1 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
5 2014-05-06 Richard Biener <rguenther@suse.de>
7 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
8 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
9 (TODO_verify_all): Adjust.
10 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
11 TODO_verify_stmts and TODO_verify_rtl_sharing.
12 * bb-reorder.c: Likewise.
13 * cfgexpand.c: Likewise.
16 * function.c: Likewise.
19 * gimple-ssa-isolate-paths.c: Likewise.
20 * gimple-ssa-strength-reduction.c: Likewise.
21 * ipa-split.c: Likewise.
22 * loop-init.c: Likewise.
23 * loop-unroll.c: Likewise.
24 * lower-subreg.c: Likewise.
25 * modulo-sched.c: Likewise.
26 * postreload-gcse.c: Likewise.
27 * predict.c: Likewise.
29 * sched-rgn.c: Likewise.
30 * store-motion.c: Likewise.
32 * trans-mem.c: Likewise.
33 * tree-call-cdce.c: Likewise.
34 * tree-cfg.c: Likewise.
35 * tree-cfgcleanup.c: Likewise.
36 * tree-complex.c: Likewise.
37 * tree-eh.c: Likewise.
38 * tree-emutls.c: Likewise.
39 * tree-if-conv.c: Likewise.
40 * tree-into-ssa.c: Likewise.
41 * tree-loop-distribution.c: Likewise.
42 * tree-object-size.c: Likewise.
43 * tree-parloops.c: Likewise.
44 * tree-pass.h: Likewise.
45 * tree-sra.c: Likewise.
46 * tree-ssa-ccp.c: Likewise.
47 * tree-ssa-copy.c: Likewise.
48 * tree-ssa-copyrename.c: Likewise.
49 * tree-ssa-dce.c: Likewise.
50 * tree-ssa-dom.c: Likewise.
51 * tree-ssa-dse.c: Likewise.
52 * tree-ssa-forwprop.c: Likewise.
53 * tree-ssa-ifcombine.c: Likewise.
54 * tree-ssa-loop-ch.c: Likewise.
55 * tree-ssa-loop-ivcanon.c: Likewise.
56 * tree-ssa-loop.c: Likewise.
57 * tree-ssa-math-opts.c: Likewise.
58 * tree-ssa-phiopt.c: Likewise.
59 * tree-ssa-phiprop.c: Likewise.
60 * tree-ssa-pre.c: Likewise.
61 * tree-ssa-reassoc.c: Likewise.
62 * tree-ssa-sink.c: Likewise.
63 * tree-ssa-strlen.c: Likewise.
64 * tree-ssa-tail-merge.c: Likewise.
65 * tree-ssa-uncprop.c: Likewise.
66 * tree-switch-conversion.c: Likewise.
67 * tree-tailcall.c: Likewise.
68 * tree-vect-generic.c: Likewise.
69 * tree-vectorizer.c: Likewise.
70 * tree-vrp.c: Likewise.
72 * var-tracking.c: Likewise.
73 * bt-load.c: Likewise.
74 * cfgcleanup.c: Likewise.
75 * combine-stack-adj.c: Likewise.
76 * combine.c: Likewise.
77 * compare-elim.c: Likewise.
78 * config/epiphany/resolve-sw-modes.c: Likewise.
79 * config/i386/i386.c: Likewise.
80 * config/mips/mips.c: Likewise.
81 * config/s390/s390.c: Likewise.
82 * config/sh/sh_treg_combine.cc: Likewise.
83 * config/sparc/sparc.c: Likewise.
88 * mode-switching.c: Likewise.
90 * postreload.c: Likewise.
92 * reg-stack.c: Likewise.
93 * regcprop.c: Likewise.
94 * regrename.c: Likewise.
97 2014-05-06 Richard Biener <rguenther@suse.de>
100 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
101 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
103 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
106 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
108 2014-05-05 Radovan Obradovic <robradovic@mips.com>
109 Tom de Vries <tom@codesourcery.com>
111 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
112 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
114 (@node Miscellaneous Register Hooks): New node.
115 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
116 * doc/tm.texi: Regenerate.
118 2014-05-05 Marek Polacek <polacek@redhat.com>
121 * opts.c (common_handle_option): Call error_at instead of warning_at.
123 2014-05-05 Richard Biener <rguenther@suse.de>
125 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
126 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
127 under the TODO_verify_il umbrella.
129 2014-05-05 Richard Biener <rguenther@suse.de>
131 * passes.c (execute_function_todo): Move TODO_verify_flow under
132 the TODO_verify_ul umbrella.
134 2014-05-05 Richard Biener <rguenther@suse.de>
137 * fold-const.c (fold_binary_loc): Consistently avoid
138 canonicalizing X & CST away from a CST that is the mask
141 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
143 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
144 int argument to enum machine_mode.
145 (picochip_class_max_nregs): Ditto.
146 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
147 (picochip_class_max_nregs): Ditto.
149 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
151 * target.def: Add new target hook.
152 * doc/tm.texi: Regenerate.
153 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
154 * targhooks.c (default_keep_leaf_when_profiled): New function.
156 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
157 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
159 2014-05-05 Bin Cheng <bin.cheng@arm.com>
161 PR tree-optimization/60363
162 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
163 (copy_phi_args): New parameters. Call get_value_locus_in_path.
164 (update_destination_phis): New parameter.
165 (create_edge_and_update_destination_phis): Ditto.
166 (ssa_fix_duplicate_block_edges): Pass new arguments.
167 (thread_single_edge): Ditto.
169 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
171 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
172 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
173 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
174 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
175 Use RS6000_BTM_HARD_FLOAT.
176 (BU_MISC_2): Likewise.
177 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
178 RS6000_BTM_HARD_FLOAT.
179 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
181 (rs6000_invalid_builtin): Add hard floating builtin support.
182 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
184 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
186 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
188 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
189 Add missing function* argument.
191 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
193 * lra-constraints.c (valid_address_p): Move earlier in file.
194 Add a constraint argument to the address_info version.
195 (satisfies_memory_constraint_p): New function.
196 (satisfies_address_constraint_p): Likewise.
197 (process_alt_operands, curr_insn_transform): Use them.
198 (process_address): Pass the constraint to valid_address_p when
199 checking address operands.
201 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
203 * config/mips/mips.c (mips_isa_rev): New variable.
204 (mips_set_architecture): Set it.
205 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
207 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
208 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
209 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
210 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
211 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
212 conditions in terms of mips_isa_rev.
213 (mips_isa_rev): Declare.
215 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
217 * config/sh/sh-mem.cc: Use tabs instead of spaces.
218 (prob_unlikely, prob_likely): Make variables const.
220 2014-05-03 Denis Chertykov <chertykov@gmail.com>
222 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
224 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
226 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
228 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
230 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
231 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
232 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
234 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
235 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
237 Replace usage of ROUND_REG with sh_round_reg.
238 Use CEIL instead of ROUND_ADVANCE.
240 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
243 * config/sh/sh.c: Include stdlib headers before everything else.
245 2014-05-02 Jakub Jelinek <jakub@redhat.com>
247 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
248 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
249 (gimplify_adjust_omp_clauses): Simd region is never
250 directly nested in combined parallel. Instead, for linear
251 with copyin/copyout, if in combined for simd loop, make decl
252 firstprivate/lastprivate on OMP_FOR.
253 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
254 expand_omp_for_static_chunk): When setting endvar, also set
255 fd->loop.v to the same value.
257 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
259 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
261 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
263 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
266 2014-05-02 Marek Polacek <polacek@redhat.com>
268 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
270 2014-02-26 Kito Cheng <kito@0xlab.org>
272 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
273 to a C expression marco.
274 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
275 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
276 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
277 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
278 HONOR_REG_ALLOC_ORDER.
279 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
281 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
283 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
285 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
287 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
289 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
291 * tree-if-conv.c (is_cond_scalar_reduction): New function.
292 (convert_scalar_cond_reduction): Likewise.
293 (predicate_scalar_phi): Add recognition and transformation
294 of simple conditioanl reduction to be vectorizable.
296 2014-05-01 Marek Polacek <polacek@redhat.com>
299 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
301 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
303 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
304 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
305 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
306 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
307 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
308 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
309 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
310 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
312 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
314 * config/arc/arc.opt (mlra): Move comment above option name
315 to avoid mis-parsing as language options.
317 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
319 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
320 * config/sol2.h: ... here.
321 * config/sol2-10.h: Remove.
323 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
324 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
325 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
326 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
327 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
328 * config/sol2.h: ... here.
329 (SECTION_NAME_FORMAT): Don't redefine.
330 (STARTFILE_ARCH32_SPEC): Rename to ...
331 (STARTFILE_ARCH_SPEC): ... this.
332 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
333 * config/sparc/sol2.h: ... here.
334 (SECTION_NAME_FORMAT): Don't undef.
335 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
336 (SUBTARGET_EXTRA_SPECS): Remove.
337 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
339 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
340 (MD_STARTFILE_PREFIX): Remove.
341 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
342 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
343 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
344 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
345 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
346 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
347 * config/i386/sol2.h: ... here.
348 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
349 * config/i386/sol2-bi.h: Remove.
350 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
351 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
353 * config/i386/t-sol2-64: Rename to ...
354 * config/i386/t-sol2: ... this.
355 * config/sparc/t-sol2-64: Rename to ...
356 * config/sparc/t-sol2: ... this.
358 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
359 sol2_tm_file_head, sol2_tm_file_tail.
360 Include ${cpu_type}/sol2.h before sol2.h.
362 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
363 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
364 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
365 Reflect i386/t-sol2-64 renaming.
366 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
367 Reflect sparc/t-sol2-64 renaming.
369 2014-04-30 Richard Biener <rguenther@suse.de>
371 * passes.c (execute_function_todo): Move TODO_verify_stmts
372 and TODO_verify_ssa under the TODO_verify_il umbrella.
373 * tree-ssa.h (verify_ssa): Adjust prototype.
374 * tree-ssa.c (verify_ssa): Add parameter to tell whether
375 we should verify SSA operands.
376 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
377 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
378 whether we should verify whether not throwing stmts have EH info.
379 * graphite-scop-detection.c (create_sese_edges): Adjust.
380 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
381 * tree-eh.c (lower_try_finally_switch): Do not add the
382 default case label twice.
