfun->machine->fp_save_offset.
4300 (nios2_expand_prologue): Correct setting of frame pointer register
4302 (nios2_expand_epilogue): Update recovery of stack pointer from
4303 frame pointer accordingly.
4304 (nios2_initial_elimination_offset): Update calculation of offset
4305 for eliminating to HARD_FRAME_POINTER_REGNUM.
4307 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4310 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4311 cgraph_get_create_node on VAR_DECLs.
4313 2014-03-10 Richard Biener <rguenther@suse.de>
4316 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4318 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4320 * config/vms/vms.opt (vms_float_format): New variable.
4322 2014-03-08 Tobias Burnus <burnus@net-b.de>
4324 * doc/invoke.texi (-fcilkplus): Update implementation status.
4326 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4327 Richard Biener <rguenther@suse.de>
4329 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4330 consistently accross all TUs.
4331 (run_gcc): Enable -fshort-double automatically at link at link-time
4332 and disallow override.
4334 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4337 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4338 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4339 if they can't be used.
4341 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4343 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4344 for Solaris 11/x86 ld.
4345 * configure: Regenerate.
4347 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4349 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4350 (LIB_TLS_SPEC): Save as ld_tls_libs.
4351 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4352 (HAVE_AS_IX86_TLSLDM): New test.
4353 * configure, config.in: Regenerate.
4354 * config/i386/i386.c (legitimize_tls_address): Fall back to
4355 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4356 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4357 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4358 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4360 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4362 * common.opt (fira-loop-pressure): Mark as optimization.
4364 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4366 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4367 an OpenMP mappable type.
4369 2014-03-06 Matthias Klose <doko@ubuntu.com>
4371 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4372 MULTILIB_OSDIRNAMES is not defined.
4374 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4375 Meador Inge <meadori@codesourcery.com>
4378 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4379 (arm_legitimize_address): Call legitimize_tls_address for any
4380 arm_tls_referenced_p expression, handle constant addend. Call it
4381 before testing for !TARGET_ARM.
4382 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4384 2014-03-06 Richard Biener <rguenther@suse.de>
4390 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4392 * tree-streamer.c (record_common_node): Assert we don't record
4393 nodes with type double.
4394 (preload_common_node): Skip type double, complex double and double
4395 pointer since it is now frontend dependent due to fshort-double option.
4397 2014-03-06 Richard Biener <rguenther@suse.de>
4399 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4400 or -fno-lto is specified and the linker has full plugin support.
4401 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4402 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4403 * lto-wrapper.c (merge_and_complain): Merge compile-time
4404 optimization levels.
4405 (run_gcc): And pass it through to the link options.
4407 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4411 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4413 * cselib.c (remove_useless_values): Skip to avoid quadratic
4414 behavior if the condition moved from...
4415 (cselib_process_insn): ... here holds.
4417 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4420 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4421 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4424 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4425 (TM_H): Add x86-tune.def.
4427 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4429 * config/aarch64/aarch64.c (generic_tunings):
4430 Use cortexa57_extra_costs.
4432 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4435 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4436 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4437 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4440 2014-03-04 Heiher <r@hev.cc>
4442 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4443 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4445 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4447 * config/i386/predicates.md (const2356_operand): Change to ...
4448 (const2367_operand): ... this.
4449 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4451 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4452 (*avx512pf_scatterpf<mode>sf): Ditto.
4453 (avx512pf_scatterpf<mode>df): Ditto.
4454 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4455 (*avx512pf_scatterpf<mode>df): Ditto.
4456 * config/i386/i386.c (ix86_expand_builtin): Update
4457 incorrect hint operand error message.
4459 2014-03-04 Richard Biener <rguenther@suse.de>
4461 * lto-section-in.c (lto_get_section_data): Fix const cast.
4463 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4465 * tree-streamer.c (record_common_node): Assert we don't record
4466 nodes with type double.
4467 (preload_common_node): Skip type double, complex double and double
4468 pointer since it is now frontend dependent due to fshort-double option.
4470 2014-03-04 Richard Biener <rguenther@suse.de>
4473 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4474 (lto_input_toplevel_asms): Likewise.
4475 * lto-section-in.c (lto_get_section_data): Instead do it here
4478 2014-03-04 Richard Biener <rguenther@suse.de>
4480 PR tree-optimization/60382
4481 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4482 dead PHIs a reduction.
4484 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4486 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4488 (_mm_prefetch): Move out of GCC target("sse") pragma.
4489 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4490 GCC target("prfchw") pragma.
4491 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4493 * config/i386/i386.c (ix86_option_override_internal): Enable
4494 -mprfchw with -mprefetchwt1.
4496 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4498 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4501 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4503 * opts.h (CL_PCH_IGNORE): Define.
4504 * targhooks.c (option_affects_pch_p):
4505 Return false for options that have CL_PCH_IGNORE set.
4506 * opt-functions.awk: Process PchIgnore.
4507 * doc/options.texi: Document PchIgnore.
4509 * config/arc/arc.opt (misize): Add PchIgnore property.
4511 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4513 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4514 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4515 constraint on constants to permit them being loaded into
4516 GENERAL_REGS or BASE_REGS.
4518 2014-03-03 Nick Clifton <nickc@redhat.com>
4520 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4521 anti-cacnonical alternatives.
4522 (negandhi3_real): New pattern.
4523 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4525 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4527 * config/avr/avr-mcus.def: Remove atxmega16x1.
4528 * config/avr/avr-tables.opt: Regenerate.
4529 * config/avr/t-multilib: Regenerate.
4530 * doc/avr-mmcu.texi: Regenerate.
4532 2014-03-03 Tobias Grosser <tobias@grosser.es>
4533 Mircea Namolaru <mircea.namolaru@inria.fr>
4535 PR tree-optimization/58028
4536 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4539 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4541 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4542 not handled by recognizers.
4544 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4547 * function.c (expand_function_end): Don't emit
4548 clobber_return_register sequence if clobber_after is a BARRIER.
4549 * cfgexpand.c (construct_exit_block): Append instructions before
4550 return_label to prev_bb.
4552 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4554 * config/rs6000/constraints.md: Document reserved use of "wc".
4556 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4559 * ipa.c (function_and_variable_visibility): When dissolving comdat
4560 group, also set all symbols to local.
4562 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4567 2013-12-14 Jan Hubicka <jh@suse.cz>
4569 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4571 2014-03-02 Jon Beniston <jon@beniston.com>
4577 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4578 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4579 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4580 (simple_return, *simple_return): New patterns
4581 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4582 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4584 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4586 * dwarf2out.c (gen_subprogram_die): Tidy.
4588 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4591 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4592 (*mov_t_msb_neg_negc): ... this new insn.
4594 2014-02-28 Jason Merrill <jason@redhat.com>
4597 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4600 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4603 * dwarf2out.c (decltype_auto_die): New static.
4604 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4605 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4606 (is_cxx_auto): Likewise.
4608 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4610 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4611 we are not using general regs only.
4613 2014-02-28 Richard Biener <rguenther@suse.de>
4616 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4617 previous fix and only allow to remove trivial pre-headers
4618 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4619 (remove_forwarder_block): Properly update the latch of a loop.
4621 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4624 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4625 (cselib_preserved_hash_table): New.
4626 (preserve_constants_and_equivs): Move preserved vals to it.
4627 (cselib_find_slot): Look it up first.
4628 (cselib_init): Initialize it.
4629 (cselib_finish): Release it.
4630 (dump_cselib_table): Dump it.
4632 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4635 * cselib.c (remove_useless_values): Skip to avoid quadratic
4636 behavior if the condition moved from...
4637 (cselib_process_insn): ... here holds.
4639 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4642 * var-tracking.c (vt_initialize): Apply the same condition to
4643 preserve the CFA base value.
4645 2014-02-28 Joey Ye <joey.ye@arm.com>
4648 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4649 if reload in progress or completed.
4651 2014-02-28 Tobias Burnus <burnus@net-b.de>
4654 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4657 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4659 * doc/tm.texi.in (Condition Code Status): Update documention for
4660 relative locations of cc0-setter and cc0-user.
4662 2014-02-27 Jeff Law <law@redhat.com>
4664 PR rtl-optimization/52714
4665 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4666 into two independent simple sets, if I3 is a jump, ensure the
4667 pattern we place into I3 is a (set (pc) ...).
4669 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4670 Jeff Law <law@redhat.com>
4672 PR rtl-optimization/49847
4673 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4674 are in different blocks.
4675 * doc/tm.texi (Condition Code Status): Update documention for
4676 relative locations of cc0-setter and cc0-user.
4678 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4681 * lra.c (lra_emit_add): Check SUBREG too.
4683 2014-02-27 Andreas Schwab <schwab@suse.de>
4685 * config/m68k/m68k.c (m68k_option_override): Disable
4686 -flive-range-shrinkage for classic m68k.
4687 (m68k_override_options_after_change): Likewise.
4689 2014-02-27 Marek Polacek <polacek@redhat.com>
4692 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4693 -Wmaybe-uninitialized.
4695 2014-02-27 Alan Modra <amodra@gmail.com>
4698 * reload1.c (emit_input_reload_insns): When reload_override_in,
4699 set old to rl->in_reg when rl->in_reg is a subreg.
4701 2014-02-26 Richard Biener <rguenther@suse.de>
4704 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4706 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4708 * common/config/i386/predicates.md (const1256_operand): Remove.
4709 (const2356_operand): New.
4710 (const_1_to_2_operand): Remove.
4711 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4712 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4713 (*avx512pf_gatherpf<mode>sf): Ditto.
4714 (avx512pf_gatherpf<mode>df): Ditto.
4715 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4716 (*avx512pf_gatherpf<mode>df): Ditto.
4717 (avx512pf_scatterpf<mode>sf): Ditto.
4718 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4719 (*avx512pf_scatterpf<mode>sf): Ditto.
4720 (avx512pf_scatterpf<mode>df): Ditto.
4721 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4722 (*avx512pf_scatterpf<mode>df): Ditto.
4723 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4725 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4727 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4728 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4729 (_mm512_mask_testn_epi64_mask): Move to ...
4730 * config/i386/avx512cdintrin.h: Here.
4731 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4732 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4733 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4734 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4735 TARGET_AVX512F from TARGET_AVX512CD.
4737 2014-02-26 Richard Biener <rguenther@suse.de>
4740 * ipa.c (walk_polymorphic_call_targets): Properly guard
4741 call to inline_update_overall_summary.
4743 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4746 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4747 and latches only if requested. Fix latch if it is removed.
4748 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4749 LOOPS_HAVE_PREHEADERS.
4751 2014-02-25 Andrew Pinski <apinski@cavium.com>
4753 * builtins.c (expand_builtin_thread_pointer): Create a new target
4754 when the target is NULL.
4756 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4758 PR rtl-optimization/60317
4759 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4760 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4761 * lra-assigns.c: Include params.h.
4762 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4763 other reload pseudos considerations.
4765 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4767 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4768 to use canonical form for nor<mode>3.
4770 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4773 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4776 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4778 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4779 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4780 (ix86_handle_option): Handle OPT_mprefetchwt1.
4781 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4782 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4784 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4785 OPTION_MASK_ISA_PREFETCHWT1.
4786 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4787 (PTA_PREFETCHWT1): New.
4788 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4789 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4790 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4791 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4792 (*prefetch_avx512pf_<mode>_: Change into ...
4793 (*prefetch_prefetchwt1_<mode>: This.
4794 * config/i386/i386.opt (mprefetchwt1): New.
4795 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4796 (_mm_prefetch): Handle intent to write.
4797 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4799 2014-02-25 Richard Biener <rguenther@suse.de>
4802 * emit-rtl.c (mem_attrs_htab): Remove.
4803 (mem_attrs_htab_hash): Likewise.
4804 (mem_attrs_htab_eq): Likewise.
4805 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4806 (init_emit_once): Do not allocate mem_attrs_htab.
4808 2014-02-25 Richard Biener <rguenther@suse.de>
4811 * lto-opts.c (lto_write_options): Output non-explicit conservative
4812 -fwrapv, -fno-trapv and -fno-strict-overflow.
4813 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4814 (run_gcc): And pass them through.
4816 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4818 * sel-sched.c (calculate_new_fences): New parameter ptime.
4819 Calculate it as a maximum over all fence cycles.
4820 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4821 Print the final schedule timing when sched_verbose.
4823 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4825 PR rtl-optimization/60292
4826 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4827 bit fot the fence instruction.
4829 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4831 * calls.h: Fix typo in comment.
4833 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4835 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4836 adjusting offsetable addresses.
4838 2014-02-24 Guozhi Wei <carrot@google.com>
4840 * sparseset.h (sparseset_pop): Fix the wrong index.
4842 2014-02-24 Walter Lee <walt@tilera.com>
4844 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4845 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4847 * common/config/tilegx/tilegx-common.c
4848 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4849 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4851 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4852 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4853 (tilegx_gimplify_va_arg_expr): Handle big endian.
4854 (tilegx_expand_unaligned_load): Ditto.
4855 (tilegx_expand_unaligned_store): Ditto.
4856 (TARGET_RETURN_IN_MSB): New.
4857 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4858 (TARGET_ENDIAN_DEFAULT): New.
4859 (TARGET_BIG_ENDIAN): Handle big endian.
4860 (BYTES_BIG_ENDIAN): Ditto.
4861 (WORDS_BIG_ENDIAN): Ditto.
4862 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4865 * config/tilegx/tilegx.md (extv): Handle big endian.
4867 (insn_st<n>): Ditto.
4868 (insn_st<n>_add<bitsuffix>): Ditto.
4869 (insn_stnt<n>): Ditto.
4870 (insn_stnt<n>_add<bitsuffix>):Ditto.
4871 (vec_interleave_highv8qi): Handle big endian.
4872 (vec_interleave_highv8qi_be): New.
4873 (vec_interleave_highv8qi_le): New.
4874 (insn_v1int_h): Handle big endian.
4875 (vec_interleave_lowv8qi): Handle big endian.
4876 (vec_interleave_lowv8qi_be): New.
4877 (vec_interleave_lowv8qi_le): New.
4878 (insn_v1int_l): Handle big endian.
4879 (vec_interleave_highv4hi): Handle big endian.
4880 (vec_interleave_highv4hi_be): New.
4881 (vec_interleave_highv4hi_le): New.
4882 (insn_v2int_h): Handle big endian.
4883 (vec_interleave_lowv4hi): Handle big endian.
4884 (vec_interleave_lowv4hi_be): New.
4885 (vec_interleave_lowv4hi_le): New.
4886 (insn_v2int_l): Handle big endian.
4887 (vec_interleave_highv2si): Handle big endian.
4888 (vec_interleave_highv2si_be): New.
4889 (vec_interleave_highv2si_le): New.
4890 (insn_v4int_h): Handle big endian.
4891 (vec_interleave_lowv2si): Handle big endian.
4892 (vec_interleave_lowv2si_be): New.
4893 (vec_interleave_lowv2si_le): New.
4894 (insn_v4int_l): Handle big endian.
4895 * config/tilegx/tilegx.opt (mbig-endian): New option.
4896 (mlittle-endian): New option.
4897 * doc/install.texi: Document tilegxbe-linux.
4898 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4900 2014-02-24 Martin Jambor <mjambor@suse.cz>
4903 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4904 there are no parameter descriptors.
4906 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4908 PR rtl-optimization/60268
4909 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4910 initialization to ...
4911 (sched_rgn_init): ... here.
4912 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4914 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4916 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4919 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4921 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4924 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4926 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4927 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4929 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4931 * config/microblaze/predicates.md: Add cmp_op predicate.
4932 * config/microblaze/microblaze.md: Add branch_compare instruction
4933 which uses cmp_op predicate and emits cmp insn before branch.
4934 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4935 to microblaze_expand_conditional_branch and consolidate logic.
4936 (microblaze_expand_conditional_branch): emit branch_compare
4937 insn instead of handling cmp op separate from branch insn.
4939 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4941 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4944 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4946 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4947 define_insn with define_expand and new define_insn
4948 *altivec_lve<VI_char>x_internal.
4949 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4950 and new define_insn *altivec_stve<VI_char>x_internal.
4951 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4953 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4955 (altivec_expand_stvex_be): New function.
4957 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4959 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4960 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4961 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4962 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4964 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4967 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4968 instead of emit_move_insn.
4970 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4972 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4974 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4975 gen_altivec_vsumsws.
4977 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4979 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4980 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4981 (altivec_lvxl_<mode>): New define_expand incorporating
4982 -maltivec=be semantics where needed.
4983 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4984 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4985 semantics where needed.
4986 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4987 (altivec_stvx_<mode>): New define_expand incorporating
4988 -maltivec=be semantics where needed.
4989 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4990 VM2 iterator instead of V4SI.
4991 (altivec_stvxl_<mode>): New define_expand incorporating
4992 -maltivec=be semantics where needed.
4993 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4994 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4995 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4996 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4997 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4998 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4999 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5000 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5001 ALTIVEC_BUILTIN_STVXL.
5002 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5003 (altivec_expand_stvx_be): Likewise.
5004 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5005 (altivec_expand_lvx_be): Likewise.
5006 (altivec_expand_stvx_be): Likewise.
5007 (altivec_expand_builtin): Add cases for
5008 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5009 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5010 (altivec_init_builtins): Add definitions for
5011 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5012 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5014 2014-02-21 Catherine Moore <clm@codesourcery.com>
5016 * doc/invoke.texi (mvirt, mno-virt): Document.
5017 * config/mips/mips.opt (mvirt): New option.
5018 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5020 2014-02-21 Richard Biener <rguenther@suse.de>
5022 PR tree-optimization/60276
5023 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5024 (STMT_VINFO_MIN_NEG_DIST): New macro.
5025 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5026 STMT_VINFO_MIN_NEG_DIST.
5027 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5028 made for negative dependence distances still hold.
5030 2014-02-21 Richard Biener <rguenther@suse.de>
5033 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5034 DECL_INITIAL for globals not in the current function context.
5036 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5038 PR tree-optimization/56490
5039 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5040 * tree-ssa-uninit.c: Include params.h.
5041 (compute_control_dep_chain): Add num_calls argument, return false
5042 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
5043 num_calls to recursive call.
5044 (find_predicates): Change dep_chain into normal array,
5045 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
5046 variable and adjust compute_control_dep_chain caller.
5047 (find_def_preds): Likewise.
5049 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
5051 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
5052 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
5054 2014-02-21 Nick Clifton <nickc@redhat.com>
5056 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
5057 (pushhi1): Likewise.
5058 (popqi1): Add mode to pre_dec.
5061 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5063 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
5064 mode for mask of V8SFmode permutation.
5066 2014-02-20 Richard Henderson <rth@redhat.com>
5069 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
5070 a new pseudo for OLDVAL.
5072 2014-02-20 Jakub Jelinek <jakub@redhat.com>
5075 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
5076 gen_reg_rtx if d->testing_p.
5077 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
5078 if d->testing_p and we will certainly return true.
5079 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
5082 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
5084 * emit-rtl.c (gen_reg_rtx): Assert that
5085 crtl->emit.regno_pointer_align_length is non-zero.
5087 2014-02-20 Richard Henderson <rth@redhat.com>
5090 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
5091 on failure the store back into EXPECT.
5093 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
5094 Sandra Loosemore <sandra@codesourcery.com>
5096 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
5097 * config/nios2/nios2.c (nios2_function_profiler): Add
5098 -fPIC (flag_pic == 2) support.
5099 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
5100 (nios2_large_offset_p): New function.
5101 (nios2_unspec_reloc_p): Move up position, update to use
5102 nios2_large_offset_p.
5103 (nios2_unspec_address): Remove function.
5104 (nios2_unspec_offset): New function.
5105 (nios2_large_got_address): New function.
5106 (nios2_got_address): Add large offset support.
5107 (nios2_legitimize_tls_address): Update usage of removed and new
5109 (nios2_symbol_binds_local_p): New function.
5110 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
5111 (nios2_legitimize_address): Update to use nios2_large_offset_p.
5112 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
5113 (nios2_print_operand): Merge H/L processing, add hiadj/lo
5114 processing for (const (unspec ...)).
5115 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
5117 2014-02-20 Richard Biener <rguenther@suse.de>
5119 * tree-cfg.c (replace_uses_by): Mark altered BBs before
5120 doing the substitution.
5121 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
5123 2014-02-20 Martin Jambor <mjambor@suse.cz>
5126 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
5127 info when checking whether lattices are bottom.
5129 2014-02-20 Richard Biener <rguenther@suse.de>
5132 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
5135 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
5138 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
5139 parameter specifying the scaling.
5140 (inline_call): Update.
5141 (want_inline_recursively): Guard division by zero.
5142 (recursive_inlining): Update.
5143 * ipa-inline.h (clone_inlined_nodes): Update.
5145 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5148 * config/i386/i386.c (classify_argument): Pass structures of size
5149 64 bytes or less in register.
5151 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5152 Kirill Yukhin <kirill.yukhin@intel.com>
5154 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5155 (_mm_rcp28_round_ss): Ditto.
5156 (_mm_rsqrt28_round_sd): Ditto.
5157 (_mm_rsqrt28_round_ss): Ditto.
5158 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5159 (_mm_rcp14_round_ss): Ditto.
5160 (_mm_rsqrt14_round_sd): Ditto.
5161 (_mm_rsqrt14_round_ss): Ditto.
5162 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5163 the first input operand, get rid of match_dup.
5164 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5166 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5168 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5169 operand as the first input operand, set type attribute.
5170 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5172 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5173 operand as the first input operand, set type attribute.
5175 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5177 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5180 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5183 * config/i386/i386.c (construct_container): Remove TFmode check
5184 for X86_64_INTEGER_CLASS.
5186 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5189 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5190 only when -Wpsabi is enabled.
5192 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5195 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5196 passing arrays in registers are the same as for structs, so remove the
5197 special case for them.
5199 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5201 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5202 destination type, extract only the valid bits if the source type is not
5203 integral and has a different mode.
5205 2014-02-19 Richard Biener <rguenther@suse.de>
5208 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5211 2014-02-19 Richard Biener <rguenther@suse.de>
5214 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5215 (ipa_modify_call_arguments): Emit an argument load explicitely and
5216 preserve virtual SSA form there and for the replacement call.
5217 Do not update SSA form nor free dominance info.
5219 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5221 * ipa.c (function_and_variable_visibility): Also clear WEAK
5222 flag when disolving COMDAT_GROUP.
5224 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5226 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5227 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5229 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5230 do more sanity checks.
5231 (detect_type_change): Return true when giving up early.
5232 (compute_complex_assign_jump_func): Fix type parameter of
5233 ipa_set_ancestor_jf.
5234 (compute_complex_ancestor_jump_func): Likewise.
5235 (update_jump_functions_after_inlining): Fix updating of
5237 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5239 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5241 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5242 inline clones when edge disappears.
5244 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5247 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5248 Split 64-bit moves into 2 patterns. Do not allow the use of
5249 direct move for TDmode in little endian, since the decimal value
5250 has little endian bytes within a word, but the 64-bit pieces are
5251 ordered in a big endian fashion, and normal subreg's of TDmode are
5253 (mov<mode>_64bit_dm): Likewise.
5254 (movtd_64bit_nodm): Likewise.
5256 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5258 PR tree-optimization/60174
5259 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5260 statement of an SSA_NAME that occurs in an abnormal PHI node.
5262 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5265 * final.c (SEEN_BB): Remove.
5266 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5267 (final_scan_insn): Don't force_source_line on second
5268 NOTE_INSN_BASIC_BLOCK.
5270 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5273 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5274 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5275 (type_natural_mode): Warn ABI change when %zmm register is not
5276 available for AVX512F vector value passing.
5278 2014-02-18 Kai Tietz <ktietz@redhat.com>
5281 * config/i386/i386.c (ix86_expand_prologue): Use value in
5282 rax register as displacement when restoring %r10 or %rax.
5283 Fix wrong offset when restoring both registers.
5285 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5287 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5288 assertion with conditional return.
5290 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5291 Uros Bizjak <ubizjak@gmail.com>
5294 * config/i386/driver-i386.c (host_detect_local_cpu): If
5295 YMM state is not saved by the OS, also clear has_f16c. Move
5296 CPUID 0x80000001 handling before YMM state saving checking.
5298 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5300 PR rtl-optimization/58960
5301 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5302 factored out from ...
5303 (sched_init): ... here.
5304 (free_global_sched_pressure_data): New, factored out from ...
5305 (sched_finish): ... here.
5306 * sched-int.h (free_global_sched_pressure_data): Declare.
5307 * sched-rgn.c (nr_regions_initial): New static global.
5308 (haifa_find_rgns): Initialize it.
5309 (schedule_region): Disable sched-pressure for the newly
5312 2014-02-17 Richard Biener <rguenther@suse.de>
5314 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5315 release SSA defs of pattern stmts.
5317 2014-02-17 Richard Biener <rguenther@suse.de>
5319 * tree-inline.c (expand_call_inline): Release the virtual
5320 operand defined by the call we are about to inline.
5322 2014-02-17 Richard Biener <rguenther@suse.de>
5324 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5326 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5327 Ilya Tocar <ilya.tocar@intel.com>
5329 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5330 arguments order in builtin.
5331 (_mm512_permutexvar_epi64): Ditto.
5332 (_mm512_mask_permutexvar_epi64): Ditto
5333 (_mm512_maskz_permutexvar_epi32): Ditto
5334 (_mm512_permutexvar_epi32): Ditto
5335 (_mm512_mask_permutexvar_epi32): Ditto
5337 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5339 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5340 (p8_vmrgow): Likewise.
5342 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5344 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5347 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5350 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5351 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5352 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5353 using direct move instructions on ISA 2.07. Also adjust
5354 instruction length for 64-bit.
5355 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5356 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5358 2014-02-15 Alan Modra <amodra@gmail.com>
5362 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5363 find_replacement on parts of insn rtl that might be reloaded.
5365 2014-02-15 Richard Biener <rguenther@suse.de>
5367 PR tree-optimization/60183
5368 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5369 (tree_ssa_phiprop): Calculate and free post-dominators.
5371 2014-02-14 Jeff Law <law@redhat.com>
5373 PR rtl-optimization/60131
5374 * ree.c (get_extended_src_reg): New function.
5375 (combine_reaching_defs): Use it rather than assuming location of REG.
5376 (find_and_remove_re): Verify first operand of extension is
5377 a REG before adding the insns to the copy list.
5379 2014-02-14 Roland McGrath <mcgrathr@google.com>
5381 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5382 * configure: Regenerated.
5383 * config.in: Regenerated.
5384 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5385 instead of ASM_SHORT.
5387 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5388 Richard Earnshaw <rearnsha@arm.com>
5390 PR rtl-optimization/59535
5391 * lra-constraints.c (process_alt_operands): Encourage alternative
5392 when unassigned pseudo class is superset of the alternative class.
5393 (inherit_reload_reg): Don't inherit when optimizing for code size.
5394 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5395 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5396 modes not less than 4 for Thumb1.
5398 2014-02-14 Kyle McMartin <kyle@redhat.com>
5401 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5403 2014-02-14 Richard Biener <rguenther@suse.de>
5405 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5406 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5407 Do not strip INDIRECT_REFs.
5409 2014-02-14 Richard Biener <rguenther@suse.de>
5412 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5413 DECL_FUNCTION_SPECIFIC_TARGET.
5414 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5415 * tree-streamer-out.c (pack_ts_target_option): Remove.
5416 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5417 (write_ts_function_decl_tree_pointers): Do not stream
5418 DECL_FUNCTION_SPECIFIC_TARGET.
5419 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5420 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5421 (lto_input_ts_function_decl_tree_pointers): Do not stream
5422 DECL_FUNCTION_SPECIFIC_TARGET.
5424 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5426 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5427 (get_initial_def_for_induction, vectorizable_induction): Ignore
5428 debug stmts when looking for exit_phi.
5429 (vectorizable_live_operation): Fix up condition.
5431 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5433 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5434 nreverse() because it changes the content of original tree list.
5436 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5438 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5439 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5441 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5443 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5444 GNU coding standards.
5446 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5449 * dwarf2out.c (gen_subprogram_die): Don't call
5450 add_calling_convention_attribute if subr_die is old_die.
5452 2014-02-13 Sharad Singhai <singhai@google.com>
5454 * doc/optinfo.texi: Fix order of nodes.
5456 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5458 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5459 operands[2], not operands[3].
5461 2014-02-13 Richard Biener <rguenther@suse.de>
5464 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5465 mention the possibility of an in-tree build.
5466 (CLooG): Update recommended version to 0.18.1, mention the
5467 possibility of an in-tree build and clarify that the ISL
5468 bundled with CLooG does not work.
5470 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5473 * expr.c (compress_float_constant): If x is a hard register,
5474 extend into a pseudo and then move to x.
5476 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5478 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5479 caused by bad second argument to warning_at() with -mhotpatch and
5480 nested functions (e.g. with gfortran).
5482 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5484 * opts.c (option_name): Remove "enabled by default" rider.
5486 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5488 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5490 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5491 Uros Bizjak <ubizjak@gmail.com>
5494 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5495 * configure: Regenerated.