384 2014-04-30 Marek Polacek <polacek@redhat.com>
386 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
387 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
388 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
389 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
391 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
393 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
394 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
395 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
396 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
397 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
398 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
399 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
400 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
402 2014-04-29 David Malcolm <dmalcolm@redhat.com>
404 * tree-cfg.c (dump_function_to_file): Dump the return type of
405 functions, in a line to itself before the function body, mimicking
406 the layout of a C function.
408 2014-04-29 Jakub Jelinek <jakub@redhat.com>
410 PR tree-optimization/60971
411 * tree-tailcall.c (process_assignment): Reject conversions which
414 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
416 * calls.c (initialize_argument_information): Always treat
417 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
418 (expand_call): Likewise.
419 (emit_library_call_calue_1): Likewise.
420 * expr.c (PUSH_ARGS_REVERSED): Do not define.
421 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
424 2014-04-29 Nick Clifton <nickc@redhat.com>
426 * config/msp430/msp430.md (umulsidi): Fix typo.
427 (mulhisi3): Enable even inside interrupt handlers.
428 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
429 bigger return address pushed in large mode.
431 2014-04-29 Nick Clifton <nickc@redhat.com>
433 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
434 (arc_init_reg_tables): Use a machine_mode enum to iterate over
436 * config/m32r/m32r.c (init_reg_tables): Likewise.
437 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
438 enum to hold the modes.
440 2014-04-29 Richard Biener <rguenther@suse.de>
442 * dominance.c (free_dominance_info): Add overload with
444 (dom_info_state): Likewise.
445 (dom_info_available_p): Likewise.
446 * basic-block.h (free_dominance_info, dom_info_state,
447 dom_info_available_p): Declare overloads.
448 * passes.c (execute_function_todo): Verify that verifiers
449 don't change dominator info state. Drop dominator info
450 for IPA pass invocations.
451 * cgraph.c (release_function_body): Restore asserts that
452 dominator information is released.
454 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
456 * doc/invoke.texi: Fix typo.
457 * tree-vrp.c: Fix typos.
458 * gimple.c (infer_nonnull_range): Reorder operands of an &&
461 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
463 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
465 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
467 * config/aarch64/aarch64-builtins.c
468 (aarch64_types_storestruct_lane_qualifiers): New.
469 (TYPES_STORESTRUCT_LANE): Likewise.
470 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
471 (st3_lane): Likewise.
472 (st4_lane): Likewise.
473 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
474 (vec_store_lanesci_lane<mode>): Likewise.
475 (vec_store_lanesxi_lane<mode>): Likewise.
476 (aarch64_st2_lane<VQ:mode>): Likewise.
477 (aarch64_st3_lane<VQ:mode>): Likewise.
478 (aarch64_st4_lane<VQ:mode>): Likewise.
479 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
480 * config/aarch64/arm_neon.h
481 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
482 use new macro arguments.
483 (__ST3_LANE_FUNC): Likewise.
484 (__ST4_LANE_FUNC): Likewise.
485 * config/aarch64/iterators.md (V_TWO_ELEM): New.
486 (V_THREE_ELEM): Likewise.
487 (V_FOUR_ELEM): Likewise.
489 2014-04-28 David Malcolm <dmalcolm@redhat.com>
491 * doc/gimple.texi: Replace the description of the now-defunct
492 union gimple_statement_d with a diagram showing the
493 gimple_statement_base class hierarchy and its relationships to
494 the GSS_ and GIMPLE_ enums.
496 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
498 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
499 * config/aarch64/aarch64.c
500 (aarch64_cannot_change_mode_class): Weaken conditions.
501 (aarch64_modes_tieable_p): New.
502 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
504 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
506 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
507 (loadsync_<mode>): Change mode.
508 (load_quadpti, store_quadpti): New.
509 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
510 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
512 2014-04-28 Martin Jambor <mjambor@suse.cz>
514 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
515 same alias type as the original statement.
516 (subreplacement_assignment_data): New type.
517 (handle_unscalarized_data_in_subtree): New type of parameter,
518 generate new memory accesses with same alias type as the original
520 (load_assign_lhs_subreplacements): Likewise.
521 (sra_modify_constructor_assign): Generate new memory accesses with
522 same alias type as the original statement.
524 2014-04-28 Richard Biener <rguenther@suse.de>
526 * tree-pass.h (TODO_verify_il): Define.
527 (TODO_verify_all): Complete properly.
528 * passes.c (execute_function_todo): Move existing loop-closed
529 SSA verification under TODO_verify_il.
530 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
531 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
532 Fix tree sharing issue.
534 2014-04-28 Richard Biener <rguenther@suse.de>
537 * builtins.def (DEF_C11_BUILTIN): Add.
538 (BUILT_IN_ALIGNED_ALLOC): Likewise.
539 * coretypes.h (enum function_class): Add function_c11_misc.
540 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
541 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
542 (call_may_clobber_ref_p_1): Likewise.
543 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
544 (mark_all_reaching_defs_necessary_1): Likewise.
545 (propagate_necessity): Likewise.
546 (eliminate_unnecessary_stmts): Likewise.
547 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
549 2014-04-28 Richard Biener <rguenther@suse.de>
551 * tree-vrp.c (vrp_var_may_overflow): Remove.
552 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
553 with overflow immediately bump to one before that value and
554 let iteration figure out overflow status.
556 2014-04-28 Richard Biener <rguenther@suse.de>
558 * configure.ac: Do valgrind header checks unconditionally.
559 Add --enable-valgrind-annotations.
560 * system.h: Guard valgrind header inclusion with
561 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
562 * alloc-pool.c (pool_alloc, pool_free): Use
563 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
564 to guard possibly dead code.
565 * config.in: Regenerated.
566 * configure: Likewise.
568 2014-04-28 Jeff Law <law@redhat.com>
570 PR tree-optimization/60902
571 * tree-ssa-threadedge.c
572 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
573 over real defs when invalidating outputs from statements that do not
574 produce useful outputs for threading.
576 2014-04-28 Richard Biener <rguenther@suse.de>
578 PR tree-optimization/60979
579 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
580 SCOPs that end in a block with a successor with abnormal
583 2014-04-28 Richard Biener <rguenther@suse.de>
585 * tree-pass.h (execute_pass_list): Adjust prototype.
586 * passes.c (pass_manager::execute_early_local_passes):
588 (do_per_function): Change callback signature, push all actual
589 work to the callbals.
590 (do_per_function_toporder): Likewise.
591 (execute_function_dump): Adjust.
592 (execute_function_todo): Likewise.
593 (clear_last_verified): Likewise.
594 (verify_curr_properties): Likewise.
595 (update_properties_after_pass): Likewise.
596 (execute_pass_list_1): Split out from ...
597 (execute_pass_list): ... here. Adjust.
598 (execute_ipa_pass_list): Likewise.
599 * cgraphunit.c (cgraph_add_new_function): Adjust.
600 (analyze_function): Likewise.
601 (expand_function): Likewise.
602 * cgraph.c (release_function_body): Free dominance info
603 here instead of asserting it was magically freed elsewhere.
605 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
607 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
608 * configure: Regenerate.
609 * config/sparc/sparc.opt (muser-mode): New option.
610 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
612 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
613 * doc/invoke.texi (SPARC options): Document -muser-mode.
615 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
617 * cselib.c (find_slot_memmode): Delete.
618 (cselib_hasher): Change compare_type to a struct.
619 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
621 (preserve_constants_and_equivs): Adjust for new compare_type.
622 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
623 (wrap_constant): Delete.
624 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
626 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
628 * doc/install.texi (Building with profile feedback): Remove
631 2014-04-26 Tom de Vries <tom@codesourcery.com>
633 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
636 2014-04-25 Cary Coutant <ccoutant@google.com>
639 * dwarf2out.c (should_move_die_to_comdat): A type definition
640 can contain a subprogram definition, but don't move it to a
642 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
643 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
645 (clone_tree_hash): Rename to...
646 (clone_tree_partial): ...this; change callers. Copy
647 DW_TAG_subprogram DIEs as declarations.
648 (copy_decls_walk): Don't copy children of a declaration into a
651 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
654 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
657 2014-04-25 Jiong Wang <jiong.wang@arm.com>
659 * config/arm/predicates.md (call_insn_operand): Add long_call check.
660 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
662 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
665 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
667 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
670 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
672 PR tree-optimization/60930
673 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
674 creating a multiply candidate by folding two constant
675 multiplicands when the result overflows.
677 2014-04-25 Jakub Jelinek <jakub@redhat.com>
679 PR tree-optimization/60960
680 * tree-vect-generic.c (expand_vector_operation): Only call
681 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
683 2014-04-25 Tom de Vries <tom@codesourcery.com>
685 * expr.c (clobber_reg_mode): New function.
686 * expr.h (clobber_reg): New function.
688 2014-04-25 Tom de Vries <tom@codesourcery.com>
690 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
693 2014-04-25 Radovan Obradovic <robradovic@mips.com>
694 Tom de Vries <tom@codesourcery.com>
696 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
698 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
699 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
700 new argument to find_all_hard_reg_sets call.
702 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
704 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
705 Use HOST_WIDE_INT_C for mask literal.
706 (aarch_rev16_shleft_mask_imm_p): Likewise.
708 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
711 * config/sparc/sparc.md (ashlsi3_extend): Delete.
713 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
715 PR preprocessor/56540
716 * config/i386/i386-c.c (ix86_target_macros): Define
717 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
719 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
721 * configure.ac (tga_func): Remove.
722 (LIB_TLS_SPEC): Remove.
723 * configure: Regenerate.
724 * config.in: Regenerate.
725 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
727 2014-04-25 Richard Biener <rguenther@suse.de>
730 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
731 call stmt use/clobber sets during stmt walk instead of
732 walking the possibly incomplete set of caller edges.
734 2014-04-25 Richard Biener <rguenther@suse.de>
737 * passes.c (apply_ipa_transforms): Inline into only caller ...
738 (execute_one_pass): ... here. Properly bring in function
739 bodies for nodes we want to apply IPA transforms to.
741 2014-04-24 Cong Hou <congh@google.com>
743 PR tree-optimization/60896
744 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
745 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
746 (vect_mark_pattern_stmts): Set the def type of all statements in
747 PATTERN_DEF_SEQ as vect_internal_def.
749 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
751 * doc/extend.texi (PowerPC Built-in Functions): Document new
752 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
754 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
756 * config/rs6000/predicates.md (const_0_to_3_operand): New
757 predicate to match 0..3 integer constants.
759 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
760 to support adding miscellaneous builtin functions.
761 (BU_DFP_MISC_2): Likewise.
762 (BU_P7_MISC_1): Likewise.