5497 2014-02-12 Richard Biener <rguenther@suse.de>
5499 * vec.c (vec_prefix::calculate_allocation): Move as
5500 inline variant to vec.h.
5501 (vec_prefix::calculate_allocation_1): New out-of-line version.
5502 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5503 (vec_prefix::m_has_auto_buf): Rename to ...
5504 (vec_prefix::m_using_auto_storage): ... this.
5505 (vec_prefix::calculate_allocation): Inline the easy cases
5506 and dispatch to calculate_allocation_1 which doesn't need the
5508 (va_heap::reserve): Use gcc_checking_assert.
5509 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5510 m_using_auto_storage.
5511 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5513 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5514 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5515 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5517 2014-02-12 Richard Biener <rguenther@suse.de>
5519 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5520 when we found a dependence.
5522 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5524 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5526 (maybe_fold_stmt): ... into this new function.
5527 * omp-low.c (lower_omp): Update comment.
5529 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5532 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5535 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5537 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5538 identifiers in comments.
5539 (cortexa53_extra_costs): Likewise.
5540 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5541 (cortexa7_extra_costs): Likewise.
5542 (cortexa12_extra_costs): Likewise.
5543 (cortexa15_extra_costs): Likewise.
5544 (v7m_extra_costs): Likewise.
5546 2014-02-12 Richard Biener <rguenther@suse.de>
5549 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5550 of posix_memalign being successful.
5551 (lower_stmt): Restrict lowering of posix_memalign to when
5552 -ftree-bit-ccp is enabled.
5554 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5556 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5558 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5560 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5562 PR rtl-optimization/60116
5563 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5564 other_insn once the combination has been validated.
5566 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5569 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5571 * ipa-devirt.c: Include demangle.h
5572 (odr_violation_reported): New static variable.
5573 (add_type_duplicate): Update odr_violations.
5574 (maybe_record_node): Add completep parameter; update it.
5575 (record_target_from_binfo): Add COMPLETEP parameter;
5576 update it as needed.
5577 (possible_polymorphic_call_targets_1): Likewise.
5578 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5579 rename FINAL to COMPLETE.
5580 (record_targets_from_bases): Sanity check we found the binfo;
5581 fix COMPLETEP updating.
5582 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5583 parameter, fix computing of COMPLETEP.
5584 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5585 at LTO time do demangling.
5586 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5587 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5589 (gimple_get_virt_method_for_binfo): Likewise.
5590 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5591 gimple_get_virt_method_for_vtable): Update prototypes.
5593 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5596 * genautomata.c (add_presence_absence): Fix typo with
5597 {final_}presence_list.
5599 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5602 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5603 for VSX/Altivec vectors that land in GPR registers.
5605 2014-02-11 Richard Henderson <rth@redhat.com>
5606 Jakub Jelinek <jakub@redhat.com>
5609 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5610 around drhs if type conversion to lacc->type is not useless.
5612 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5614 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5616 (cortex-a57.cortex-a53): Likewise.
5617 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5619 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5621 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5624 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5626 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5627 add_options_for_arm_vfp3.
5629 2014-02-11 Jeff Law <law@redhat.com>
5632 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5633 object with an undesirable mode.
5635 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5638 * config/i386/sol2-9.h: New file.
5639 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5640 *-*-solaris2.9*): Use it.
5642 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5644 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5645 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5647 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5649 * config/microblaze/microblaze.c: Extend mcpu version format
5651 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5653 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5655 2014-02-10 Richard Henderson <rth@redhat.com>
5658 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5659 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5660 ms-abi vs -mno-accumulate-outgoing-args.
5661 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5662 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5665 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5668 * cfgexpand.c (expand_asm_operands): Attach source location to
5669 ASM_INPUT rtx objects.
5670 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5672 2014-02-10 Nick Clifton <nickc@redhat.com>
5674 * config/mn10300/mn10300.c (popcount): New function.
5675 (mn10300_expand_prologue): Include saved registers in stack usage
5678 2014-02-10 Jeff Law <law@redhat.com>
5681 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5682 when changing the SET_DEST of a prior insn to avoid an input reload.
5684 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5686 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5687 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5688 -mcall-openbsd, or -mcall-linux.
5689 (CC1_ENDIAN_BIG_SPEC): Remove.
5690 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5691 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5692 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5693 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5694 and %cc1_endian_default.
5695 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5697 2014-02-10 Richard Biener <rguenther@suse.de>
5699 PR tree-optimization/60115
5700 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5701 MEM_REF handling. Properly verify that the accesses are not
5702 out of the objects bound.
5704 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5706 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5709 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5711 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5712 proper constants and fix formatting.
5713 (possible_polymorphic_call_targets): Fix formatting.
5715 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5716 Ilya Tocar <ilya.tocar@intel.com>
5718 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5719 (_mm512_loadu_epi32): Renamed into...
5720 (_mm512_loadu_si512): This.
5721 (_mm512_storeu_epi32): Renamed into...
5722 (_mm512_storeu_si512): This.
5723 (_mm512_maskz_ceil_ps): Removed.
5724 (_mm512_maskz_ceil_pd): Ditto.
5725 (_mm512_maskz_floor_ps): Ditto.
5726 (_mm512_maskz_floor_pd): Ditto.
5727 (_mm512_floor_round_ps): Ditto.
5728 (_mm512_floor_round_pd): Ditto.
5729 (_mm512_ceil_round_ps): Ditto.
5730 (_mm512_ceil_round_pd): Ditto.
5731 (_mm512_mask_floor_round_ps): Ditto.
5732 (_mm512_mask_floor_round_pd): Ditto.
5733 (_mm512_mask_ceil_round_ps): Ditto.
5734 (_mm512_mask_ceil_round_pd): Ditto.
5735 (_mm512_maskz_floor_round_ps): Ditto.
5736 (_mm512_maskz_floor_round_pd): Ditto.
5737 (_mm512_maskz_ceil_round_ps): Ditto.
5738 (_mm512_maskz_ceil_round_pd): Ditto.
5739 (_mm512_expand_pd): Ditto.
5740 (_mm512_expand_ps): Ditto.
5741 * config/i386/i386.c (ix86_builtins): Remove
5742 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5743 (bdesc_args): Ditto.
5744 * config/i386/predicates.md (const1256_operand): New.
5745 (const_1_to_2_operand): Ditto.
5746 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5747 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5748 (*avx512pf_gatherpf<mode>sf): Ditto.
5749 (avx512pf_gatherpf<mode>df): Ditto.
5750 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5751 (*avx512pf_gatherpf<mode>df): Ditto.
5752 (avx512pf_scatterpf<mode>sf): Ditto.
5753 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5754 (*avx512pf_scatterpf<mode>sf): Ditto.
5755 (avx512pf_scatterpf<mode>df): Ditto.
5756 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5757 (*avx512pf_scatterpf<mode>df): Ditto.
5758 (avx512f_expand<mode>): Removed.
5759 (<shift_insn><mode>3<mask_name>): Change predicate type.
5761 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5763 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5764 not at the end of datarefs vector use ordered_remove to avoid
5765 reordering datarefs vector.
5768 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5769 mark local addressable non-static vars as GOVD_PRIVATE
5770 instead of GOVD_LOCAL.
5771 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5772 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5776 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5777 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5778 assume_aligned or alloc_align attributes.
5779 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5780 arguments. Handle also assume_aligned and alloc_align attributes.
5781 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5782 calls to functions with assume_aligned or alloc_align attributes.
5783 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5785 2014-02-08 Terry Guo <terry.guo@arm.com>
5787 * doc/invoke.texi: Document ARM -march=armv7e-m.
5789 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5791 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5792 flag on __cilkrts_rethrow builtin.
5795 * ipa-cp.c (determine_versionability): Fail at -O0
5796 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5797 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5800 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5803 * tree-inline.c (copy_forbidden): Fail for
5804 __attribute__((optimize (0))) functions.
5806 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5808 * varpool.c: Include pointer-set.h.
5809 (varpool_remove_unreferenced_decls): Variables in other partitions
5810 will not be output; be however careful to not lose information
5813 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5815 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5816 lookup in the vtable constructor.
5818 2014-02-07 Jeff Law <law@redhat.com>
5821 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5822 define_insn_and_split.
5824 * ipa-inline.c (inline_small_functions): Fix typos.
5826 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5828 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5829 (s390_can_use_return_insn): Declare.
5830 * config/s390/s390.h (EPILOGUE_USES): Define.
5831 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5833 (s390_chunkify_start): Handle return JUMP_LABELs.
5834 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5835 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5836 (s390_can_use_return_insn): New functions.
5837 (s390_fix_long_loop_prediction): Handle conditional returns.
5838 (TARGET_SET_UP_BY_PROLOGUE): Define.
5839 * config/s390/s390.md (ANY_RETURN): New code iterator.
5840 (*creturn, *csimple_return, return, simple_return): New patterns.
5842 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5844 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5845 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5846 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5847 REG_CFA_RESTORE list when deciding not to restore a register.
5849 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5851 * config/s390/s390.c: Include tree-pass.h and context.h.
5852 (s390_early_mach): New function, split out from...
5853 (s390_emit_prologue): ...here.
5854 (pass_data_s390_early_mach): New pass structure.
5855 (pass_s390_early_mach): New class.
5856 (s390_option_override): Create and register early_mach pass.
5857 Move to end of file.
5859 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5861 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5862 to match for the exit block.
5864 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5866 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5867 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5868 Reject misaligned operands.
5870 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5872 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5874 2014-02-07 Richard Biener <rguenther@suse.de>
5877 * gimple-low.c (lower_builtin_posix_memalign): New function.
5878 (lower_stmt): Call it to lower posix_memalign in a way
5879 to make alignment info accessible.
5881 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5884 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5885 __builtin_setjmp_receiver.
5887 2014-02-07 Richard Biener <rguenther@suse.de>
5890 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5891 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5892 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5893 Handle BUILT_IN_POSIX_MEMALIGN.
5894 (find_func_clobbers): Likewise.
5895 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5896 (call_may_clobber_ref_p_1): Likewise.
5898 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5901 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5904 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5907 * lto-cgraph.c (lto_output_node): Use
5908 symtab_get_symbol_partitioning_class.
5909 (lto_output_varpool_node): likewise.
5910 (symtab_get_symbol_partitioning_class): Move here from
5912 * cgraph.h (symbol_partitioning_class): Likewise.
5913 (symtab_get_symbol_partitioning_class): Declare.
5915 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5917 * ggc.h (ggc_internal_cleared_alloc): New macro.
5918 * vec.h (vec_safe_copy): Handle memory stats.
5919 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5920 * target-globals.c (save_target_globals): Likewise.
5922 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5925 * expr.c (emit_move_resolve_push): Export; be bit more selective
5926 on when to clear alias set.
5927 * expr.h (emit_move_resolve_push): Declare.
5928 * function.h (struct function): Add tail_call_marked.
5929 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5930 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5931 * config/i386/i386.md (TImode move expander): De not call
5933 (FP push expanders): Preserve memory attributes.
5934 * config/i386/sse.md (push<mode>1): Remove.
5935 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5936 (ix86_expand_push): Remove.
5937 * config/i386/mmx.md (push<mode>1): Remove.
5939 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5941 PR rtl-optimization/60030
5942 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5943 lopart with paradoxical subreg before shifting it up by hprec.
5945 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5947 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5948 Remove extra newline at end of file.
5949 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5950 (arm_issue_rate): Handle cortexa57.
5951 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5952 (cortex-a57.cortex-a53): Likewise.
5954 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5957 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5958 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5960 (arm_expand_prologue): Adjust all callers.
5961 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5962 info, registers also at the lowest numbered registers side. Use
5963 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5967 * var-tracking.c (adjust_mems): Before adding a SET to
5968 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5970 2014-02-06 Alan Modra <amodra@gmail.com>
5973 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5974 change SDmode to DDmode when lra_in_progress.
5976 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5979 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5980 free_data_ref on the dr first, and before goto again also set dr
5981 to the next dr. For simd_lane_access, free old datarefs[i] before
5982 overwriting it. For get_vectype_for_scalar_type failure, don't
5983 free_data_ref if simd_lane_access.
5985 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5988 * tree.h (opts_for_fn): New inline function.
5989 (opt_for_fn): Define.
5990 * config/i386/i386.c (ix86_function_regparm): Use
5991 opt_for_fn (decl, optimize) instead of optimize.
5993 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5995 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5996 for SYMBOL_REF in large memory model.
5998 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6000 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6002 (cortex-a57): Likewise.
6003 (cortex-a57.cortex-a53): Likewise.
6005 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6006 Kugan Vivekanandarajah <kuganv@linaro.org>
6008 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6010 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6012 2014-02-06 Richard Biener <rguenther@suse.de>
6014 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6015 set_loop_copy and initialize_original_copy_tables.
6017 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6019 * config/aarch64/aarch64-simd.md
6020 (aarch64_ashr_simddi): Change QI to SI.
6022 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6023 Jakub Jelinek <jakub@redhat.com>
6026 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6029 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6031 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6032 CODE_FOR_altivec_vpku[hw]um to
6033 CODE_FOR_altivec_vpku[hw]um_direct.
6034 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6035 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6036 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6037 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6039 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6041 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6042 generation for -maltivec=be.
6043 (altivec_vsumsws): Simplify redundant test.
6045 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6047 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
6048 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
6049 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
6050 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
6051 gen_altivec_vpkuwum.
6052 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
6054 (altivec_vpks<VI_char>ss): Likewise.
6055 (altivec_vpks<VI_char>us): Likewise.
6056 (altivec_vpku<VI_char>us): Likewise.
6057 (altivec_vpku<VI_char>um): Likewise.
6058 (altivec_vpku<VI_char>um_direct): New (copy of
6059 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
6061 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
6062 target is little endian and -maltivec=be is not specified.
6063 (*altivec_vupkhs<VU_char>_direct): New (copy of
6064 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
6065 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
6066 target is little endian and -maltivec=be is not specified.
6067 (*altivec_vupkls<VU_char>_direct): New (copy of
6068 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
6069 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
6070 little endian and -maltivec=be is not specified.
6071 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
6072 little endian and -maltivec=be is not specified.
6074 2014-02-05 Richard Henderson <rth@redhat.com>
6077 * combine-stack-adj.c: Revert r206943.
6078 * sched-int.h (struct deps_desc): Add last_args_size.
6079 * sched-deps.c (init_deps): Initialize it.
6080 (sched_analyze_insn): Add OUTPUT dependencies between insns that
6081 contain REG_ARGS_SIZE notes.
6083 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6085 * lto-cgraph.c (asm_nodes_output): Make global.
6086 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
6087 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
6088 (driver_handle_option): Handle OPT_fwpa.
6090 2014-02-05 Jakub Jelinek <jakub@redhat.com>
6093 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
6094 a comment typo and formatting issue. If odr_hash hasn't been
6095 created, return vNULL and set *completep to false.
6098 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
6099 bb with no successors.
6101 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
6104 * doc/invoke.texi (-march): Clarify documentation for ARM.
6108 2014-02-05 Richard Biener <rguenther@suse.de>
6110 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
6111 when not vectorizing because of too many alias checks.
6112 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
6113 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
6115 2014-02-05 Nick Clifton <nickc@redhat.com>
6117 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
6118 accept extended registers in any mode when compiling for the MN10300.
6120 2014-02-05 Yury Gribov <y.gribov@samsung.com>
6122 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
6123 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
6124 sanitization attributes.
6125 (can_inline_edge_p): Likewise.
6126 (sanitize_attrs_match_for_inline_p): New function.
6128 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6130 * ipa-prop.c (detect_type_change): Shor circuit testing of
6131 type changes on THIS pointer.
6133 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
6136 * config/pa/pa.c (legitimize_tls_address): Return original address
6137 if not passed a SYMBOL_REF rtx.
6138 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
6140 (pa_emit_move_sequence): Simplify TLS source operands.
6141 (pa_legitimate_constant_p): Reject all TLS constants.
6142 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
6143 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
6145 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6147 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
6148 groups when we know they are controlled by LTO.
6149 * varasm.c (default_binds_local_p_1): If object is in other partition,
6150 it will be resolved locally.
6152 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6154 * config/host-linux.c (linux_gt_pch_use_address): Don't
6155 use SSIZE_MAX because it is not always defined.
6157 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6160 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6161 threshold for pseudo splitting.
6162 (update_ebb_live_info): Process call argument hard registers and
6163 hard registers from insn definition too.
6164 (max_small_class_regs_num): New constant.
6165 (inherit_in_ebb): Update live hard regs through EBBs. Update
6166 reloads_num only for small register classes. Don't split for
6169 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6172 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6175 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6177 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6180 2014-02-04 Marek Polacek <polacek@redhat.com>
6182 * gdbinit.in (pel): Define.
6184 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6186 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6189 2014-02-04 Richard Biener <rguenther@suse.de>
6192 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6193 in function context local.
6194 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6195 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6196 similar to LTO_imported_decl_ref.
6198 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6200 PR tree-optimization/60002
6201 * cgraphclones.c (build_function_decl_skip_args): Clear
6204 PR tree-optimization/60023
6205 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6206 false to gsi_replace.
6207 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6208 has been in some EH region and vec_stmt could throw, add
6209 vec_stmt into the same EH region.
6210 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6211 has no lhs, ignore it.
6212 * internal-fn.c (expand_MASK_LOAD): Likewise.
6215 * tree-inline.c (copy_forbidden): Fail for
6216 __attribute__((optimize (0))) functions.
6219 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6220 is set, copy one less argument.
6221 (expand_simd_clones): Don't subtract clone_info->inbranch
6222 from simd_clone_struct_alloc argument.
6224 PR rtl-optimization/57915
6225 * recog.c (simplify_while_replacing): If all unary/binary/relational
6226 operation arguments are constant, attempt to simplify those.
6229 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6230 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6232 2014-02-04 Richard Biener <rguenther@suse.de>
6234 PR tree-optimization/60012
6235 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6236 TBAA disambiguation to all DDRs.
6238 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6241 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6242 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6244 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6247 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6249 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6251 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6252 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6254 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6257 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6258 to figure out targets of polymorphic calls with known decl.
6259 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6260 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6261 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6262 (get_polymorphic_call_info): ... here.
6263 (get_polymorphic_call_info_from_invariant): New function.
6265 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6267 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6268 lookup via vtable pointer; check for type consistency
6269 and turn inconsitent facts into UNREACHABLE.
6270 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6271 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6272 type inconsistent querries; return UNREACHABLE instead.
6274 2014-02-03 Richard Henderson <rth@twiddle.net>
6277 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6278 already processed this node.
6279 (normalize_one_pred_1): Pass along mark_set.
6280 (normalize_one_pred): Create and destroy a pointer_set_t.
6281 (normalize_one_pred_chain): Likewise.
6283 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6285 PR gcov-profile/58602
6286 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6288 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6291 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6292 -fno-devirtualize; try to devirtualize by the knowledge of
6293 virtual table pointer given by aggregate propagation.
6294 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6295 (ipa_print_node_jump_functions): Dump also offset that
6296 is relevant for polymorphic calls.
6297 (determine_known_aggregate_parts): Add arg_type parameter; use it
6298 instead of determining the type from pointer type.
6299 (ipa_compute_jump_functions_for_edge): Update call of
6300 determine_known_aggregate_parts.
6301 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6302 (gimple_get_virt_method_for_binfo): ... here; simplify using
6303 vtable_pointer_value_to_vtable.
6304 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6305 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6306 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6307 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6309 (vtable_pointer_value_to_binfo): ... here.
6310 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6312 2014-02-03 Teresa Johnson <tejohnson@google.com>
6314 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6315 redef of outer loop index variable.
6317 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6321 * doc/extend.texi (Function Attributes): Typo.
6323 2014-02-03 Cong Hou <congh@google.com>
6325 PR tree-optimization/60000
6326 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6327 if the vectorized statement is a store. A store statement can only
6328 appear at the end of pattern statements.
6330 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6332 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6333 (ix86_option_override_internal): Default long double to 64-bit for
6334 32-bit Bionic and to 128-bit for 64-bit Bionic.
6336 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6337 TARGET_LONG_DOUBLE_128 is true.
6338 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6340 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6341 (mlong-double-64): Negate -mlong-double-128.
6342 (mlong-double-128): New option.
6344 * config/i386/i386-c.c (ix86_target_macros): Define
6345 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6347 * doc/invoke.texi: Document -mlong-double-128.
6349 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6351 PR rtl-optimization/60024
6352 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6354 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6356 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6358 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6360 PR rtl-optimization/57662
6361 * sel-sched.c (code_motion_path_driver): Do not mark already not
6362 existing blocks in the visiting bitmap.
6364 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6366 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6367 on the insn being emitted.
6369 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6370 Will Deacon <will.deacon@arm.com>
6372 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6374 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6376 * config/arm/arm-tables.opt: Regenerate.
6378 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6380 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6381 for vector types other than V16QImode.
6382 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6383 define_expand, and call altivec_expand_vec_perm_le when producing
6384 code with little endian element order.
6385 (*altivec_vperm_<mode>_internal): New insn having previous
6386 behavior of altivec_vperm_<mode>.
6387 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6388 altivec_expand_vec_perm_le when producing code with little endian
6390 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6391 behavior of altivec_vperm_<mode>_uns.
6393 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6395 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6396 (altivec_vsumsws): Add handling for -maltivec=be with a little
6398 (altivec_vsumsws_direct): New.
6399 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6400 gen_altivec_vsumsws.
6402 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6404 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6405 vtable_pointer_value_to_binfo): New functions.
6406 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6407 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6409 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6411 * config/nios2/nios2.md (load_got_register): Initialize GOT
6412 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6413 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6415 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6417 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6418 preserverd by passthrough, do not propagate the type.
6420 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6422 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6423 (mips_atomic_assign_expand_fenv): New function.
6424 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6426 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6428 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6429 (__builtin_mips_set_fcsr): Likewise.
6430 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6431 MIPS_USI_FTYPE_VOID.
6432 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6433 (mips16_expand_set_fcsr): Likewise.
6434 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6435 (mips16_set_fcsr_stub): Likewise.
6436 (mips16_get_fcsr_one_only_stub): New class.
6437 (mips16_set_fcsr_one_only_stub): Likewise.
6438 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6439 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6440 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6441 (hard_float): New availability predicate.
6442 (mips_builtins): Add get_fcsr and set_fcsr.
6443 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6444 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6445 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6446 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6447 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6450 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6452 * config/mips/mips.c (mips_one_only_stub): New class.
6453 (mips_need_mips16_rdhwr_p): Replace with...
6454 (mips16_rdhwr_stub): ...this new variable.
6455 (mips16_stub_call_address): New function.
6456 (mips16_rdhwr_one_only_stub): New class.
6457 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6458 (mips_output_mips16_rdhwr): Delete.
6459 (mips_finish_stub): New function.
6460 (mips_code_end): Use it to handle rdhwr stubs.
6462 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6465 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6466 when calculating size of integer atomic types.
6468 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6470 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6472 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6474 PR tree-optimization/60003
6475 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6476 * profile.c (branch_prob): Use gimple_call_builtin_p
6477 to check for BUILT_IN_SETJMP_RECEIVER.
6478 * tree-inline.c (copy_bb): Call notice_special_calls.
6480 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6483 * lra-constraints.c (process_alt_operands): Update reload_sum only
6486 2014-01-31 Richard Henderson <rth@redhat.com>
6489 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6490 until after else_eh is processed.
6492 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6494 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6495 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6496 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6497 in smmintrin.h, remove them.
6498 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6499 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6500 * config/i386/i386.md (ROUND_SAE): Fix value.
6501 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6502 (const48_operand): New.
6503 * config/i386/subst.md (round), (round_expand): Use
6504 const_4_or_8_to_11_operand.
6505 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6507 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6509 * config/i386/constraints.md (Yk): Swap meaning with k.
6510 * config/i386/i386.md (movhi_internal): Change Yk to k.
6511 (movqi_internal): Ditto.
6512 (*k<logic><mode>): Ditto.
6515 (kandn<mode>): Ditto.
6516 (*<code>hi_1): Ditto.
6517 (*<code>qi_1): Ditto.
6518 (kxnor<mode>): Ditto.
6519 (kortestzhi): Ditto.
6520 (kortestchi): Ditto.
6522 (*one_cmplhi2_1): Ditto.
6523 (*one_cmplqi2_1): Ditto.
6524 * config/i386/sse.md (): Change k to Yk.
6525 (avx512f_load<mode>_mask): Ditto.
6526 (avx512f_blendm<mode>): Ditto.
6527 (avx512f_store<mode>_mask): Ditto.
6528 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6529 (avx512f_storedqu<mode>_mask): Ditto.
6530 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6532 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6533 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6534 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6535 (avx512f_maskcmp<mode>3): Ditto.
6536 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6537 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6538 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6539 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6540 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6541 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6542 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6543 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6544 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6545 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6546 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6547 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6548 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6549 (vec_extract_lo_<mode>_maskm): Ditto.
6550 (vec_extract_hi_<mode>_maskm): Ditto.
6551 (avx512f_vternlog<mode>_mask): Ditto.
6552 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6553 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6554 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6555 (avx512f_<code>v8div16qi2_mask): Ditto.
6556 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6557 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6558 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6559 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6560 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6561 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6562 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6563 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6564 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6565 (avx512cd_maskb_vec_dupv8di): Ditto.
6566 (avx512cd_maskw_vec_dupv16si): Ditto.
6567 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6568 (avx512f_vpermi2var<mode>3_mask): Ditto.
6569 (avx512f_vpermi2var<mode>3_mask): Ditto.
6570 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6571 (*avx512f_gathersi<mode>): Ditto.
6572 (*avx512f_gathersi<mode>_2): Ditto.
6573 (*avx512f_gatherdi<mode>): Ditto.
6574 (*avx512f_gatherdi<mode>_2): Ditto.
6575 (*avx512f_scattersi<mode>): Ditto.
6576 (*avx512f_scatterdi<mode>): Ditto.
6577 (avx512f_compress<mode>_mask): Ditto.
6578 (avx512f_compressstore<mode>_mask): Ditto.
6579 (avx512f_expand<mode>_mask): Ditto.
6580 * config/i386/subst.md (mask): Change k to Yk.
6581 (mask_scalar_merge): Ditto.
6584 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6586 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6588 2014-01-31 Richard Biener <rguenther@suse.de>
6591 * builtins.c (fold_builtin_memory_op): Make sure to not
6592 use a floating-point mode or a boolean or enumeral type for
6595 2014-01-30 DJ Delorie <dj@redhat.com>
6597 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6598 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6599 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6600 whenever main() has an epilogue.
6602 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6604 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6605 unused variable "field".
6606 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6607 (vsx_mergeh_<mode>): Likewise.
6608 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6609 (altivec_vmrghh): Likewise.
6610 (altivec_vmrghw): Likewise.
6611 (altivec_vmrglb): Likewise.
6612 (altivec_vmrglh): Likewise.
6613 (altivec_vmrglw): Likewise.
6614 (altivec_vspltb): Add missing uses.
6615 (altivec_vsplth): Likewise.
6616 (altivec_vspltw): Likewise.
6617 (altivec_vspltsf): Likewise.
6619 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6622 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6623 frame related instructions.
6625 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6627 PR rtl-optimization/59959
6628 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6629 any reload of register whose subreg is invalid.
6631 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6633 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6634 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6635 Add missing return type - void.
6637 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6639 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6640 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6641 remove element index adjustment for endian (now handled in vsx.md
6643 (altivec_expand_vec_perm_const): Use
6644 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6645 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6646 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6647 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6648 define_expand and a new define_insn *altivec_vspltb_internal;
6649 adjust for -maltivec=be on a little endian target.
6650 (altivec_vspltb_direct): New.
6651 (altivec_vsplth): Divide into a define_expand and a new
6652 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6653 little endian target.
6654 (altivec_vsplth_direct): New.
6655 (altivec_vspltw): Divide into a define_expand and a new
6656 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6657 little endian target.
6658 (altivec_vspltw_direct): New.
6659 (altivec_vspltsf): Divide into a define_expand and a new
6660 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6661 a little endian target.
6663 2014-01-30 Richard Biener <rguenther@suse.de>
6665 PR tree-optimization/59993
6666 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6667 can propagate form the earlier stmt and avoid the transform
6668 when the intermediate result is needed.
6670 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6672 * README.Portability: Fix typo.
6674 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6676 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6677 comparison_operator with ordered_comparison_operator.
6679 2014-01-30 Nick Clifton <nickc@redhat.com>
6681 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6682 Rename to mn10300_store_multiple_regs.
6683 * config/mn10300/mn10300.c: Likewise.
6684 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6685 store_multiple_regs.
6686 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6687 Call mn10300_store_multiple_regs.
6689 2014-01-30 Nick Clifton <nickc@redhat.com>
6690 DJ Delorie <dj@redhat.com>
6692 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6693 %fp 2 to keep registers after it properly word-aligned.
6694 (rl78_alloc_physical_registers_umul): Handle the case where both
6695 input operands are the same.