763 (BU_P7_MISC_2): Likewise.
764 (BU_P8V_MISC_3): Likewise.
765 (BU_MISC_1): Likewise.
766 (BU_MISC_2): Likewise.
767 (DIVWE): Add extended divide builtin functions.
775 (DXEX): Add decimal floating-point builtin functions.
787 (CDTBCD): Add new BCD builtin functions.
791 (BCDADD_LT): Likewise.
792 (BCDADD_EQ): Likewise.
793 (BCDADD_GT): Likewise.
794 (BCDADD_OV): Likewise.
796 (BCDSUB_LT): Likewise.
797 (BCDSUB_EQ): Likewise.
798 (BCDSUB_GT): Likewise.
799 (BCDSUB_OV): Likewise.
800 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
801 (UNPACK_TD): Likewise.
803 (UNPACK_TF): Likewise.
804 (UNPACK_TF_0): Likewise.
805 (UNPACK_TF_1): Likewise.
806 (PACK_V1TI): Likewise.
807 (UNPACK_V1TI): Likewise.
809 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
810 support for decimal floating point builtin functions.
811 (rs6000_expand_ternop_builtin): Add checks for the new builtin
812 functions that take constant arguments.
813 (rs6000_invalid_builtin): Add decimal floating point builtin
815 (rs6000_init_builtins): Setup long double, _Decimal64, and
816 _Decimal128 types for new builtin functions.
817 (builtin_function_type): Set the unsigned flags appropriately for
818 the new builtin functions.
819 (rs6000_opt_masks): Add support for decimal floating point builtin
822 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
823 floating point builtin functions.
824 (RS6000_BTM_COMMON): Likewise.
825 (RS6000_BTI_long_double): Likewise.
826 (RS6000_BTI_dfloat64): Likewise.
827 (RS6000_BTI_dfloat128): Likewise.
828 (long_double_type_internal_node): Likewise.
829 (dfloat64_type_internal_node): Likewise.
830 (dfloat128_type_internal_node): Likewise.
832 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
833 2.07 bcd arithmetic instructions.
834 (UNSPEC_BCDSUB): Likewise.
835 (UNSPEC_BCD_OVERFLOW): Likewise.
836 (UNSPEC_BCD_ADD_SUB): Likewise.
837 (bcd_add_sub): Likewise.
838 (BCD_TEST): Likewise.
839 (bcd<bcd_add_sub>): Likewise.
840 (bcd<bcd_add_sub>_test): Likewise.
841 (bcd<bcd_add_sub>_test2): Likewise.
842 (bcd<bcd_add_sub>_<code>): Likewise.
843 (peephole2 for combined bcd ops): Likewise.
845 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
846 decimal floating point builtin functions.
847 (UNSPEC_DENBCD): Likewise.
848 (UNSPEC_DXEX): Likewise.
849 (UNSPEC_DIEX): Likewise.
850 (UNSPEC_DSCLI): Likewise.
851 (UNSPEC_DSCRI): Likewise.
852 (D64_D128): Likewise.
853 (dfp_suffix): Likewise.
854 (dfp_ddedpd_<mode>): Likewise.
855 (dfp_denbcd_<mode>): Likewise.
856 (dfp_dxex_<mode>): Likewise.
857 (dfp_diex_<mode>): Likewise.
858 (dfp_dscli_<mode>): Likewise.
859 (dfp_dscri_<mode>): Likewise.
861 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
863 (UNSPEC_CDTBCD): Likewise.
864 (UNSPEC_CBCDTD): Likewise.
865 (UNSPEC_DIVE): Add support for new extended divide builtin
867 (UNSPEC_DIVEO): Likewise.
868 (UNSPEC_DIVEU): Likewise.
869 (UNSPEC_DIVEUO): Likewise.
870 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
871 pack/unpack 128-bit types.
872 (UNSPEC_PACK_128BIT): Likewise.
873 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
874 (udiv<mode>3): Use idiv_ldiv mode attribute.
875 (div<mode>3): Likewise.
876 (addg6s): Add new BCD builtin functions.
879 (UNSPEC_DIV_EXTEND): Add support for new extended divide
881 (div_extend): Likewise.
882 (div<div_extend>_<mode>"): Likewise.
883 (FP128_64): Add support for new builtin functions to pack/unpack
885 (unpack<mode>): Likewise.
886 (unpacktf_0): Likewise.
887 (unpacktf_1): Likewise.
888 (unpack<mode>_dm): Likewise.
889 (unpack<mode>_nodm): Likewise.
890 (pack<mode>): Likewise.
891 (unpackv1ti): Likewise.
892 (packv1ti): Likewise.
894 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
896 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
899 2014-04-24 Jakub Jelinek <jakub@redhat.com>
901 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
902 * gimplify.c (omp_is_private): Change last argument's type to int.
903 Only diagnose lastprivate if the simd argument is 1, only diagnose
904 linear if the simd argument is 2.
905 (gimplify_omp_for): Adjust omp_is_private callers. When adding
906 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
907 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
908 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
909 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
910 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
911 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
912 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
913 * tree-nested.c (convert_nonlocal_omp_clauses,
914 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
916 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
919 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
922 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
924 * flag-types.h (enum ivar_visibility): Add.
926 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
928 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
931 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
933 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
935 2014-04-24 Radovan Obradovic <robradovic@mips.com>
936 Tom de Vries <tom@codesourcery.com>
938 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
939 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
941 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
942 * emit-rtl.c (try_split): Same.
944 2014-04-24 Radovan Obradovic <robradovic@mips.com>
945 Tom de Vries <tom@codesourcery.com>
947 * common.opt (fuse-caller-save): New option.
949 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
951 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
954 2014-04-24 Richard Biener <rguenther@suse.de>
956 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
957 during TER and instead use the sepops interface for expanding
958 non-GIMPLE_SINGLE_RHS.
960 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
962 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
963 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
965 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
967 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
968 assembler 64-bit option.
969 * configure: Regenerate.
971 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
973 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
974 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
975 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
976 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
977 (TARGET_CRYPTO): Take TARGET_SIMD into account.
979 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
981 * config/aarch64/aarch64-builtins.c
982 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
983 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
984 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
985 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
987 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
988 (Vrevsuff): New mode attribute.
990 2014-04-24 Terry Guo <terry.guo@arm.com>
992 * config/arm/arm.h (machine_function): Define variable
993 after_arm_reorg here.
994 * config/arm/arm.c (after_arm_reorg): Remove the definition.
995 (arm_split_constant): Update the way to access variable
998 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
1000 2014-04-23 Tom de Vries <tom@codesourcery.com>
1002 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
1004 2014-04-23 David Malcolm <dmalcolm@redhat.com>
1006 * is-a.h: Update comments to reflect the following changes to the
1007 "pointerness" of the API, making the template parameter match the
1008 return type, allowing use of is-a.h with typedefs of pointers.
1009 (is_a_helper::cast): Return a T rather then a pointer to a T, so
1010 that the return type matches the parameter to the is_a_helper.
1012 (dyn_cast): Likewise.
1014 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
1015 pointer from the is-a.h API.
1017 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
1018 (is_a_helper <cgraph_node *>::test): ...this, matching change to
1020 (is_a_helper <varpool_node>::test): Likewise, convert to...
1021 (is_a_helper <varpool_node *>::test): ...this.
1023 (varpool_first_variable): Update for removal of implicit pointer
1024 from the is-a.h API.
1025 (varpool_next_variable): Likewise.
1026 (varpool_first_static_initializer): Likewise.
1027 (varpool_next_static_initializer): Likewise.
1028 (varpool_first_defined_variable): Likewise.
1029 (varpool_next_defined_variable): Likewise.
1030 (cgraph_first_defined_function): Likewise.
1031 (cgraph_next_defined_function): Likewise.
1032 (cgraph_first_function): Likewise.
1033 (cgraph_next_function): Likewise.
1034 (cgraph_first_function_with_gimple_body): Likewise.
1035 (cgraph_next_function_with_gimple_body): Likewise.
1036 (cgraph_alias_target): Likewise.
1037 (varpool_alias_target): Likewise.
1038 (cgraph_function_or_thunk_node): Likewise.
1039 (varpool_variable_node): Likewise.
1040 (symtab_real_symbol_p): Likewise.
1041 * cgraphunit.c (referred_to_p): Likewise.
1042 (analyze_functions): Likewise.
1043 (handle_alias_pairs): Likewise.
1044 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1045 * gimple-ssa.h (gimple_vuse_op): Likewise.
1046 (gimple_vdef_op): Likewise.
1047 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
1048 * gimple.c (gimple_build_asm_1): Likewise.
1049 (gimple_build_try): Likewise.
1050 (gimple_build_resx): Likewise.
1051 (gimple_build_eh_dispatch): Likewise.
1052 (gimple_build_omp_for): Likewise.
1053 (gimple_omp_for_set_clauses): Likewise.
1055 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
1056 (is_a_helper <gimple_statement_asm *>::test): ...this.
1057 (is_a_helper <gimple_statement_bind>::test): Convert to...
1058 (is_a_helper <gimple_statement_bind *>::test): ...this.
1059 (is_a_helper <gimple_statement_call>::test): Convert to...
1060 (is_a_helper <gimple_statement_call *>::test): ...this.
1061 (is_a_helper <gimple_statement_catch>::test): Convert to...
1062 (is_a_helper <gimple_statement_catch *>::test): ...this.
1063 (is_a_helper <gimple_statement_resx>::test): Convert to...
1064 (is_a_helper <gimple_statement_resx *>::test): ...this.
1065 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
1066 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
1067 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
1068 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
1069 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
1070 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
1071 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
1072 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
1073 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
1074 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
1075 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
1076 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
1077 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
1078 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
1079 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
1080 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
1081 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
1082 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
1083 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
1084 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
1085 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
1086 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
1087 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
1088 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
1089 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
1090 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
1091 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
1092 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
1093 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
1094 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
1095 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
1096 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
1097 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
1098 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
1099 (is_a_helper <gimple_statement_phi>::test): Convert to...
1100 (is_a_helper <gimple_statement_phi *>::test): ...this.
1101 (is_a_helper <gimple_statement_transaction>::test): Convert to...
1102 (is_a_helper <gimple_statement_transaction *>::test): ...this.
1103 (is_a_helper <gimple_statement_try>::test): Convert to...
1104 (is_a_helper <gimple_statement_try *>::test): ...this.
1105 (is_a_helper <gimple_statement_wce>::test): Convert to...