6697 2014-01-30 Richard Biener <rguenther@suse.de>
6699 PR tree-optimization/59903
6700 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6703 2014-01-30 Jason Merrill <jason@redhat.com>
6706 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6709 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6711 2014-01-30 Richard Biener <rguenther@suse.de>
6713 PR tree-optimization/59951
6714 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6716 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6719 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6720 SFmode to DFmode case.
6722 2014-01-29 DJ Delorie <dj@redhat.com>
6724 * config/msp430/msp430.opt (-minrt): New.
6725 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6727 (ENDFILE_SPEC): Likewise.
6729 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6731 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6732 (estimate_function_body_sizes): Use it.
6734 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6737 * dwarf2out.c (is_cxx_auto): New.
6738 (is_base_type): Use it.
6739 (gen_type_die_with_usage): Likewise.
6741 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6743 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6744 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6745 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6746 -maltivec=be with LE targets.
6747 (vsx_mergeh_<mode>): Likewise.
6748 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6749 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6750 (altivec_vmrghb): Replace with define_expand and new
6751 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6752 (altivec_vmrghb_direct): New define_insn.
6753 (altivec_vmrghh): Replace with define_expand and new
6754 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6755 (altivec_vmrghh_direct): New define_insn.
6756 (altivec_vmrghw): Replace with define_expand and new
6757 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6758 (altivec_vmrghw_direct): New define_insn.
6759 (*altivec_vmrghsf): Adjust for endianness.
6760 (altivec_vmrglb): Replace with define_expand and new
6761 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6762 (altivec_vmrglb_direct): New define_insn.
6763 (altivec_vmrglh): Replace with define_expand and new
6764 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6765 (altivec_vmrglh_direct): New define_insn.
6766 (altivec_vmrglw): Replace with define_expand and new
6767 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6768 (altivec_vmrglw_direct): New define_insn.
6769 (*altivec_vmrglsf): Adjust for endianness.
6770 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6771 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6772 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6773 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6774 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6775 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6776 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6777 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6779 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6781 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6782 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6785 2014-01-29 Richard Biener <rguenther@suse.de>
6787 PR tree-optimization/58742
6788 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6789 associate_pointerplus_align.
6790 (associate_pointerplus_diff): New function.
6791 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6792 and associate_pointerplus_diff.
6794 2014-01-29 Richard Biener <rguenther@suse.de>
6796 * lto-streamer.h (LTO_major_version): Bump to 3.
6797 (LTO_minor_version): Reset to 0.
6799 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6801 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6802 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6803 (arm_file_start): Generate correct asm header for armv7ve.
6804 * config/arm/bpabi.h: Add multilib support for armv7ve.
6805 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6806 and cortex-a15 to armv7ve.
6807 * config/arm/t-aprofile: Add multilib support for armv7ve.
6808 * doc/invoke.texi: Document -march=armv7ve.
6810 2014-01-29 Richard Biener <rguenther@suse.de>
6812 PR tree-optimization/58742
6813 * tree-ssa-forwprop.c (associate_plusminus): Return true
6814 if we changed sth, defer EH cleanup to ...
6815 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6816 (simplify_mult): New function.
6818 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6821 PR tree-optimization/59920
6822 * tree.c (build_common_builtin_nodes): Remove
6823 __builtin_setjmp_dispatcher initialization.
6824 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6825 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6826 instead of gsi_after_labels + manually skipping debug stmts.
6827 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6828 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6829 * tree-inline.c (copy_edges_for_bb): Remove
6830 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6831 argument. Ignore computed_goto_p stmts. Don't call
6832 make_abnormal_goto_edges. If a call might need abnormal edges
6833 for non-local gotos, see if it already has an edge to
6834 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6835 with true argument, don't do anything then, otherwise add
6836 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6837 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6839 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6840 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6841 (lower_stmt): Don't set data->calls_builtin_setjmp.
6842 (lower_builtin_setjmp): Adjust comment.
6843 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6844 * tree-cfg.c (found_computed_goto): Remove.
6845 (factor_computed_gotos): Remove.
6846 (make_goto_expr_edges): Return bool, true for computed gotos.
6847 Don't call make_abnormal_goto_edges.
6848 (build_gimple_cfg): Don't set found_computed_goto, don't call
6849 factor_computed_gotos.
6850 (computed_goto_p): No longer static.
6851 (make_blocks): Don't set found_computed_goto.
6852 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6853 (make_edges): If make_goto_expr_edges returns true, push bb
6854 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6855 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6856 vector. Record mapping between bbs and OpenMP regions if there
6857 are any, adjust make_gimple_omp_edges caller. Call
6858 handle_abnormal_edges.
6859 (make_abnormal_goto_edges): Remove.
6860 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6861 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6862 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6863 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6864 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6865 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6866 filling *region also set *region_idx to (*region)->entry->index.
6869 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6870 For REGs set ORIGINAL_REGNO.
6872 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6874 * doc/md.texi: Mention that a target shouldn't implement
6875 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6878 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6880 PR tree-optimization/59594
6881 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6882 a copy of the datarefs vector rather than the vector itself.
6884 2014-01-28 Jason Merrill <jason@redhat.com>
6887 * dwarf2out.c (auto_die): New static.
6888 (gen_type_die_with_usage): Handle C++1y 'auto'.
6889 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6892 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6895 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6896 (SPEC_X32): Likewise.
6897 (SPEC_64): Likewise.
6898 * config/i386/i386.c (ix86_option_override_internal): Turn off
6899 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6901 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6902 * config/i386/i386.h (TARGET_16BIT): New macro.
6903 (TARGET_16BIT_P): Likewise.
6904 * config/i386/i386.opt: Add m16.
6905 * doc/invoke.texi: Document -m16.
6907 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6909 PR preprocessor/59935
6910 * input.c (location_get_source_line): Bail out on when line number
6911 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6913 2014-01-28 Richard Biener <rguenther@suse.de>
6915 PR tree-optimization/58742
6916 * tree-ssa-forwprop.c (associate_plusminus): Handle
6917 pointer subtraction of the form (T)(P + A) - (T)P.
6919 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6921 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6924 2014-01-28 Richard Biener <rguenther@suse.de>
6927 2014-01-28 Richard Biener <rguenther@suse.de>
6929 PR rtl-optimization/45364
6930 PR rtl-optimization/59890
6931 * var-tracking.c (local_get_addr_clear_given_value): Handle
6932 already cleared slot.
6933 (val_reset): Handle not allocated local_get_addr_cache.
6934 (vt_find_locations): Use post-order on the inverted CFG.
6936 2014-01-28 Richard Biener <rguenther@suse.de>
6938 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6940 2014-01-28 Richard Biener <rguenther@suse.de>
6942 PR rtl-optimization/45364
6943 PR rtl-optimization/59890
6944 * var-tracking.c (local_get_addr_clear_given_value): Handle
6945 already cleared slot.
6946 (val_reset): Handle not allocated local_get_addr_cache.
6947 (vt_find_locations): Use post-order on the inverted CFG.
6949 2014-01-28 Alan Modra <amodra@gmail.com>
6951 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6952 * configure.ac <recursive call for build != host>: Define
6953 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6954 and LD_FOR_BUILD too.
6955 * configure: Regenerate.
6957 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6959 * config/i386/i386.c (get_builtin_code_for_version): Separate
6960 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6961 Broadwell from Haswell.
6963 2014-01-27 Steve Ellcey <sellcey@mips.com>
6965 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6966 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6967 * config/mips/mips.c (mips_option_override): Change setting
6969 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6970 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6971 Change from Mask to Var.
6973 2014-01-27 Jeff Law <law@redhat.com>
6975 * ipa-inline.c (inline_small_functions): Fix typo.
6977 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6979 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6980 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6981 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6982 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6983 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6984 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6985 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6986 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6987 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6988 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6989 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6990 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6991 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6992 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6993 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6994 (_mm512_storeu_epi64): Ditto.
6995 (_mm512_cmpge_epi32_mask): Ditto.
6996 (_mm512_cmpge_epu32_mask): Ditto.
6997 (_mm512_cmpge_epi64_mask): Ditto.
6998 (_mm512_cmpge_epu64_mask): Ditto.
6999 (_mm512_cmple_epi32_mask): Ditto.
7000 (_mm512_cmple_epu32_mask): Ditto.
7001 (_mm512_cmple_epi64_mask): Ditto.
7002 (_mm512_cmple_epu64_mask): Ditto.
7003 (_mm512_cmplt_epi32_mask): Ditto.
7004 (_mm512_cmplt_epu32_mask): Ditto.
7005 (_mm512_cmplt_epi64_mask): Ditto.
7006 (_mm512_cmplt_epu64_mask): Ditto.
7007 (_mm512_cmpneq_epi32_mask): Ditto.
7008 (_mm512_cmpneq_epu32_mask): Ditto.
7009 (_mm512_cmpneq_epi64_mask): Ditto.
7010 (_mm512_cmpneq_epu64_mask): Ditto.
7011 (_mm512_expand_pd): Ditto.
7012 (_mm512_expand_ps): Ditto.
7013 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7014 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7015 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7016 * config/i386/i386.c (ix86_builtins): Add
7017 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7018 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7019 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7020 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7021 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7022 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7023 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7024 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7025 IX86_BUILTIN_PMOVUSQW512_MEM.
7026 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7027 __builtin_ia32_pmovsqd512mem_mask,
7028 __builtin_ia32_pmovqd512mem_mask,
7029 __builtin_ia32_pmovusqw512mem_mask,
7030 __builtin_ia32_pmovsqw512mem_mask,
7031 __builtin_ia32_pmovqw512mem_mask,
7032 __builtin_ia32_pmovusdw512mem_mask,
7033 __builtin_ia32_pmovsdw512mem_mask,
7034 __builtin_ia32_pmovdw512mem_mask,
7035 __builtin_ia32_pmovqb512mem_mask,
7036 __builtin_ia32_pmovusqb512mem_mask,
7037 __builtin_ia32_pmovsqb512mem_mask,
7038 __builtin_ia32_pmovusdb512mem_mask,
7039 __builtin_ia32_pmovsdb512mem_mask,
7040 __builtin_ia32_pmovdb512mem_mask.
7041 (bdesc_args): Add __builtin_ia32_expanddf512,
7042 __builtin_ia32_expandsf512.
7043 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
7044 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
7045 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
7046 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
7047 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
7048 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
7049 (avx512f_<code>v8div16qi2_mask_store): This.
7050 (avx512f_expand<mode>): New.
7052 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
7054 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
7056 (_mm512_mask_prefetch_i64gather_pd): Ditto.
7057 (_mm512_prefetch_i32scatter_pd): Ditto.
7058 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
7059 (_mm512_prefetch_i64scatter_pd): Ditto.
7060 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
7061 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
7062 (_mm512_mask_prefetch_i64gather_ps): Ditto.
7063 (_mm512_prefetch_i32scatter_ps): Ditto.
7064 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
7065 (_mm512_prefetch_i64scatter_ps): Ditto.
7066 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
7067 * config/i386/i386-builtin-types.def: Define
7068 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
7069 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
7070 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
7071 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
7072 IX86_BUILTIN_SCATTERPFQPD.
7073 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
7074 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
7075 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
7076 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
7077 __builtin_ia32_scatterpfqps.
7078 (ix86_expand_builtin): Expand new built-ins.
7079 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
7080 fix memory access data type.
7081 (*avx512pf_gatherpf<mode>_mask): Ditto.
7082 (*avx512pf_gatherpf<mode>): Ditto.
7083 (avx512pf_scatterpf<mode>): Ditto.
7084 (*avx512pf_scatterpf<mode>_mask): Ditto.
7085 (*avx512pf_scatterpf<mode>): Ditto.
7086 (GATHER_SCATTER_SF_MEM_MODE): New.
7087 (avx512pf_gatherpf<mode>df): Ditto.
7088 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7089 (*avx512pf_scatterpf<mode>df): Ditto.
7091 2014-01-27 Jakub Jelinek <jakub@redhat.com>
7094 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
7095 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
7098 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7100 * common/config/arm/arm-common.c
7101 (arm_rewrite_mcpu): Handle multiple names.
7103 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7105 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7107 * gimple-builder.h (create_gimple_tmp): Delete.
7109 2014-01-27 Christian Bruel <christian.bruel@st.com>
7111 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
7114 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
7116 * config/pa/pa.md (call): Generate indirect long calls to non-local
7117 functions when outputing 32-bit code.
7118 (call_value): Likewise except for special call to buggy powf function.
7120 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
7121 portable runtime and PIC indirect calls.
7122 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
7123 and PIC call sequences. Use ldo instead of blr to set return register
7124 in PIC call sequence.
7126 2014-01-25 Walter Lee <walt@tilera.com>
7128 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
7129 avoid clobbering a live register.
7131 2014-01-25 Walter Lee <walt@tilera.com>
7133 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
7134 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
7135 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
7136 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
7138 2014-01-25 Walter Lee <walt@tilera.com>
7140 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
7141 arguments on even registers.
7142 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
7144 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
7145 (BIGGEST_ALIGNMENT): Ditto.
7146 (BIGGEST_FIELD_ALIGNMENT): Ditto.
7148 2014-01-25 Walter Lee <walt@tilera.com>
7150 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
7151 insns before bundling.
7152 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
7154 2014-01-25 Walter Lee <walt@tilera.com>
7156 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7157 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7158 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7160 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7162 * config/mips/constraints.md (kl): Delete.
7163 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7164 define expands, using...
7165 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7166 instructions for MIPS16.
7167 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7168 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7170 2014-01-25 Walter Lee <walt@tilera.com>
7172 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7176 2014-01-25 Walter Lee <walt@tilera.com>
7178 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7179 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7181 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7183 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7186 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7188 * print-rtl.c (in_call_function_usage): New var.
7189 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7190 EXPR_LIST mode as mode and not as reg note name.
7193 * cfgloopmanip.c (copy_loop_info): If
7194 loop->warned_aggressive_loop_optimizations, make sure
7195 the flag is set in target loop too.
7197 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7199 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7201 * builtins.def: Likewise.
7202 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7203 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7204 * ira.c (ira_setup_eliminable_regset): Likewise.
7205 * omp-low.c (gate_expand_omp): Likewise.
7206 (execute_lower_omp): Likewise.
7207 (diagnose_sb_0): Likewise.
7208 (gate_diagnose_omp_blocks): Likewise.
7209 (simd_clone_clauses_extract): Likewise.
7212 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7214 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7215 correction for little endian...
7216 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7219 2014-01-24 Jeff Law <law@redhat.com>
7221 PR tree-optimization/59919
7222 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7223 for non-returning calls.
7225 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7227 * common/config/aarch64/aarch64-common.c
7228 (aarch64_rewrite_mcpu): Handle multiple names.
7229 * config/aarch64/aarch64.h
7230 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7232 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7234 * input.c (add_file_to_cache_tab): Handle the case where fopen
7237 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7240 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7241 from push operand if code of push isn't PRE_DEC.
7243 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7246 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7247 -mquad-memory-atomic. Update -mquad-memory documentation to say
7248 it is only used for non-atomic loads/stores.
7250 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7251 -mquad-memory or -mquad-memory-atomic switches.
7253 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7254 -mquad-memory-atomic to ISA 2.07 support.
7256 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7257 to separate support of normal quad word memory operations (ldq, stq)
7258 from the atomic quad word memory operations.
7260 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7261 support to separate non-atomic quad word operations from atomic
7262 quad word operations. Disable non-atomic quad word operations in
7263 little endian mode so that we don't have to swap words after the
7264 load and before the store.
7265 (quad_load_store_p): Add comment about atomic quad word support.
7266 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7267 options printed with -mdebug=reg.
7269 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7270 -mquad-memory-atomic as the test for whether we have quad word
7271 atomic instructions.
7272 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7273 or -mp8-vector are used, allow byte/half-word atomic operations.
7275 * config/rs6000/sync.md (load_lockedti): Insure that the address
7276 is a proper indexed or indirect address for the lqarx instruction.
7277 On little endian systems, swap the hi/lo registers after the lqarx
7279 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7280 insure the address is valid for the lqarx instruction.
7281 (store_conditionalti): Insure that the address is a proper indexed
7282 or indirect address for the stqcrx. instruction. On little endian
7283 systems, swap the hi/lo registers before doing the stqcrx.
7285 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7286 insure the address is valid for the stqcrx. instruction.
7288 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7289 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7290 type of quad memory support is available.
7292 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7295 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7296 there is a danger of looping.
7298 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7300 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7301 force flag_ira_loop_pressure if set via command line.
7303 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7305 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7306 (ashr_simd): New builtin handling DI mode.
7307 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7308 (aarch64_sshr_simddi): New match pattern.
7309 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7310 (vshrd_n_s64): Likewise.
7311 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7313 2014-01-23 Nick Clifton <nickc@redhat.com>
7315 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7316 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7317 favour of mcu specific scripts.
7318 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7321 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7322 Alex Velenko <Alex.Velenko@arm.com>
7324 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7325 (vaddv_s16): Likewise.
7326 (vaddv_s32): Likewise.
7327 (vaddv_u8): Likewise.
7328 (vaddv_u16): Likewise.
7329 (vaddv_u32): Likewise.
7330 (vaddvq_s8): Likewise.
7331 (vaddvq_s16): Likewise.
7332 (vaddvq_s32): Likewise.
7333 (vaddvq_s64): Likewise.
7334 (vaddvq_u8): Likewise.
7335 (vaddvq_u16): Likewise.
7336 (vaddvq_u32): Likewise.
7337 (vaddvq_u64): Likewise.
7338 (vaddv_f32): Likewise.
7339 (vaddvq_f32): Likewise.
7340 (vaddvq_f64): Likewise.
7341 (vmaxv_f32): Likewise.
7342 (vmaxv_s8): Likewise.
7343 (vmaxv_s16): Likewise.
7344 (vmaxv_s32): Likewise.
7345 (vmaxv_u8): Likewise.
7346 (vmaxv_u16): Likewise.
7347 (vmaxv_u32): Likewise.
7348 (vmaxvq_f32): Likewise.
7349 (vmaxvq_f64): Likewise.
7350 (vmaxvq_s8): Likewise.
7351 (vmaxvq_s16): Likewise.
7352 (vmaxvq_s32): Likewise.
7353 (vmaxvq_u8): Likewise.
7354 (vmaxvq_u16): Likewise.
7355 (vmaxvq_u32): Likewise.
7356 (vmaxnmv_f32): Likewise.
7357 (vmaxnmvq_f32): Likewise.
7358 (vmaxnmvq_f64): Likewise.
7359 (vminv_f32): Likewise.
7360 (vminv_s8): Likewise.
7361 (vminv_s16): Likewise.
7362 (vminv_s32): Likewise.
7363 (vminv_u8): Likewise.
7364 (vminv_u16): Likewise.
7365 (vminv_u32): Likewise.
7366 (vminvq_f32): Likewise.
7367 (vminvq_f64): Likewise.
7368 (vminvq_s8): Likewise.
7369 (vminvq_s16): Likewise.
7370 (vminvq_s32): Likewise.
7371 (vminvq_u8): Likewise.
7372 (vminvq_u16): Likewise.
7373 (vminvq_u32): Likewise.
7374 (vminnmv_f32): Likewise.
7375 (vminnmvq_f32): Likewise.
7376 (vminnmvq_f64): Likewise.
7378 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7380 * config/aarch64/aarch64-simd.md
7381 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7382 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7383 (*aarch64_mul3_elt<mode>): Likewise.
7384 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7385 (*aarch64_mul3_elt_to_64v2df): Likewise.
7386 (*aarch64_mla_elt<mode>): Likewise.
7387 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7388 (*aarch64_mls_elt<mode>): Likewise.
7389 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7390 (*aarch64_fma4_elt<mode>): Likewise.
7391 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7392 (*aarch64_fma4_elt_to_64v2df): Likewise.
7393 (*aarch64_fnma4_elt<mode>): Likewise.
7394 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7395 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7396 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7397 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7398 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7399 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7400 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7401 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7402 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7404 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7406 * config/aarch64/aarch64-simd.md
7407 (aarch64_be_checked_get_lane<mode>): New define_expand.
7408 * config/aarch64/aarch64-simd-builtins.def
7409 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7410 New builtin definition.
7411 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7412 Use new safe be builtin.
7414 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7416 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7418 (aarch64_be_st1<mode>): Likewise.
7419 (aarch_ld1<VALL:mode>): Define_expand modified.
7420 (aarch_st1<VALL:mode>): Likewise.
7421 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7422 (UNSPEC_ST1): Likewise.
7424 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7426 * config/microblaze/microblaze.md: Add trap insn and attribute
7428 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7430 PR preprocessor/58580
7431 * input.h (location_get_source_line): Take an additional line_size
7433 (void diagnostics_file_cache_fini): Declare new function.
7434 * input.c (struct fcache): New type.
7435 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7436 New static constants.
7437 (diagnostic_file_cache_init, total_lines_num)
7438 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7439 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7440 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7441 (get_next_line, read_next_line, goto_next_line, read_line_num):
7442 New static function definitions.
7443 (diagnostic_file_cache_fini): New function.
7444 (location_get_source_line): Take an additional output line_len
7445 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7447 * diagnostic.c (diagnostic_finish): Call
7448 diagnostic_file_cache_fini.
7449 (adjust_line): Take an additional input parameter for the length
7450 of the line, rather than calculating it with strlen.
7451 (diagnostic_show_locus): Adjust the use of
7452 location_get_source_line and adjust_line with respect to their new
7453 signature. While displaying a line now, do not stop at the first
7454 null byte. Rather, display the zero byte as a space and keep
7455 going until we reach the size of the line.
7456 * Makefile.in: Add vec.o to OBJS-libcommon
7458 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7459 Ilya Tocar <ilya.tocar@intel.com>
7461 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7462 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7463 (__builtin_ia32_kmov16): Ditto.
7464 * config/i386/i386.md (UNSPEC_KMOV): New.
7467 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7469 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7470 (_mm512_storeu_si512): Ditto.
7472 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7475 * rtl.h (get_referenced_operands): Declare.
7476 * recog.c (get_referenced_operands): New function.
7477 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7478 operands have been referenced when recording LO_SUM references.
7480 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7482 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7484 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7486 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7487 Enable for generic and recent AMD targets.
7489 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7491 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7492 ARG_SIZE note when adjustment was eliminated.
7494 2014-01-22 Jeff Law <law@redhat.com>
7496 PR tree-optimization/59597
7497 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7498 in file. Accept new argument REGISTERING and use it to modify
7499 dump output appropriately.
7500 (register_jump_thread): Corresponding changes.
7501 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7502 thread paths involving joiner blocks. Add code to dump cancelled
7503 jump threading paths.
7505 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7507 PR rtl-optimization/59477
7508 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7509 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7511 2014-01-22 Tom Tromey <tromey@redhat.com>
7513 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7515 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7517 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7519 PR rtl-optimization/59896
7520 * lra-constraints.c (process_alt_operands): Check unused note for
7521 matched operands of insn with no output reloads.
7523 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7525 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7526 (mips_move_from_gpr_cost): Likewise.
7528 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7530 PR rtl-optimization/59858
7531 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7532 ira_class_hard_regs_num.
7533 (process_alt_operands): Increase reject for dying matched operand.
7535 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7538 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7539 smaller than size, perform several stores or loads and stores
7540 at dst + count - size to store or copy all of size bytes, rather
7541 than just last modesize bytes.
7543 2014-01-20 DJ Delorie <dj@redhat.com>
7545 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7546 that CLOBBERs are REGs before propogating their values.
7548 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7551 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7552 (cgraph_inline_failed_type): New function.
7553 * cgraph.h (DEFCIFCODE): Add type.
7554 (cgraph_inline_failed_type_t): New enum.
7555 (cgraph_inline_failed_type): New prototype.
7556 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7557 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7558 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7559 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7560 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7561 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7562 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7563 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7564 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7565 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7566 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7567 OPTIMIZATION_MISMATCH.
7568 * tree-inline.c (expand_call_inline): Emit errors during
7569 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7571 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7574 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7575 mode attribute in insn output.
7577 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7579 * output.h (output_constant): Delete.
7580 * varasm.c (output_constant): Make private.
7582 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7584 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7586 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7589 * tree.h (fold_builtin_strcat): New prototype.
7590 * builtins.c (fold_builtin_strcat): No longer static. Add len
7591 argument, if non-NULL, don't call c_strlen. Optimize
7592 directly into __builtin_memcpy instead of __builtin_strcpy.
7593 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7594 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7596 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7598 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7599 for SImode_address_operand operands, having only a REG argument.
7601 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7603 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7604 loader name using mbig-endian.
7605 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7607 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7609 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7613 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7615 * config/aarch64/aarch64-protos.h
7616 (aarch64_cannot_change_mode_class_ptr): Declare.
7617 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7618 aarch64_cannot_change_mode_class_ptr): New.
7619 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7620 backend hook aarch64_cannot_change_mode_class.
7622 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7624 * common/config/aarch64/aarch64-common.c
7625 (aarch64_handle_option): Don't handle any option order logic here.
7626 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7627 selected_cpu, warn on architecture version mismatch.
7628 (aarch64_override_options): Fix parsing order for option strings.
7630 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7631 Iain Sandoe <iain@codesourcery.com>
7634 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7635 warning. Amend comment to reflect current functionality.
7637 2014-01-20 Richard Biener <rguenther@suse.de>
7640 * builtins.c (fold_builtin_strcat): Remove case better handled
7641 by tree-ssa-strlen.c.
7643 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7645 * config/aarch64/aarch64.opt
7646 (mcpu, march, mtune): Make case-insensitive.
7648 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7651 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7652 if operands[1] is a REG or ZERO_EXTEND of a REG.
7654 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7656 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7658 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7660 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7661 long non-pic millicode calls.
7663 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7665 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7667 2014-01-19 Kito Cheng <kito@0xlab.org>
7669 * builtins.c (expand_movstr): Check movstr expand done or fail.
7671 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7672 H.J. Lu <hongjiu.lu@intel.com>
7675 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7676 to DImode for zero-extended addresses.
7678 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7680 PR rtl-optimization/57763
7681 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7682 on the new indirect jump_insn and increment LABEL_NUSES (label).
7684 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7688 * config.gcc (x86_archs): New variable.
7689 (x86_64_archs): Likewise.
7690 (x86_cpus): Likewise.
7691 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7692 --with-arch/--with-cpu= options.
7693 Support --with-arch=/--with-cpu={nehalem,westmere,
7694 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7696 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7698 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7699 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7701 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7703 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7705 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7708 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7709 clear cpp_get_options (parse_in)->warn_unused_macros for
7710 ix86_target_macros_internal with cpp_define.
7712 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7714 * jump.c (delete_related_insns): Keep (use (insn))s.
7715 * reorg.c (redundant_insn): Check for barriers too.
7717 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7719 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7721 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7723 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7724 call to $$dyncall when TARGET_LONG_CALLS is true.
7726 2014-01-17 Jeff Law <law@redhat.com>
7728 * ree.c (combine_set_extension): Temporarily disable test for
7729 changing number of hard registers.
7731 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7734 * ipa-inline-analysis.c (inline_free_summary):
7735 Do not free summary of aliases.
7737 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7740 * gimplify.c (gimplify_expr): Use create_tmp_var
7741 instead of create_tmp_var_raw. If cond doesn't have
7742 integral type, don't add the IFN_ANNOTATE builtin at all.
7744 2014-01-17 Martin Jambor <mjambor@suse.cz>
7747 * ipa-cp.c (prev_edge_clone): New variable.
7748 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7749 Also resize prev_edge_clone vector.
7750 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7751 (ipcp_edge_removal_hook): New function.
7752 (ipcp_driver): Register ipcp_edge_removal_hook.
7754 2014-01-17 Andrew Pinski <apinski@cavium.com>
7755 Steve Ellcey <sellcey@mips.com>
7758 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7761 2014-01-17 Jeff Law <law@redhat.com>
7764 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7765 so that pass_ccp runs first.
7767 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7769 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7770 (ix86_adjust_cost): Use !TARGET_XXX.
7771 (do_reorder_for_imul): Likewise.
7772 (swap_top_of_ready_list): Likewise.
7773 (ix86_sched_reorder): Likewise.
7775 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7777 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7778 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7779 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7780 (intel_memset): New. Duplicate slm_memset.
7781 (intel_cost): New. Duplicate slm_cost.
7782 (m_INTEL): New macro.
7783 (processor_target_table): Add "intel".
7784 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7785 with PROCESSOR_INTEL for "intel".
7786 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7787 PROCESSOR_SILVERMONT.
7788 (ix86_issue_rate): Likewise.
7789 (ix86_adjust_cost): Likewise.
7790 (ia32_multipass_dfa_lookahead): Likewise.
7791 (swap_top_of_ready_list): Likewise.
7792 (ix86_sched_reorder): Likewise.
7793 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7794 instead of TARGET_OPT_AGU.
7795 * config/i386/i386.h (TARGET_INTEL): New.
7796 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7797 (processor_type): Add PROCESSOR_INTEL.
7798 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7799 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7801 2014-01-17 Marek Polacek <polacek@redhat.com>
7804 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7807 2014-01-17 Richard Biener <rguenther@suse.de>
7809 PR tree-optimization/46590
7810 * opts.c (default_options_table): Add entries for
7811 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7812 all enabled at -O1 but not for -Og.