1106 (is_a_helper <gimple_statement_wce *>::test): ...this.
1107 (is_a_helper <const gimple_statement_asm>::test): Convert to...
1108 (is_a_helper <const gimple_statement_asm *>::test): ...this.
1109 (is_a_helper <const gimple_statement_bind>::test): Convert to...
1110 (is_a_helper <const gimple_statement_bind *>::test): ...this.
1111 (is_a_helper <const gimple_statement_call>::test): Convert to...
1112 (is_a_helper <const gimple_statement_call *>::test): ...this.
1113 (is_a_helper <const gimple_statement_catch>::test): Convert to...
1114 (is_a_helper <const gimple_statement_catch *>::test): ...this.
1115 (is_a_helper <const gimple_statement_resx>::test): Convert to...
1116 (is_a_helper <const gimple_statement_resx *>::test): ...this.
1117 (is_a_helper <const gimple_statement_eh_dispatch>::test):
1119 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
1121 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
1123 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
1124 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
1126 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1128 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1130 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1132 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1134 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1135 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1137 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1138 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1140 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1141 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1142 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1143 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1145 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1146 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1148 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1149 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1151 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1152 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1154 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1155 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1157 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1158 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1160 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1161 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1163 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1164 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1165 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1166 (is_a_helper <const gimple_statement_transaction>::test): Convert
1168 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1169 (is_a_helper <const gimple_statement_with_ops>::test): Convert
1171 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
1172 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
1173 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
1174 (is_a_helper <const gimple_statement_with_memory_ops>::test):
1176 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
1178 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
1180 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
1182 (gimple_use_ops): Update for removal of implicit pointer from the
1184 (gimple_set_use_ops): Likewise.
1185 (gimple_vuse): Likewise.
1186 (gimple_vdef): Likewise.
1187 (gimple_vuse_ptr): Likewise.
1188 (gimple_vdef_ptr): Likewise.
1189 (gimple_set_vuse): Likewise.
1190 (gimple_set_vdef): Likewise.
1191 (gimple_omp_return_set_lhs): Likewise.
1192 (gimple_omp_return_lhs): Likewise.
1193 (gimple_omp_return_lhs_ptr): Likewise.
1194 (gimple_call_fntype): Likewise.
1195 (gimple_call_set_fntype): Likewise.
1196 (gimple_call_set_internal_fn): Likewise.
1197 (gimple_call_use_set): Likewise.
1198 (gimple_call_clobber_set): Likewise.
1199 (gimple_bind_vars): Likewise.
1200 (gimple_bind_set_vars): Likewise.
1201 (gimple_bind_body_ptr): Likewise.
1202 (gimple_bind_set_body): Likewise.
1203 (gimple_bind_add_stmt): Likewise.
1204 (gimple_bind_block): Likewise.
1205 (gimple_bind_set_block): Likewise.
1206 (gimple_asm_ninputs): Likewise.
1207 (gimple_asm_noutputs): Likewise.
1208 (gimple_asm_nclobbers): Likewise.
1209 (gimple_asm_nlabels): Likewise.
1210 (gimple_asm_input_op): Likewise.
1211 (gimple_asm_input_op_ptr): Likewise.
1212 (gimple_asm_output_op): Likewise.
1213 (gimple_asm_output_op_ptr): Likewise.
1214 (gimple_asm_set_output_op): Likewise.
1215 (gimple_asm_clobber_op): Likewise.
1216 (gimple_asm_set_clobber_op): Likewise.
1217 (gimple_asm_label_op): Likewise.
1218 (gimple_asm_set_label_op): Likewise.
1219 (gimple_asm_string): Likewise.
1220 (gimple_catch_types): Likewise.
1221 (gimple_catch_types_ptr): Likewise.
1222 (gimple_catch_handler_ptr): Likewise.
1223 (gimple_catch_set_types): Likewise.
1224 (gimple_catch_set_handler): Likewise.
1225 (gimple_eh_filter_types): Likewise.
1226 (gimple_eh_filter_types_ptr): Likewise.
1227 (gimple_eh_filter_failure_ptr): Likewise.
1228 (gimple_eh_filter_set_types): Likewise.
1229 (gimple_eh_filter_set_failure): Likewise.
1230 (gimple_eh_must_not_throw_fndecl): Likewise.
1231 (gimple_eh_must_not_throw_set_fndecl): Likewise.
1232 (gimple_eh_else_n_body_ptr): Likewise.
1233 (gimple_eh_else_e_body_ptr): Likewise.
1234 (gimple_eh_else_set_n_body): Likewise.
1235 (gimple_eh_else_set_e_body): Likewise.
1236 (gimple_try_eval_ptr): Likewise.
1237 (gimple_try_cleanup_ptr): Likewise.
1238 (gimple_try_set_eval): Likewise.
1239 (gimple_try_set_cleanup): Likewise.
1240 (gimple_wce_cleanup_ptr): Likewise.
1241 (gimple_wce_set_cleanup): Likewise.
1242 (gimple_phi_capacity): Likewise.
1243 (gimple_phi_num_args): Likewise.
1244 (gimple_phi_result): Likewise.
1245 (gimple_phi_result_ptr): Likewise.
1246 (gimple_phi_set_result): Likewise.
1247 (gimple_phi_arg): Likewise.
1248 (gimple_phi_set_arg): Likewise.
1249 (gimple_resx_region): Likewise.
1250 (gimple_resx_set_region): Likewise.
1251 (gimple_eh_dispatch_region): Likewise.
1252 (gimple_eh_dispatch_set_region): Likewise.
1253 (gimple_omp_critical_name): Likewise.
1254 (gimple_omp_critical_name_ptr): Likewise.
1255 (gimple_omp_critical_set_name): Likewise.
1256 (gimple_omp_for_clauses): Likewise.
1257 (gimple_omp_for_clauses_ptr): Likewise.
1258 (gimple_omp_for_set_clauses): Likewise.
1259 (gimple_omp_for_collapse): Likewise.
1260 (gimple_omp_for_index): Likewise.
1261 (gimple_omp_for_index_ptr): Likewise.
1262 (gimple_omp_for_set_index): Likewise.
1263 (gimple_omp_for_initial): Likewise.
1264 (gimple_omp_for_initial_ptr): Likewise.
1265 (gimple_omp_for_set_initial): Likewise.
1266 (gimple_omp_for_final): Likewise.
1267 (gimple_omp_for_final_ptr): Likewise.
1268 (gimple_omp_for_set_final): Likewise.
1269 (gimple_omp_for_incr): Likewise.
1270 (gimple_omp_for_incr_ptr): Likewise.
1271 (gimple_omp_for_set_incr): Likewise.
1272 (gimple_omp_for_pre_body_ptr): Likewise.
1273 (gimple_omp_for_set_pre_body): Likewise.
1274 (gimple_omp_parallel_clauses): Likewise.
1275 (gimple_omp_parallel_clauses_ptr): Likewise.
1276 (gimple_omp_parallel_set_clauses): Likewise.
1277 (gimple_omp_parallel_child_fn): Likewise.
1278 (gimple_omp_parallel_child_fn_ptr): Likewise.
1279 (gimple_omp_parallel_set_child_fn): Likewise.
1280 (gimple_omp_parallel_data_arg): Likewise.
1281 (gimple_omp_parallel_data_arg_ptr): Likewise.
1282 (gimple_omp_parallel_set_data_arg): Likewise.
1283 (gimple_omp_task_clauses): Likewise.
1284 (gimple_omp_task_clauses_ptr): Likewise.
1285 (gimple_omp_task_set_clauses): Likewise.
1286 (gimple_omp_task_child_fn): Likewise.
1287 (gimple_omp_task_child_fn_ptr): Likewise.
1288 (gimple_omp_task_set_child_fn): Likewise.
1289 (gimple_omp_task_data_arg): Likewise.
1290 (gimple_omp_task_data_arg_ptr): Likewise.
1291 (gimple_omp_task_set_data_arg): Likewise.
1292 (gimple_omp_taskreg_clauses): Likewise.
1293 (gimple_omp_taskreg_clauses_ptr): Likewise.
1294 (gimple_omp_taskreg_set_clauses): Likewise.
1295 (gimple_omp_taskreg_child_fn): Likewise.
1296 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1297 (gimple_omp_taskreg_set_child_fn): Likewise.
1298 (gimple_omp_taskreg_data_arg): Likewise.
1299 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1300 (gimple_omp_taskreg_set_data_arg): Likewise.
1301 (gimple_omp_task_copy_fn): Likewise.
1302 (gimple_omp_task_copy_fn_ptr): Likewise.
1303 (gimple_omp_task_set_copy_fn): Likewise.
1304 (gimple_omp_task_arg_size): Likewise.
1305 (gimple_omp_task_arg_size_ptr): Likewise.
1306 (gimple_omp_task_set_arg_size): Likewise.
1307 (gimple_omp_task_arg_align): Likewise.
1308 (gimple_omp_task_arg_align_ptr): Likewise.
1309 (gimple_omp_task_set_arg_align): Likewise.
1310 (gimple_omp_single_clauses): Likewise.
1311 (gimple_omp_single_clauses_ptr): Likewise.
1312 (gimple_omp_single_set_clauses): Likewise.
1313 (gimple_omp_target_clauses): Likewise.
1314 (gimple_omp_target_clauses_ptr): Likewise.
1315 (gimple_omp_target_set_clauses): Likewise.
1316 (gimple_omp_target_child_fn): Likewise.
1317 (gimple_omp_target_child_fn_ptr): Likewise.
1318 (gimple_omp_target_set_child_fn): Likewise.
1319 (gimple_omp_target_data_arg): Likewise.
1320 (gimple_omp_target_data_arg_ptr): Likewise.
1321 (gimple_omp_target_set_data_arg): Likewise.
1322 (gimple_omp_teams_clauses): Likewise.
1323 (gimple_omp_teams_clauses_ptr): Likewise.
1324 (gimple_omp_teams_set_clauses): Likewise.
1325 (gimple_omp_sections_clauses): Likewise.
1326 (gimple_omp_sections_clauses_ptr): Likewise.
1327 (gimple_omp_sections_set_clauses): Likewise.
1328 (gimple_omp_sections_control): Likewise.
1329 (gimple_omp_sections_control_ptr): Likewise.
1330 (gimple_omp_sections_set_control): Likewise.
1331 (gimple_omp_for_set_cond): Likewise.
1332 (gimple_omp_for_cond): Likewise.