7813 * common.opt (fbranch-count-reg): Remove Init(1).
7814 (fmove-loop-invariants): Likewise.
7815 (ftree-pta): Likewise.
7817 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7819 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7820 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7821 decls to at least the GCC 4.8 used alignments.
7824 * tree-nested.c (convert_nonlocal_reference_stmt,
7825 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7826 of GIMPLE_BIND stmts, adjust associated decls.
7828 2014-01-17 Richard Biener <rguenther@suse.de>
7830 PR tree-optimization/46590
7831 * vec.h (vec<>::bseach): New member function implementing
7832 binary search according to C89 bsearch.
7833 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7834 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7835 bitmap pointer again. Make accesses_in_loop a flat array.
7836 (mem_ref_obstack): New global.
7837 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7838 (mark_ref_stored): Likewise.
7839 (ref_indep_loop_p_2): Likewise.
7840 (set_ref_stored_in_loop): New helper function.
7841 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7842 (memref_free): Adjust.
7843 (record_mem_ref_loc): Simplify.
7844 (gather_mem_refs_stmt): Adjust.
7845 (sort_locs_in_loop_postorder_cmp): New function.
7846 (analyze_memory_references): Sort accesses_in_loop after
7847 loop postorder number.
7848 (find_ref_loc_in_loop_cmp): New function.
7849 (for_all_locs_in_loop): Find relevant cluster of locs in
7850 accesses_in_loop and iterate without recursion.
7851 (execute_sm): Avoid uninit warning.
7852 (struct ref_always_accessed): Simplify.
7853 (ref_always_accessed::operator ()): Likewise.
7854 (ref_always_accessed_p): Likewise.
7855 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7856 loop postorder numbers here.
7857 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7860 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7863 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7864 on decls for which assemble_alias has been called.
7866 2014-01-17 Nick Clifton <nickc@redhat.com>
7868 * config/msp430/msp430.opt: (mcpu): New option.
7869 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7870 (msp430_option_override): Parse target_cpu. If the MCU name
7871 matches a generic string, clear target_mcu.
7872 (msp430_attr): Allow numeric interrupt values up to 63.
7873 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7874 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7876 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7878 * config/msp430/msp430.md (popm): Use %J rather than %I.
7879 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7880 (addhi_cy_i): Use immediate_operand for operand 2.
7881 * doc/invoke.texi: Document -mcpu option.
7883 2014-01-17 Richard Biener <rguenther@suse.de>
7885 PR rtl-optimization/38518
7886 * df.h (df_analyze_loop): Declare.
7887 * df-core.c: Include cfgloop.h.
7888 (df_analyze_1): Split out main part of df_analyze.
7889 (df_analyze): Adjust.
7890 (loop_inverted_post_order_compute): New function.
7891 (loop_post_order_compute): Likewise.
7892 (df_analyze_loop): New function avoiding whole-function
7894 * loop-invariant.c (find_defs): Use df_analyze_loop.
7895 (find_invariants): Adjust.
7896 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7898 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7900 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7901 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7903 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7905 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7906 traversal when removing references.
7908 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7911 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7913 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7916 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7917 pushing a reload for an autoinc when we had previously reloaded an
7918 inner part of the address.
7920 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7922 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7924 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7925 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7926 when not giving up or versioning for alias only because of
7928 (vect_analyze_data_ref_dependences): Set to true.
7929 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7931 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7932 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7936 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7939 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7940 operand 0 predicate for gathers, use a new pseudo as subtarget.
7942 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7945 * lra-constraints.c (process_alt_operands): Add printing debug info.
7946 Check absence of input/output reloads for matched operands too.
7948 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7950 PR rtl-optimization/59835
7951 * ira.c (ira_init_register_move_cost): Increase cost for
7954 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7956 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7958 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7961 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7962 non-register objects. Use gen_(high/low)part more consistently.
7965 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7968 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7969 endian support, remove tests for WORDS_BIG_ENDIAN.
7970 (p8_mfvsrd_3_<mode>): Likewise.
7971 (reload_gpr_from_vsx<mode>): Likewise.
7972 (reload_gpr_from_vsxsf): Likewise.
7973 (p8_mfvsrd_4_disf): Likewise.
7975 2014-01-16 Richard Biener <rguenther@suse.de>
7977 PR rtl-optimization/46590
7978 * lcm.c (compute_antinout_edge): Use postorder iteration.
7979 (compute_laterin): Use inverted postorder iteration.
7981 2014-01-16 Nick Clifton <nickc@redhat.com>
7984 * varasm.c (output_constant): Return the number of bytes actually
7986 (output_constructor_array_range): Update the field size with the
7987 number of bytes emitted by output_constant.
7988 (output_constructor_regular_field): Likewise. Also do not
7989 complain if the total number of bytes emitted is now greater
7990 than the expected fieldpos.
7991 * output.h (output_constant): Update prototype and descriptive comment.
7993 2014-01-16 Marek Polacek <polacek@redhat.com>
7996 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7997 it is error_mark_node.
7999 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8001 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8002 VALID_AVX256_REG_OR_OI_MODE.
8004 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8006 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8007 current procedure should be profiled.
8009 2014-01-15 Andrew Pinski <apinski@cavium.com>
8011 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8012 of moving from/to the STACK_REG register class.
8014 2014-01-15 Richard Henderson <rth@redhat.com>
8017 * reginfo.c (global_regs_decl): Globalize.
8018 * rtl.h (global_regs_decl): Declare.
8019 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8020 reserved via global_regs.
8022 2014-01-15 Teresa Johnson <tejohnson@google.com>
8024 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8026 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8028 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8029 and vmulosh rather than call gen_vec_widen_smult_*.
8030 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8031 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8032 (vec_widen_smult_even_v16qi): Likewise.
8033 (vec_widen_umult_even_v8hi): Likewise.
8034 (vec_widen_smult_even_v8hi): Likewise.
8035 (vec_widen_umult_odd_v16qi): Likewise.
8036 (vec_widen_smult_odd_v16qi): Likewise.
8037 (vec_widen_umult_odd_v8hi): Likewise.
8038 (vec_widen_smult_odd_v8hi): Likewise.
8039 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8040 vmuloub rather than call gen_vec_widen_umult_*.
8041 (vec_widen_umult_lo_v16qi): Likewise.
8042 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
8043 vmulosb rather than call gen_vec_widen_smult_*.
8044 (vec_widen_smult_lo_v16qi): Likewise.
8045 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
8046 rather than call gen_vec_widen_umult_*.
8047 (vec_widen_umult_lo_v8hi): Likewise.
8048 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
8049 rather than call gen_vec_widen_smult_*.
8050 (vec_widen_smult_lo_v8hi): Likewise.
8052 2014-01-15 Jeff Law <law@redhat.com>
8054 PR tree-optimization/59747
8055 * ree.c (find_and_remove_re): Properly handle case where a second
8056 eliminated extension requires widening a copy created for elimination
8057 of a prior extension.
8058 (combine_set_extension): Ensure that the number of hard regs needed
8059 for a destination register does not change when we widen it.
8061 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
8063 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
8064 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
8065 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
8066 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
8067 (avr-*-rtems*): Likewise.
8068 (bfin*-rtems*): Likewise.
8069 (moxie-*-rtems*): Likewise.
8070 (h8300-*-rtems*): Likewise.
8071 (i[34567]86-*-rtems*): Likewise.
8072 (lm32-*-rtems*): Likewise.
8073 (m32r-*-rtems*): Likewise.
8074 (m68k-*-rtems*): Likewise.
8075 (microblaze*-*-rtems*): Likewise.
8076 (mips*-*-rtems*): Likewise.
8077 (powerpc-*-rtems*): Likewise.
8078 (sh-*-rtems*): Likewise.
8079 (sparc-*-rtems*): Likewise.
8080 (sparc64-*-rtems*): Likewise.
8081 (v850-*-rtems*): Likewise.
8082 (m32c-*-rtems*): Likewise.
8084 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
8086 PR rtl-optimization/59511
8087 * ira.c (ira_init_register_move_cost): Use memory costs for some
8088 cases of register move cost calculations.
8089 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
8090 instead of BB frequency.
8091 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
8092 * lra-assigns.c (find_hard_regno_for): Ditto.
8094 2014-01-15 Richard Biener <rguenther@suse.de>
8096 PR tree-optimization/59822
8097 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
8098 (vectorizable_load): Use it to hoist defs of uses of invariant
8099 loads out of the loop.
8101 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8102 Kugan Vivekanandarajah <kuganv@linaro.org>
8105 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
8108 2014-01-15 Richard Biener <rguenther@suse.de>
8110 PR rtl-optimization/59802
8111 * lcm.c (compute_available): Use inverted postorder to seed
8112 the initial worklist.
8114 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8117 * config/s390/s390.c (s390_preferred_reload_class): Don't return
8118 ADDR_REGS for invalid symrefs in non-PIC code.
8120 2014-01-15 Jakub Jelinek <jakub@redhat.com>
8123 * builtins.c (determine_block_size): Initialize *probable_max_size
8124 even if len_rtx is CONST_INT.
8126 2014-01-14 Andrew Pinski <apinski@cavium.com>
8128 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
8129 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
8130 (cortexa53_tunings): Likewise.
8131 (aarch64_sched_issue_rate): New function.
8132 (TARGET_SCHED_ISSUE_RATE): Define.
8134 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8136 * ira-costs.c (find_costs_and_classes): Add missed
8137 ira_init_register_move_cost_if_necessary.
8139 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8142 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
8144 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
8147 * config/i386/i386.c (type_natural_mode): Add a bool parameter
8148 to indicate if type is used for function return value. Warn ABI
8149 change if the vector mode isn't available for function return value.
8150 (ix86_function_arg_advance): Pass false to type_natural_mode.
8151 (ix86_function_arg): Likewise.
8152 (ix86_gimplify_va_arg): Likewise.
8153 (function_arg_32): Don't warn ABI change.
8154 (ix86_function_value): Pass true to type_natural_mode.
8155 (ix86_return_in_memory): Likewise.
8156 (ix86_struct_value_rtx): Removed.
8157 (TARGET_STRUCT_VALUE_RTX): Likewise.
8159 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8161 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8162 converting a conditional jump into a conditional return.
8164 2014-01-14 Richard Biener <rguenther@suse.de>
8166 PR tree-optimization/58921
8167 PR tree-optimization/59006
8168 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8169 hoisting invariant stmts.
8170 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8171 invariant loads on the preheader edge if possible.
8173 2014-01-14 Joey Ye <joey.ye@arm.com>
8175 * doc/plugin.texi (Building GCC plugins): Update to C++.
8177 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8179 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8180 (_mm_rcp28_round_ss): Ditto.
8181 (_mm_rsqrt28_round_sd): Ditto.
8182 (_mm_rsqrt28_round_ss): Ditto.
8183 (_mm_rcp28_sd): Ditto.
8184 (_mm_rcp28_ss): Ditto.
8185 (_mm_rsqrt28_sd): Ditto.
8186 (_mm_rsqrt28_ss): Ditto.
8187 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8188 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8189 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8190 (IX86_BUILTIN_RCP28SD): Ditto.
8191 (IX86_BUILTIN_RCP28SS): Ditto.
8192 (IX86_BUILTIN_RSQRT28SD): Ditto.
8193 (IX86_BUILTIN_RSQRT28SS): Ditto.
8194 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8195 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8196 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8197 (ix86_expand_special_args_builtin): Expand new FTYPE.
8198 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8199 (srcp14<mode>): Make insn unary.
8200 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8201 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8202 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8203 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8204 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8205 Fix rounding: make it SAE only.
8206 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8208 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8210 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8211 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8212 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8213 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8214 (round_saeonly_mask_scalar_operand4): Ditto.
8215 (round_saeonly_mask_scalar_op3): Ditto.
8216 (round_saeonly_mask_scalar_op4): Ditto.
8218 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8220 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8221 Implement -maltivec=be for vec_insert and vec_extract.
8223 2014-01-10 DJ Delorie <dj@redhat.com>
8225 * config/msp430/msp430.md (call_internal): Don't allow memory
8226 references with SP as the base register.
8227 (call_value_internal): Likewise.
8228 * config/msp430/constraints.md (Yc): New. For memory references
8229 that don't use SP as a base register.
8231 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8232 "an integer without a # prefix"
8233 * config/msp430/msp430.md (epilogue_helper): Use it.
8235 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8238 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8239 AVX512F gather builtins.
8240 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8241 on gather decls with INTEGER_TYPE masktype.
8242 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8243 directly into the builtin rather than hoisting it before loop.
8245 PR tree-optimization/59387
8246 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8247 (scev_const_prop): If folded_casts and type has undefined overflow,
8248 use force_gimple_operand instead of force_gimple_operand_gsi and
8249 for each added stmt if it is assign with
8250 arith_code_with_undefined_signed_overflow, call
8251 rewrite_to_defined_overflow.
8252 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8253 gimple-fold.h instead.
8254 (arith_code_with_undefined_signed_overflow,
8255 rewrite_to_defined_overflow): Moved to ...
8256 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8257 rewrite_to_defined_overflow): ... here. No longer static.
8258 Include gimplify-me.h.
8259 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8260 rewrite_to_defined_overflow): New prototypes.
8262 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8264 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8266 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8268 * builtins.c (get_object_alignment_2): Minor tweak.
8269 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8271 2014-01-13 Christian Bruel <christian.bruel@st.com>
8273 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8274 optimized non constant lengths.
8276 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8279 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8280 load as __atomic_load_N if possible.
8282 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8284 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8286 (rs6000_expand_builtin): Adjust call.
8288 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8291 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8292 * config/rs6000/rs6000.c: Include target-globals.h.
8293 (rs6000_set_current_function): Instead of doing target_reinit
8294 unconditionally, use save_target_globals_default_opts and
8295 restore_target_globals.
8297 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8299 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8300 (rs6000_expand_builtin): Handle mffs and mtfsf.
8301 (rs6000_init_builtins): Define mffs and mtfsf.
8302 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8303 (rs6000_mffs): New pattern.
8304 (rs6000_mtfsf): New pattern.
8306 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8308 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8309 Start narrowing with START. Apply candidate-use pair
8310 and check overall cost in narrowing.
8311 (iv_ca_prune): Pass new argument.
8313 2014-01-10 Jeff Law <law@redhat.com>
8316 * ree.c (combine_reaching_defs): Ensure the defining statement
8317 occurs before the extension when optimizing extensions with
8318 different source and destination hard registers.
8320 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8323 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8324 vtables into the type inheritance graph.
8326 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8328 PR rtl-optimization/59754
8329 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8330 modes in the REGNO != REGNO case.
8332 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8334 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8336 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8338 PR tree-optimization/59745
8339 * tree-predcom.c (tree_predictive_commoning_loop): Call
8340 free_affine_expand_cache if giving up because components is NULL.
8342 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8343 GC in payload of target_globals struct instead of allocating them on
8344 the heap and the larger structs separately using GC.
8345 * target-globals.h (struct target_globals): Make regs, hard_regs,
8346 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8347 of GTY((skip)) and change type to void *.
8348 (reset_target_globals): Cast loads from those fields to corresponding
8351 2014-01-10 Steve Ellcey <sellcey@mips.com>
8354 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8355 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8356 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8358 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8361 * aarch64-modes.def (CC_Zmode): New flags mode.
8362 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8363 represents an equality.
8364 (aarch64_get_condition_code): Handle CC_Zmode.
8365 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8367 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8369 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8370 extraction in good case.
8372 2014-01-10 Richard Biener <rguenther@suse.de>
8374 PR tree-optimization/59374
8375 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8376 checking after SLP discovery. Mark stmts not participating
8377 in any SLP instance properly.
8379 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8381 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8382 when handling a SET rtx.
8384 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8386 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8387 (cortex-a57): Likewise.
8388 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8390 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8392 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8393 non-iwmmxt builtins.
8395 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8399 * ipa-devirt.c record_target_from_binfo): Take as argument
8400 stack of binfos and lookup matching one for virtual inheritance.
8401 (possible_polymorphic_call_targets_1): Update.
8403 2014-01-10 Huacai Chen <chenhc@lemote.com>
8405 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8406 kernel strings for Loongson-2E/2F/3A.
8408 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8411 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8412 is_gimple_call before calling gimple_call_internal_p.
8414 2014-01-09 Steve Ellcey <sellcey@mips.com>
8416 * Makefile.in (TREE_FLOW_H): Remove.
8417 (TREE_SSA_H): Add file names from tree-flow.h.
8418 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8419 * tree.h: Remove tree-flow.h reference.
8420 * hash-table.h: Remove tree-flow.h reference.
8421 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8422 reference with tree-ssa-loop.h.
8424 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8426 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8427 default element-order behavior for -maltivec.
8428 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8429 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8430 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8431 when targeting big endian, at least for now.
8432 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8434 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8437 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8438 var satisfies use_register_for_decl, just take into account type
8439 alignment, rather than decl alignment.
8441 PR tree-optimization/59622
8442 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8443 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8444 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8445 Don't devirtualize for inplace at all. For targets.length () == 1,
8446 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8448 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8450 * config/i386/i386.md (cpu): Remove the unused btver1.
8452 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8454 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8456 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8459 * tree-core.h (struct target_globals): New forward declaration.
8460 (struct tree_target_option): Add globals field.
8461 * tree.h (TREE_TARGET_GLOBALS): Define.
8462 (prepare_target_option_nodes_for_pch): New prototype.
8463 * target-globals.h (struct target_globals): Define even if
8465 * tree.c (prepare_target_option_node_for_pch,
8466 prepare_target_option_nodes_for_pch): New functions.
8467 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8468 * config/i386/i386.c: Include target-globals.h.
8469 (ix86_set_current_function): Instead of doing target_reinit
8470 unconditionally, use save_target_globals_default_opts and
8471 restore_target_globals.
8473 2014-01-09 Richard Biener <rguenther@suse.de>
8475 PR tree-optimization/59715
8476 * tree-cfg.h (split_critical_edges): Declare.
8477 * tree-cfg.c (split_critical_edges): Export.
8478 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8480 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8482 * cfgexpand.c (expand_stack_vars): Optionally disable
8483 asan stack protection.
8484 (expand_used_vars): Likewise.
8485 (partition_stack_vars): Likewise.
8486 * asan.c (asan_emit_stack_protection): Optionally disable
8487 after return stack usage.
8488 (instrument_derefs): Optionally disable memory access instrumentation.
8489 (instrument_builtin_call): Likewise.
8490 (instrument_strlen_call): Likewise.
8491 (asan_protect_global): Optionally disable global variables protection.
8492 * doc/invoke.texi: Added doc for new options.
8493 * params.def: Added new options.
8494 * params.h: Likewise.
8496 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8498 PR rtl-optimization/59724
8499 * ifcvt.c (cond_exec_process_if_block): Don't call
8500 flow_find_head_matching_sequence with 0 longest_match.
8501 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8502 non-active insns if !stop_after.
8503 (try_head_merge_bb): Revert 2014-01-07 changes.
8505 2014-01-08 Jeff Law <law@redhat.com>
8507 * ree.c (get_sub_rtx): New function, extracted from...
8508 (merge_def_and_ext): Here.
8509 (combine_reaching_defs): Use get_sub_rtx.
8511 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8513 * cgraph.h (varpool_variable_node): Do not choke on null node.
8515 2014-01-08 Catherine Moore <clm@codesourcery.com>
8517 * config/mips/mips.md (simple_return): Attempt to use JRC
8519 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8521 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8523 PR rtl-optimization/59137
8524 * reorg.c (steal_delay_list_from_target): Call update_block for
8526 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8528 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8530 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8531 two duplicate entries.
8533 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8536 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8538 * config/mips/mips.c (mips_truncated_op_cost): New function.
8539 (mips_rtx_costs): Adjust test for BADDU.
8540 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8542 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8544 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8545 (*baddu_si): ...this new pattern.
8547 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8550 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8552 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8555 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8557 (expand_expr, expand_normal): Adjust.
8558 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8559 inner_reference_p. Use inner_reference_p to expand inner references.
8560 (store_expr): Adjust.
8561 * cfgexpand.c (expand_call_stmt): Adjust.
8563 2014-01-08 Rong Xu <xur@google.com>
8565 * gcov-io.c (gcov_var): Move from gcov-io.h.
8566 (gcov_position): Ditto.
8567 (gcov_is_error): Ditto.
8568 (gcov_rewrite): Ditto.
8569 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8570 only part to libgcc/libgcov.h.
8572 2014-01-08 Marek Polacek <polacek@redhat.com>
8575 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8577 2014-01-08 Marek Polacek <polacek@redhat.com>
8580 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8582 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8584 PR rtl-optimization/59649
8585 * stor-layout.c (get_mode_bounds): For BImode return
8586 0 and STORE_FLAG_VALUE.
8588 2014-01-08 Richard Biener <rguenther@suse.de>
8591 * gimple.h (is_gimple_builtin_call): Remove.
8592 (gimple_builtin_call_types_compatible_p): New.
8593 (gimple_call_builtin_p): New overload.
8594 * gimple.c (is_gimple_builtin_call): Remove.
8595 (validate_call): Rename to ...
8596 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8598 (validate_type): New static function.
8599 (gimple_call_builtin_p): New overload and adjust.
8600 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8601 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8602 (gimple_fold_stmt_to_constant_1): Likewise.
8603 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8605 2014-01-08 Richard Biener <rguenther@suse.de>
8608 * gimplify.c (gimplify_expr): Gimplify register-register type
8609 VIEW_CONVERT_EXPRs to separate stmts.
8611 2014-01-07 Jeff Law <law@redhat.com>
8614 * ree.c (combine_set_extension): Handle case where source
8615 and destination registers in an extension insn are different.
8616 (combine_reaching_defs): Allow source and destination registers
8617 in extension to be different under limited circumstances.
8618 (add_removable_extension): Remove restriction that the
8619 source and destination registers in the extension are the same.
8620 (find_and_remove_re): Emit a copy from the extension's
8621 destination to its source after the defining insn if
8622 the source and destination registers are different.
8625 * ifcvt.c (merge_if_block): If we are merging a block with more than
8626 one successor with a block with no successors, remove any BARRIER
8627 after the second block.
8629 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8631 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8633 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8636 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8637 for 14-bit register offsets when INT14_OK_STRICT is false.
8639 2014-01-07 Roland Stigge <stigge@antcom.de>
8640 Michael Meissner <meissner@linux.vnet.ibm.com>
8643 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8644 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8646 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8648 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8651 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8653 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8654 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8655 rtx is const0_rtx or not.
8657 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8660 * target-globals.c (save_target_globals): Remove this_fn_optab
8662 * toplev.c: Include optabs.h.
8663 (target_reinit): Temporarily restore the global options if another
8664 set of options are in force.
8666 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8668 PR rtl-optimization/58668
8669 * cfgcleanup.c (flow_find_cross_jump): Don't count
8670 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8671 to determine what is counted.
8672 (flow_find_head_matching_sequence): Use active_insn_p to determine
8674 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8676 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8677 determine what is counted.
8679 PR tree-optimization/59643
8680 * tree-predcom.c (split_data_refs_to_components): If one dr is
8681 read and one write, determine_offset fails and the write isn't
8682 in the bad component, just put the read into the bad component.
8684 2014-01-07 Mike Stump <mikestump@comcast.net>
8685 Jakub Jelinek <jakub@redhat.com>
8688 * tree-core.h (struct tree_optimization_option): Change optabs
8689 type from unsigned char * to void *.
8690 * optabs.c (init_tree_optimization_optabs): Adjust
8691 TREE_OPTIMIZATION_OPTABS initialization.
8693 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8696 * config/i386/i386.h (struct machine_function): Add
8697 no_drap_save_restore field.
8698 * config/i386/i386.c (ix86_save_reg): Use
8699 !cfun->machine->no_drap_save_restore instead of
8700 crtl->stack_realign_needed.
8701 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8702 this function clears frame_pointer_needed. Set
8703 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8704 and DRAP reg is needed.
8706 2014-01-06 Marek Polacek <polacek@redhat.com>
8709 * doc/implement-c.texi: Mention that other integer types are
8710 permitted as bit-field types in strictly conforming mode.
8712 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8714 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8717 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8719 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8721 2014-01-06 Martin Jambor <mjambor@suse.cz>
8724 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8726 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8728 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8732 * var-tracking.c (add_stores): Preserve the value of the source even if
8733 we don't record the store.
8735 2014-01-06 Terry Guo <terry.guo@arm.com>
8737 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8739 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8742 * config/darwin.c (darwin_function_section): Adjust return values to
8743 correspond to optimisation changes made in r206070.
8745 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8747 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8748 from prefetch_block tune setting.
8749 (nocona_cost): Correct size of prefetch block to 64.
8751 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8753 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8754 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8755 used to save the static chain register in the computation of the offset
8756 from which the FP registers need to be restored.
8758 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8760 PR tree-optimization/59519
8761 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8762 ICE if get_current_def (current_new_name) is already non-NULL, as long
8763 as it is a phi result of some other phi in *new_exit_bb that has
8766 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8767 or vmovdqu* for misaligned_operand.
8768 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8769 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8770 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8771 aligned_mem for AVX512F masked aligned load and store builtins and for
8774 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8776 PR tree-optimization/59651
8777 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8778 Address range for negative step should be added by TYPE_SIZE_UNIT.
8780 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8782 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8783 overlapping registers also for registers other than the stack pointer.
8785 2014-01-03 Marek Polacek <polacek@redhat.com>
8788 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8791 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8794 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8797 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8798 (push splitter): Use <P:MODE_SIZE> instead of
8799 GET_MODE_SIZE (<P:MODE>mode).
8800 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8801 (mov -1, reg peephole2): Likewise.
8802 * config/i386/sse.md (*mov<mode>_internal,
8803 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8804 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8805 *<code><mode>3, *andnot<mode>3<mask_name>,
8806 <mask_codefor><code><mode>3<mask_name>): Likewise.
8807 * config/i386/subst.md (mask_mode512bit_condition,
8808 sd_mask_mode512bit_condition): Likewise.
8810 2014-01-02 Xinliang David Li <davidxl@google.com>
8812 PR tree-optimization/59303
8813 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8814 (dump_predicates): Better output format.
8815 (pred_equal_p): New function.
8816 (is_neq_relop_p): Ditto.
8817 (is_neq_zero_form_p): Ditto.
8818 (pred_expr_equal_p): Ditto.
8819 (pred_neg_p): Ditto.
8820 (simplify_pred): Ditto.
8821 (simplify_preds_2): Ditto.
8822 (simplify_preds_3): Ditto.
8823 (simplify_preds_4): Ditto.
8824 (simplify_preds): Ditto.
8826 (push_to_worklist): Ditto.
8827 (get_pred_info_from_cmp): Ditto.
8828 (is_degenerated_phi): Ditto.
8829 (normalize_one_pred_1): Ditto.
8830 (normalize_one_pred): Ditto.
8831 (normalize_one_pred_chain): Ditto.
8832 (normalize_preds): Ditto.
8833 (normalize_cond_1): Remove function.
8834 (normalize_cond): Ditto.
8835 (is_gcond_subset_of): Ditto.
8836 (is_subset_of_any): Ditto.
8837 (is_or_set_subset_of): Ditto.
8838 (is_and_set_subset_of): Ditto.
8839 (is_norm_cond_subset_of): Ditto.
8840 (pred_chain_length_cmp): Ditto.
8841 (convert_control_dep_chain_into_preds): Type change.
8842 (find_predicates): Ditto.
8843 (find_def_preds): Ditto.
8844 (destroy_predicates_vecs): Ditto.
8845 (find_matching_predicates_in_rest_chains): Ditto.
8846 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8847 (is_pred_expr_subset): Ditto.
8848 (is_pred_chain_subset_of): Ditto.
8849 (is_included_in): Ditto.
8850 (is_superset_of): Ditto.
8852 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8854 Update copyright years.
8856 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8858 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8859 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8860 config/arc/arc.md, config/arc/arc.opt,
8861 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8862 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8863 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8864 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8865 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8866 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8867 vtable-verify.c, vtable-verify.h: Use the standard form for the
8870 2014-01-02 Tobias Burnus <burnus@net-b.de>
8872 * gcc.c (process_command): Update copyright notice dates.
8873 * gcov-dump.c: Ditto.
8875 * doc/cpp.texi: Bump @copying's copyright year.
8876 * doc/cppinternals.texi: Ditto.
8877 * doc/gcc.texi: Ditto.
8878 * doc/gccint.texi: Ditto.
8879 * doc/gcov.texi: Ditto.
8880 * doc/install.texi: Ditto.
8881 * doc/invoke.texi: Ditto.
8883 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8885 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8887 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8889 * config/i386/sse.md (*mov<mode>_internal): Guard
8890 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8892 PR rtl-optimization/59647
8893 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8894 new_rtx into UNSIGNED_FLOAT rtxes.
8896 Copyright (C) 2014 Free Software Foundation, Inc.
8898 Copying and distribution of this file, with or without modification,
8899 are permitted in any medium without royalty provided the copyright
8900 notice and this notice are preserved.