1333 (gimple_omp_atomic_store_set_val): Likewise.
1334 (gimple_omp_atomic_store_val): Likewise.
1335 (gimple_omp_atomic_store_val_ptr): Likewise.
1336 (gimple_omp_atomic_load_set_lhs): Likewise.
1337 (gimple_omp_atomic_load_lhs): Likewise.
1338 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1339 (gimple_omp_atomic_load_set_rhs): Likewise.
1340 (gimple_omp_atomic_load_rhs): Likewise.
1341 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1342 (gimple_omp_continue_control_def): Likewise.
1343 (gimple_omp_continue_control_def_ptr): Likewise.
1344 (gimple_omp_continue_set_control_def): Likewise.
1345 (gimple_omp_continue_control_use): Likewise.
1346 (gimple_omp_continue_control_use_ptr): Likewise.
1347 (gimple_omp_continue_set_control_use): Likewise.
1348 (gimple_transaction_body_ptr): Likewise.
1349 (gimple_transaction_label): Likewise.
1350 (gimple_transaction_label_ptr): Likewise.
1351 (gimple_transaction_set_body): Likewise.
1352 (gimple_transaction_set_label): Likewise.
1354 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1355 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1356 * ipa-ref.c (ipa_record_reference): Likewise.
1357 * ipa-reference.c (analyze_function): Likewise.
1358 (ipa_reference_write_optimization_summary): Likewise.
1359 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1360 (address_taken_from_non_vtable_p): Likewise.
1361 (comdat_can_be_unshared_p_1): Likewise.
1362 * lto-cgraph.c (lto_output_ref): Likewise.
1363 (add_references): Likewise.
1364 (compute_ltrans_boundary): Likewise.
1365 (output_symtab): Likewise.
1366 (input_ref): Likewise.
1367 (input_cgraph_1): Likewise.
1368 (output_cgraph_opt_summary): Likewise.
1369 * lto-streamer-out.c (lto_output): Likewise.
1370 (output_symbol_p): Likewise.
1371 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1372 (lsei_start_function_in_partition): Likewise.
1373 (lsei_next_variable_in_partition): Likewise.
1374 (lsei_start_variable_in_partition): Likewise.
1375 * symtab.c (insert_to_assembler_name_hash): Likewise.
1376 (unlink_from_assembler_name_hash): Likewise.
1377 (symtab_unregister_node): Likewise.
1378 (symtab_remove_node): Likewise.
1379 (dump_symtab_node): Likewise.
1380 (verify_symtab_base): Likewise.
1381 (verify_symtab_node): Likewise.
1382 (symtab_make_decl_local): Likewise.
1383 (symtab_alias_ultimate_target): Likewise.
1384 (symtab_resolve_alias): Likewise.
1385 (symtab_get_symbol_partitioning_class): Likewise.
1386 * tree-phinodes.c (allocate_phi_node): Likewise.
1387 (reserve_phi_args_for_new_edge): Likewise.
1388 (remove_phi_args): Likewise.
1389 * varpool.c (varpool_node_for_asm): Likewise.
1390 (varpool_remove_unreferenced_decls): Likewise.
1392 2014-04-23 Jeff Law <law@redhat.com>
1394 PR tree-optimization/60902
1395 * tree-ssa-threadedge.c
1396 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1397 invalidate outputs from statements that do not produce useful
1398 outputs for threading.
1400 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1402 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1403 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1404 machine descriptions for Stack Smashing Protector.
1406 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1408 * aarch64.md (<optab>_rol<mode>3): New pattern.
1409 (<optab>_rolsi3_uxtw): Likewise.
1410 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1412 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1414 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1415 (arm_cortex_a12_tune): Likewise.
1417 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1419 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1421 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1423 * config/arm/arm.md (arm_rev16si2): New pattern.
1424 (arm_rev16si2_alt): Likewise.
1425 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1427 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1429 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1430 (rev16<mode>2_alt): Likewise.
1431 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1432 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1433 (aarch_rev16_shleft_mask_imm_p): Likewise.
1434 (aarch_rev16_p_1): Likewise.
1435 (aarch_rev16_p): Likewise.
1436 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1437 (aarch_rev16_shright_mask_imm_p): Likewise.
1438 (aarch_rev16_shleft_mask_imm_p): Likewise.
1440 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1442 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1443 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1445 (cortex_a53_extra_costs): Likewise.
1446 (cortex_a57_extra_costs): Likewise.
1447 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1448 (cortexa7_extra_costs): Likewise.
1449 (cortexa8_extra_costs): Likewise.
1450 (cortexa12_extra_costs): Likewise.
1451 (cortexa15_extra_costs): Likewise.
1452 (v7m_extra_costs): Likewise.
1453 (arm_new_rtx_costs): Handle BSWAP.
1455 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1457 * config/arm/arm.c (cortexa8_extra_costs): New table.
1458 (arm_cortex_a8_tune): New tuning struct.
1459 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1461 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1463 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1465 2014-04-23 Richard Biener <rguenther@suse.de>
1467 * Makefile.in (OBJS): Remove loop-unswitch.o.
1468 * loop-unswitch.c: Delete.
1469 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1470 * passes.def (pass_rtl_unswitch): Likewise.
1471 * loop-init.c (gate_rtl_unswitch): Likewise.
1472 (rtl_unswitch): Likewise.
1473 (pass_data_rtl_unswitch): Likewise.
1474 (pass_rtl_unswitch): Likewise.
1475 (make_pass_rtl_unswitch): Likewise.
1476 * rtl.h (reversed_condition): Likewise.
1477 (compare_and_jump_seq): Likewise.
1478 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1480 * loop-unroll.c (compare_and_jump_seq): Likewise.
1482 2014-04-23 Richard Biener <rguenther@suse.de>
1484 PR tree-optimization/60903
1485 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1486 commented code block.
1487 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1488 loop flags to newly created BBs and edges.
1490 2014-04-23 Nick Clifton <nickc@redhat.com>
1492 * config/msp430/msp430.c (msp430_handle_option): Move function
1494 (msp430_option_override): Simplify mcu and mcpu option handling.
1495 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1496 support for -mhwmult command line option.
1497 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1498 -mhwmult command line option.
1499 (msp430_hwmult_enabled): Delete.
1500 (msp43o_output_labelref): Add support for -mhwmult command line
1502 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1503 (umulsidi3): Likewise.
1504 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1505 (mcpu, mlarge, msmall): Likewise.
1506 (mhwmult): New option.
1507 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1509 (msp430_is_f5_mcu): Remove prototype.
1510 (msp430_use_f5_series_hwmult): Add prototype.
1511 * config/msp430/msp430-opts.h: New file.
1512 * common/config/msp430: New directory.
1513 * common/config/msp430/msp430-common.c: New file.
1514 * config.gcc (msp430): Remove target_has_targetm_common.
1515 * doc/invoke.texi: Document -mhwmult command line option.
1517 2014-04-23 Nick Clifton <nickc@redhat.com>
1519 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1520 default-manifest.o if it can be found in the search path.
1521 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1523 2014-04-23 Terry Guo <terry.guo@arm.com>
1525 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1527 2014-04-23 Richard Biener <rguenther@suse.de>
1530 * tree-inline.c (declare_return_variable): Use mark_addressable.
1532 2014-04-23 Richard Biener <rguenther@suse.de>
1535 * loop-init.c (loop_optimizer_init): Make sure to apply
1536 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1538 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1541 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1543 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1544 if flag_sanitize_undefined_trap_on_error.
1545 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1546 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1547 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1548 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1549 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1550 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1551 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1552 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1553 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1554 * ubsan.c (ubsan_instrument_unreachable): Return
1555 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1556 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1557 if flag_sanitize_undefined_trap_on_error and
1558 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1559 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1560 instrument_bool_enum_load): Emit __builtin_trap () if
1561 flag_sanitize_undefined_trap_on_error and
1562 __builtin_handle_*_abort () if !flag_sanitize_recover.
1563 * doc/invoke.texi (-fsanitize-recover,
1564 -fsanitize-undefined-trap-on-error): Document.
1566 2014-04-22 Christian Bruel <christian.bruel@st.com>
1568 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1569 Force immediates to SImode.
1571 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1573 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1575 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1576 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1577 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1578 (nios2_fpu_insn): Add entry for round.
1579 (N2FPU_NO_ERRNO_P): Define.
1580 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1582 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1584 2014-04-22 Richard Henderson <rth@redhat.com>
1586 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1587 (add<GPI>3_compare0): Remove leading * from name.
1588 (add<GPI>3_carryin): Likewise.
1589 (sub<GPI>3_compare0): Likewise.
1590 (sub<GPI>3_carryin): Likewise.
1591 (<su_optab>mulditi3): New expander.
1592 (multi3): New expander.
1593 (madd<GPI>): Remove leading * from name.
1595 2014-04-22 Martin Jambor <mjambor@suse.cz>
1597 * cgraphclones.c (cgraph_function_versioning): Copy
1598 ipa_transforms_to_apply instead of asserting it is empty.
1600 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1603 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1604 on count_exp to get mode.
1606 2014-04-22 Andrew Pinski <apinski@cavium.com>
1608 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1609 Handle TLS for ILP32.
1610 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1611 (tlsie_small_<mode>): this and handle PTR.
1612 (tlsie_small_sidi): New pattern.
1613 (tlsle_small): Change to an expand to handle ILP32.
1614 (tlsle_small_<mode>): New pattern.
1615 (tlsdesc_small): Rename to ...
1616 (tlsdesc_small_<mode>): this and handle PTR.
1618 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1620 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1622 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1624 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1625 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1626 (aarch64_types_signed_poly_qualifiers): Likewise.
1627 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1628 (aarch64_types_poly_signed_qualifiers): Likewise.
1629 (TYPES_REINTERP_SS): Type macro added.
1630 (TYPES_REINTERP_SU): Likewise.
1631 (TYPES_REINTERP_SP): Likewise.
1632 (TYPES_REINTERP_US): Likewise.
1633 (TYPES_REINTERP_PS): Likewise.
1634 (aarch64_fold_builtin): New expression folding added.
1635 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1636 Declarations removed.
1637 (REINTERP_SS): Declarations added.
1638 (REINTERP_US): Likewise.
1639 (REINTERP_PS): Likewise.
1640 (REINTERP_SU): Likewise.
1641 (REINTERP_SP): Likewise.
1642 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1643 (vreinterpretq_p8_f64): Likewise.
1644 (vreinterpret_p16_f64): Likewise.
1645 (vreinterpretq_p16_f64): Likewise.
1646 (vreinterpret_f32_f64): Likewise.
1647 (vreinterpretq_f32_f64): Likewise.
1648 (vreinterpret_f64_f32): Likewise.
1649 (vreinterpret_f64_p8): Likewise.
1650 (vreinterpret_f64_p16): Likewise.
1651 (vreinterpret_f64_s8): Likewise.
1652 (vreinterpret_f64_s16): Likewise.
1653 (vreinterpret_f64_s32): Likewise.
1654 (vreinterpret_f64_s64): Likewise.
1655 (vreinterpret_f64_u8): Likewise.
1656 (vreinterpret_f64_u16): Likewise.
1657 (vreinterpret_f64_u32): Likewise.
1658 (vreinterpret_f64_u64): Likewise.
1659 (vreinterpretq_f64_f32): Likewise.
1660 (vreinterpretq_f64_p8): Likewise.
1661 (vreinterpretq_f64_p16): Likewise.
1662 (vreinterpretq_f64_s8): Likewise.
1663 (vreinterpretq_f64_s16): Likewise.
1664 (vreinterpretq_f64_s32): Likewise.
1665 (vreinterpretq_f64_s64): Likewise.
1666 (vreinterpretq_f64_u8): Likewise.
1667 (vreinterpretq_f64_u16): Likewise.
1668 (vreinterpretq_f64_u32): Likewise.
1669 (vreinterpretq_f64_u64): Likewise.
1670 (vreinterpret_s64_f64): Likewise.
1671 (vreinterpretq_s64_f64): Likewise.
1672 (vreinterpret_u64_f64): Likewise.
1673 (vreinterpretq_u64_f64): Likewise.
1674 (vreinterpret_s8_f64): Likewise.
1675 (vreinterpretq_s8_f64): Likewise.
1676 (vreinterpret_s16_f64): Likewise.
1677 (vreinterpretq_s16_f64): Likewise.
1678 (vreinterpret_s32_f64): Likewise.
1679 (vreinterpretq_s32_f64): Likewise.
1680 (vreinterpret_u8_f64): Likewise.
1681 (vreinterpretq_u8_f64): Likewise.
1682 (vreinterpret_u16_f64): Likewise.
1683 (vreinterpretq_u16_f64): Likewise.
1684 (vreinterpret_u32_f64): Likewise.
1685 (vreinterpretq_u32_f64): Likewise.
1687 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1689 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1690 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1691 (vreinterpret_p8_s8): Likewise.
1692 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1693 (vreinterpret_p8_s16): Likewise.
1694 (vreinterpret_p8_s32): Likewise.
1695 (vreinterpret_p8_s64): Likewise.
1696 (vreinterpret_p8_f32): Likewise.
1697 (vreinterpret_p8_u8): Likewise.
1698 (vreinterpret_p8_u16): Likewise.
1699 (vreinterpret_p8_u32): Likewise.
1700 (vreinterpret_p8_u64): Likewise.
1701 (vreinterpret_p8_p16): Likewise.
1702 (vreinterpretq_p8_s8): Likewise.
1703 (vreinterpretq_p8_s16): Likewise.
1704 (vreinterpretq_p8_s32): Likewise.
1705 (vreinterpretq_p8_s64): Likewise.
1706 (vreinterpretq_p8_f32): Likewise.
1707 (vreinterpretq_p8_u8): Likewise.
1708 (vreinterpretq_p8_u16): Likewise.
1709 (vreinterpretq_p8_u32): Likewise.
1710 (vreinterpretq_p8_u64): Likewise.
1711 (vreinterpretq_p8_p16): Likewise.
1712 (vreinterpret_p16_s8): Likewise.
1713 (vreinterpret_p16_s16): Likewise.
1714 (vreinterpret_p16_s32): Likewise.
1715 (vreinterpret_p16_s64): Likewise.
1716 (vreinterpret_p16_f32): Likewise.
1717 (vreinterpret_p16_u8): Likewise.
1718 (vreinterpret_p16_u16): Likewise.
1719 (vreinterpret_p16_u32): Likewise.
1720 (vreinterpret_p16_u64): Likewise.
1721 (vreinterpret_p16_p8): Likewise.
1722 (vreinterpretq_p16_s8): Likewise.
1723 (vreinterpretq_p16_s16): Likewise.
1724 (vreinterpretq_p16_s32): Likewise.
1725 (vreinterpretq_p16_s64): Likewise.
1726 (vreinterpretq_p16_f32): Likewise.
1727 (vreinterpretq_p16_u8): Likewise.
1728 (vreinterpretq_p16_u16): Likewise.
1729 (vreinterpretq_p16_u32): Likewise.
1730 (vreinterpretq_p16_u64): Likewise.
1731 (vreinterpretq_p16_p8): Likewise.
1732 (vreinterpret_f32_s8): Likewise.
1733 (vreinterpret_f32_s16): Likewise.
1734 (vreinterpret_f32_s32): Likewise.
1735 (vreinterpret_f32_s64): Likewise.
1736 (vreinterpret_f32_u8): Likewise.
1737 (vreinterpret_f32_u16): Likewise.
1738 (vreinterpret_f32_u32): Likewise.
1739 (vreinterpret_f32_u64): Likewise.
1740 (vreinterpret_f32_p8): Likewise.
1741 (vreinterpret_f32_p16): Likewise.
1742 (vreinterpretq_f32_s8): Likewise.
1743 (vreinterpretq_f32_s16): Likewise.
1744 (vreinterpretq_f32_s32): Likewise.
1745 (vreinterpretq_f32_s64): Likewise.
1746 (vreinterpretq_f32_u8): Likewise.
1747 (vreinterpretq_f32_u16): Likewise.
1748 (vreinterpretq_f32_u32): Likewise.
1749 (vreinterpretq_f32_u64): Likewise.
1750 (vreinterpretq_f32_p8): Likewise.
1751 (vreinterpretq_f32_p16): Likewise.
1752 (vreinterpret_s64_s8): Likewise.
1753 (vreinterpret_s64_s16): Likewise.
1754 (vreinterpret_s64_s32): Likewise.
1755 (vreinterpret_s64_f32): Likewise.
1756 (vreinterpret_s64_u8): Likewise.
1757 (vreinterpret_s64_u16): Likewise.
1758 (vreinterpret_s64_u32): Likewise.
1759 (vreinterpret_s64_u64): Likewise.
1760 (vreinterpret_s64_p8): Likewise.
1761 (vreinterpret_s64_p16): Likewise.
1762 (vreinterpretq_s64_s8): Likewise.
1763 (vreinterpretq_s64_s16): Likewise.
1764 (vreinterpretq_s64_s32): Likewise.
1765 (vreinterpretq_s64_f32): Likewise.
1766 (vreinterpretq_s64_u8): Likewise.
1767 (vreinterpretq_s64_u16): Likewise.
1768 (vreinterpretq_s64_u32): Likewise.
1769 (vreinterpretq_s64_u64): Likewise.
1770 (vreinterpretq_s64_p8): Likewise.
1771 (vreinterpretq_s64_p16): Likewise.
1772 (vreinterpret_u64_s8): Likewise.
1773 (vreinterpret_u64_s16): Likewise.
1774 (vreinterpret_u64_s32): Likewise.
1775 (vreinterpret_u64_s64): Likewise.
1776 (vreinterpret_u64_f32): Likewise.
1777 (vreinterpret_u64_u8): Likewise.
1778 (vreinterpret_u64_u16): Likewise.
1779 (vreinterpret_u64_u32): Likewise.
1780 (vreinterpret_u64_p8): Likewise.
1781 (vreinterpret_u64_p16): Likewise.
1782 (vreinterpretq_u64_s8): Likewise.
1783 (vreinterpretq_u64_s16): Likewise.
1784 (vreinterpretq_u64_s32): Likewise.
1785 (vreinterpretq_u64_s64): Likewise.
1786 (vreinterpretq_u64_f32): Likewise.
1787 (vreinterpretq_u64_u8): Likewise.
1788 (vreinterpretq_u64_u16): Likewise.
1789 (vreinterpretq_u64_u32): Likewise.
1790 (vreinterpretq_u64_p8): Likewise.
1791 (vreinterpretq_u64_p16): Likewise.
1792 (vreinterpret_s8_s16): Likewise.
1793 (vreinterpret_s8_s32): Likewise.
1794 (vreinterpret_s8_s64): Likewise.
1795 (vreinterpret_s8_f32): Likewise.
1796 (vreinterpret_s8_u8): Likewise.
1797 (vreinterpret_s8_u16): Likewise.
1798 (vreinterpret_s8_u32): Likewise.
1799 (vreinterpret_s8_u64): Likewise.
1800 (vreinterpret_s8_p8): Likewise.
1801 (vreinterpret_s8_p16): Likewise.
1802 (vreinterpretq_s8_s16): Likewise.
1803 (vreinterpretq_s8_s32): Likewise.
1804 (vreinterpretq_s8_s64): Likewise.
1805 (vreinterpretq_s8_f32): Likewise.
1806 (vreinterpretq_s8_u8): Likewise.
1807 (vreinterpretq_s8_u16): Likewise.
1808 (vreinterpretq_s8_u32): Likewise.
1809 (vreinterpretq_s8_u64): Likewise.
1810 (vreinterpretq_s8_p8): Likewise.
1811 (vreinterpretq_s8_p16): Likewise.
1812 (vreinterpret_s16_s8): Likewise.
1813 (vreinterpret_s16_s32): Likewise.
1814 (vreinterpret_s16_s64): Likewise.
1815 (vreinterpret_s16_f32): Likewise.
1816 (vreinterpret_s16_u8): Likewise.
1817 (vreinterpret_s16_u16): Likewise.
1818 (vreinterpret_s16_u32): Likewise.
1819 (vreinterpret_s16_u64): Likewise.
1820 (vreinterpret_s16_p8): Likewise.
1821 (vreinterpret_s16_p16): Likewise.
1822 (vreinterpretq_s16_s8): Likewise.
1823 (vreinterpretq_s16_s32): Likewise.
1824 (vreinterpretq_s16_s64): Likewise.
1825 (vreinterpretq_s16_f32): Likewise.
1826 (vreinterpretq_s16_u8): Likewise.
1827 (vreinterpretq_s16_u16): Likewise.
1828 (vreinterpretq_s16_u32): Likewise.
1829 (vreinterpretq_s16_u64): Likewise.
1830 (vreinterpretq_s16_p8): Likewise.
1831 (vreinterpretq_s16_p16): Likewise.
1832 (vreinterpret_s32_s8): Likewise.
1833 (vreinterpret_s32_s16): Likewise.
1834 (vreinterpret_s32_s64): Likewise.
1835 (vreinterpret_s32_f32): Likewise.
1836 (vreinterpret_s32_u8): Likewise.
1837 (vreinterpret_s32_u16): Likewise.
1838 (vreinterpret_s32_u32): Likewise.
1839 (vreinterpret_s32_u64): Likewise.
1840 (vreinterpret_s32_p8): Likewise.
1841 (vreinterpret_s32_p16): Likewise.
1842 (vreinterpretq_s32_s8): Likewise.
1843 (vreinterpretq_s32_s16): Likewise.
1844 (vreinterpretq_s32_s64): Likewise.
1845 (vreinterpretq_s32_f32): Likewise.
1846 (vreinterpretq_s32_u8): Likewise.
1847 (vreinterpretq_s32_u16): Likewise.
1848 (vreinterpretq_s32_u32): Likewise.
1849 (vreinterpretq_s32_u64): Likewise.
1850 (vreinterpretq_s32_p8): Likewise.
1851 (vreinterpretq_s32_p16): Likewise.
1852 (vreinterpret_u8_s8): Likewise.
1853 (vreinterpret_u8_s16): Likewise.
1854 (vreinterpret_u8_s32): Likewise.
1855 (vreinterpret_u8_s64): Likewise.
1856 (vreinterpret_u8_f32): Likewise.
1857 (vreinterpret_u8_u16): Likewise.
1858 (vreinterpret_u8_u32): Likewise.
1859 (vreinterpret_u8_u64): Likewise.
1860 (vreinterpret_u8_p8): Likewise.
1861 (vreinterpret_u8_p16): Likewise.
1862 (vreinterpretq_u8_s8): Likewise.
1863 (vreinterpretq_u8_s16): Likewise.
1864 (vreinterpretq_u8_s32): Likewise.
1865 (vreinterpretq_u8_s64): Likewise.
1866 (vreinterpretq_u8_f32): Likewise.
1867 (vreinterpretq_u8_u16): Likewise.
1868 (vreinterpretq_u8_u32): Likewise.
1869 (vreinterpretq_u8_u64): Likewise.
1870 (vreinterpretq_u8_p8): Likewise.
1871 (vreinterpretq_u8_p16): Likewise.
1872 (vreinterpret_u16_s8): Likewise.
1873 (vreinterpret_u16_s16): Likewise.
1874 (vreinterpret_u16_s32): Likewise.
1875 (vreinterpret_u16_s64): Likewise.
1876 (vreinterpret_u16_f32): Likewise.
1877 (vreinterpret_u16_u8): Likewise.
1878 (vreinterpret_u16_u32): Likewise.
1879 (vreinterpret_u16_u64): Likewise.
1880 (vreinterpret_u16_p8): Likewise.
1881 (vreinterpret_u16_p16): Likewise.
1882 (vreinterpretq_u16_s8): Likewise.
1883 (vreinterpretq_u16_s16): Likewise.
1884 (vreinterpretq_u16_s32): Likewise.
1885 (vreinterpretq_u16_s64): Likewise.
1886 (vreinterpretq_u16_f32): Likewise.
1887 (vreinterpretq_u16_u8): Likewise.
1888 (vreinterpretq_u16_u32): Likewise.
1889 (vreinterpretq_u16_u64): Likewise.
1890 (vreinterpretq_u16_p8): Likewise.
1891 (vreinterpretq_u16_p16): Likewise.
1892 (vreinterpret_u32_s8): Likewise.
1893 (vreinterpret_u32_s16): Likewise.
1894 (vreinterpret_u32_s32): Likewise.
1895 (vreinterpret_u32_s64): Likewise.
1896 (vreinterpret_u32_f32): Likewise.
1897 (vreinterpret_u32_u8): Likewise.
1898 (vreinterpret_u32_u16): Likewise.
1899 (vreinterpret_u32_u64): Likewise.
1900 (vreinterpret_u32_p8): Likewise.
1901 (vreinterpret_u32_p16): Likewise.
1902 (vreinterpretq_u32_s8): Likewise.
1903 (vreinterpretq_u32_s16): Likewise.
1904 (vreinterpretq_u32_s32): Likewise.
1905 (vreinterpretq_u32_s64): Likewise.
1906 (vreinterpretq_u32_f32): Likewise.
1907 (vreinterpretq_u32_u8): Likewise.
1908 (vreinterpretq_u32_u16): Likewise.
1909 (vreinterpretq_u32_u64): Likewise.
1910 (vreinterpretq_u32_p8): Likewise.
1911 (vreinterpretq_u32_p16): Likewise.
1913 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1915 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1917 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1920 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1921 (vqnegd_s64): Likewise.
1922 (vqabs_s64): Likewise.
1923 (vqabsd_s64): Likewise.
1925 2014-04-22 Richard Henderson <rth@redhat.com>
1927 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1928 computation to the top of the loop.
1930 2014-04-22 Renlin <renlin.li@arm.com>
1931 Jiong Wang <jiong.wang@arm.com>
1933 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1934 * config/aarch64/aarch64.c (aarch64_layout_frame)
1935 (aarch64_initial_elimination_offset): Likewise.
1937 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1939 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1942 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1944 * machmode.h (bitwise_mode_for_mode): Declare.
1945 * stor-layout.h (bitwise_type_for_mode): Likewise.
1946 * stor-layout.c (bitwise_mode_for_mode): New function.
1947 (bitwise_type_for_mode): Likewise.
1948 * builtins.c (fold_builtin_memory_op): Use it instead of
1949 int_mode_for_mode and build_nonstandard_integer_type.
1951 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1953 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1954 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1955 (*-*-solaris2*): Simplify.
1956 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1957 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1958 *-*-solaris2.9* handling.
1960 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1962 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1963 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1965 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1966 * configure: Regenerate.
1968 * config/i386/sol2-9.h: Remove.
1970 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1971 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1972 Remove Solaris 9 references.
1974 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1976 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1977 (floatuns<GPI:mode><GPF:mode>2): Remove.
1978 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1979 and floatuns conversions.
1980 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1981 and floatuns conversions.
1982 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1983 (w1,w2): New mode attributes for inequal width conversions.
1985 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1987 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1988 the output asm format.
1990 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1992 * config/aarch64/aarch64-simd.md
1993 (aarch64_cm<optab>di): Always split.
1994 (*aarch64_cm<optab>di): New.
1995 (aarch64_cmtstdi): Always split.
1996 (*aarch64_cmtstdi): New.
1998 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2000 PR tree-optimization/60823
2001 * omp-low.c (ipa_simd_modify_function_body): Go through
2002 all SSA_NAMEs and for those refering to vector arguments
2003 which are going to be replaced adjust SSA_NAME_VAR and,
2004 if it is a default definition, change it into a non-default
2005 definition assigned at the beginning of function from new_decl.
2006 (ipa_simd_modify_stmt_ops): Rewritten.
2007 * tree-dfa.c (set_ssa_default_def): When removing default def,
2008 check for NULL loc instead of NULL *loc.
2010 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2012 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
2013 restrictions on core registers for DImode values in Thumb2.
2015 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2017 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
2018 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
2020 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2022 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
2023 (*iordi_notzesidi_di): Likewise.
2024 (*iordi_notsesidi_di): Likewise.
2026 2014-04-22 Ian Bolton <ian.bolton@arm.com>
2028 * config/arm/arm-protos.h (tune_params): New struct members.
2029 * config/arm/arm.c: Initialise tune_params per processor.
2030 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
2031 for speed, based on new tune_params.
2033 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
2035 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
2037 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
2039 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
2041 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
2042 * config/aarch64/arm_neon.h (vrnd_f64): Added.
2043 (vrnda_f64): Likewise.
2044 (vrndi_f64): Likewise.
2045 (vrndm_f64): Likewise.
2046 (vrndn_f64): Likewise.
2047 (vrndp_f64): Likewise.
2048 (vrndx_f64): Likewise.
2050 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2052 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
2053 GET_MODE_SIZE argument is enum machine_mode.
2055 2014-04-22 Jakub Jelinek <jakub@redhat.com>
2058 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
2059 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
2061 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
2064 * asan.c (asan_emit_stack_protection): Force the base to align to
2065 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
2066 appropriate bits if STRICT_ALIGNMENT.
2067 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
2069 (expand_used_vars): Leave a space in the stack frame for alignment
2070 if STRICT_ALIGNMENT.
2072 2014-04-21 David Malcolm <dmalcolm@redhat.com>
2074 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
2076 (gimple_store_p): Likewise.
2077 (gimple_assign_load_p): Likewise.
2078 (gimple_assign_cast_p): Likewise.
2079 (gimple_clobber_p): Likewise.
2081 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
2082 rather than a gimple.
2083 (gimple_assign_cast_p): Likewise.
2085 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
2088 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
2089 If mode is DDmode and TARGET_E500_DOUBLE allow move.
2091 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
2092 more debug information for E500 if -mdebug=reg.
2094 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
2097 * config/i386/i386.c (ix86_expand_builtin)
2098 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
2099 register for target RTX.
2100 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
2102 2014-04-18 Cong Hou <congh@google.com>
2104 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
2105 the widen-mult pattern by handling two operands with different sizes,
2106 and operands whose size is smaller than half of the result type.
2108 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2110 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
2111 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
2112 (do_estimate_edge_time): Compute it.
2113 * ipa-inline.c (want_inline_small_function_p): Bypass
2114 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
2116 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2118 * ipa-inline.c (spec_rem): New static variable.
2119 (dump_overall_stats): New function.
2120 (dump_inline_stats): New function.
2122 2014-04-18 Richard Henderson <rth@redhat.com>
2124 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
2125 to GET_MODE_SIZE, not a reg_class_t.
2127 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2129 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2130 (vsx_xxmrglw_<mode>): Likewise.
2132 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2135 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2136 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2137 (rs6000_init_hard_regno_mode_ok): Likewise.
2139 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2141 * ipa-inline.c (inline_small_functions): Account only non-cold
2143 * doc/invoke.texi (inline-unit-growth): Update documentation.
2145 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2147 * config/rs6000/rs6000.md (addti3, subti3): New.
2149 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2152 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2153 comment. Check optimize_insn_for_size_p instead of
2154 optimize_insn_for_speed_p.
2156 2014-04-17 Martin Jambor <mjambor@suse.cz>
2158 * gimple-iterator.c (gsi_start_edge): New function.
2159 * gimple-iterator.h (gsi_start_edge): Declare.
2160 * tree-sra.c (single_non_eh_succ): New function.
2161 (disqualify_ops_if_throwing_stmt): Renamed to
2162 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2163 having one non-EH successor BB.
2164 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2165 generate loads into replacements.
2166 (sra_modify_assign): Likewise and and also use the simple path for
2168 (sra_modify_function_body): Commit statements on edges.
2170 2014-04-17 Richard Biener <rguenther@suse.de>
2173 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
2174 comparison results and add clarifying comment.
2176 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2178 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
2179 (blank_mode): Initialize it.
2180 (emit_mode_size_inline, emit_mode_nunits_inline,
2181 emit_mode_inner_inline): New functions.
2182 (emit_insn_modes_h): Call them and surround their output with
2183 #if GCC_VERSION >= 4001 ... #endif.
2184 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
2185 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
2186 mode_* arrays if the argument is __builtin_constant_p.
2187 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
2188 is enum machine_mode.
2190 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2192 * passes.c (opt_pass::execute): Adjust.
2193 (pass_manager::execute_pass_mode_switching): Likewise.
2194 (early_local_passes::execute): Likewise.
2195 (execute_one_pass): Pass cfun to the pass's execute method.
2196 * tree-pass.h (opt_pass::execute): Add function * argument.
2197 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2198 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2199 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2200 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2201 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2202 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
2203 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
2204 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2205 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2206 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
2207 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
2208 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
2209 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
2210 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
2211 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2212 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2213 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2214 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
2215 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
2216 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2217 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2218 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2219 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2220 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2221 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2222 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2223 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2224 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2225 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2228 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2230 * passes.c (opt_pass::gate): Take function * argument.
2231 (gate_all_early_local_passes): Merge into
2232 (early_local_passes::gate): this.
2233 (gate_all_early_optimizations): Merge into
2234 (all_early_optimizations::gate): this.
2235 (gate_all_optimizations): Mege into
2236 (all_optimizations::gate): this.
2237 (gate_all_optimizations_g): Merge into
2238 (all_optimizations_g::gate): this.
2239 (gate_rest_of_compilation): Mege into
2240 (rest_of_compilation::gate): this.
2241 (gate_postreload): Merge into
2242 (postreload::gate): this.
2243 (dump_one_pass): Pass cfun to the pass's gate method.
2244 (execute_ipa_summary_passes): Likewise.
2245 (execute_one_pass): Likewise.
2246 (ipa_write_summaries_2): Likewise.
2247 (ipa_write_optimization_summaries_1): Likewise.
2248 (ipa_read_summaries_1): Likewise.
2249 (ipa_read_optimization_summaries_1): Likewise.
2250 (execute_ipa_stmt_fixups): Likewise.
2251 * tree-pass.h (opt_pass::gate): Add function * argument.
2252 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
2253 combine-stack-adj.c, combine.c, compare-elim.c,
2254 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2255 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
2256 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
2257 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
2258 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2259 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2260 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2261 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
2262 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
2263 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
2264 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2265 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2266 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
2267 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2268 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2269 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2270 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2271 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2272 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2273 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2274 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2275 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2276 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2277 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
2278 var-tracking.c, vtable-verify.c, web.c: Adjust.
2280 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2282 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2283 * configure: Regenerate.
2285 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2287 * passes.c (dump_one_pass): don't check pass->has_gate.
2288 (execute_ipa_summary_passes): Likewise.
2289 (execute_one_pass): Likewise.
2290 (ipa_write_summaries_2): Likewise.
2291 (ipa_write_optimization_summaries_1): Likewise.
2292 (ipa_read_optimization_summaries_1): Likewise.
2293 (execute_ipa_stmt_fixups): Likewise.
2294 * tree-pass.h (pass_data::has_gate): Remove.
2295 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2296 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2297 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2298 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2299 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2300 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2301 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2302 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2303 gimple-low.c, gimple-ssa-isolate-paths.c,
2304 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2305 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2306 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2307 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2308 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2309 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2310 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2311 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2312 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2313 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2314 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2315 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2316 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2317 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2318 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2319 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2320 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2321 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2322 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2323 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2324 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2325 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2328 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2330 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2332 * passes.c (pass_manager::register_dump_files_1): Merge into
2333 (pass_manager::register_dump_files): this, and remove its handling of
2334 properties since the pass always has the properties anyway.
2335 (pass_manager::pass_manager): Adjust.
2337 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2339 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2340 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2341 dealing with properties.
2342 (pass_manager::register_dump_files): Adjust.
2344 2014-03-20 Mark Wielaard <mjw@redhat.com>
2346 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2347 then represent the bound as normal constant value.
2349 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2352 Forward port from 4.8 branch
2353 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2355 * config/i386/bmiintrin.h (_blsi_u32): New.
2359 (_blsmsk_u32): Ditto.
2360 (_blsmsk_u64): Ditto.
2361 (_tzcnt_u32): Ditto.
2362 (_tzcnt_u64): Ditto.
2364 2014-04-17 Kito Cheng <kito@0xlab.org>
2366 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2368 2014-04-17 Richard Biener <rguenther@suse.de>
2371 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2372 boolean results for comparisons.
2374 2014-04-17 Richard Biener <rguenther@suse.de>
2376 PR tree-optimization/60836
2377 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2378 initial PHI args to be gimple values.
2380 2014-04-17 Richard Biener <rguenther@suse.de>
2382 PR tree-optimization/60841
2383 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2384 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2385 of stmts to SLP build.
2386 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2387 (vect_analyze_slp): Likewise.
2388 (vect_analyze_slp_instance): Likewise.
2389 (vect_build_slp_tree): Limit overall SLP tree growth.
2390 * tree-vectorizer.h (vect_analyze_data_refs,
2391 vect_analyze_slp): Adjust prototypes.
2393 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2395 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2398 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2400 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2401 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2402 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2403 for TARGET_SLOW_PSHUFB
2405 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2407 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2408 * config/i386/i386.c (intel_cost): Ditto.
2410 2014-04-17 Joey Ye <joey.ye@arm.com>
2412 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2414 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2416 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2419 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2421 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2422 (type_all_derivations_known_p): New predicate.
2423 (type_all_ctors_visible_p): New predicate.
2424 (type_possibly_instantiated_p): New predicate.
2425 (get_odr_type): Compute all_derivations_known.
2426 (dump_odr_type): Dump the flag.
2427 (maybe_record_type): Cleanup.
2428 (record_target_from_binfo): Add bases_to_consider array;
2429 record bases for types w/o instances and skip CXX destructor.
2430 (possible_polymorphic_call_targets_1): Add bases_to_consider
2431 and consider_construction parameters; check if type may have instance.
2432 (get_polymorphic_call_info): Set maybe_in_construction to true
2433 when we know nothing.
2434 (record_targets_from_bases): Skip CXX destructors; they are
2435 never called for types in construction.
2436 (possible_polymorphic_call_targets): Do not record target when
2437 type may not have instance.
2439 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2442 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2443 external aliases alive, too.
2445 2014-04-16 Andrew Pinski <apinski@cavium.com>
2447 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2450 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2452 * final.c (compute_alignments): Do not apply loop alignment to a block
2453 falling through to the exit.
2455 2014-04-16 Catherine Moore <clm@codesourcery.com>
2457 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2458 Adjust constraints for microMIPS store patterns.
2460 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2462 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2464 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2466 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2467 (append_use): Run at -O0.
2468 (append_vdef): Likewise.
2469 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2470 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2472 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2474 PR tree-optimization/60844
2475 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2476 (propagate_op_to_single_use, remove_visited_stmt_chain,
2477 linearize_expr, repropagate_negates, reassociate_bb): Use it
2478 instead of gsi_remove.
2480 2014-04-16 Martin Jambor <mjambor@suse.cz>
2482 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2483 ipa_transforms_to_apply.
2484 (cgraph_function_versioning): Assert that old_node has empty
2485 ipa_transforms_to_apply.
2486 * trans-mem.c (ipa_tm_create_version): Likewise.
2487 * tree-inline.c (tree_function_versioning): Do not duplicate
2488 ipa_transforms_to_apply.
2490 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2493 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2495 Pass necessary as flags on 64-bit Solaris/x86.
2496 Use lowercase relocs for x86_64-*-*.
2497 * configure: Regenerate.
2499 2014-04-15 Jan Hubicka <jh@suse.cz>
2501 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2502 (maybe_record_node, likely_target_p): Use it.
2504 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2507 Revert following patch
2509 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2512 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2513 software floating point or no floating point registers, do not
2514 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2515 in GPRs that occurs after we tested for GPRs that would never be
2518 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2519 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2520 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2521 specifically allow DDmode, since that does not use the SPE SIMD
2524 2014-03-21 Mark Wielaard <mjw@redhat.com>
2526 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2527 as unsigned or int depending on type and value used.
2529 2014-04-15 Richard Biener <rguenther@suse.de>
2531 PR rtl-optimization/56965
2532 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2533 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2535 * alias.c (true_dependence_1): Do not call
2536 nonoverlapping_component_refs_p.
2537 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2538 nonoverlapping_component_refs_p.
2539 (indirect_refs_may_alias_p): Likewise.
2541 2014-04-15 Teresa Johnson <tejohnson@google.com>
2543 * cfg.c (dump_bb_info): Fix flags check.
2544 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2546 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2548 PR rtl-optimization/60663
2549 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2552 2014-04-15 Richard Biener <rguenther@suse.de>
2554 * lto-streamer.h (LTO_major_version): Bump to 4.
2556 2014-04-15 Richard Biener <rguenther@suse.de>
2558 * common.opt (lto_partition_model): New enum.
2559 (flto-partition=): Merge separate options with a single with argument,
2560 add -flto-partition=one support.
2561 * flag-types.h (enum lto_partition_model): Declare.
2562 * opts.c (finish_options): Remove duplicate -flto-partition=
2564 * lto-wrapper.c (run_gcc): Adjust.
2566 2014-04-15 Richard Biener <rguenther@suse.de>
2568 * alias.c (ncr_compar): New function.
2569 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2571 2014-04-15 Richard Biener <rguenther@suse.de>
2573 * alias.c (record_component_aliases): Do not walk BINFOs.
2575 2014-04-15 Richard Biener <rguenther@suse.de>
2577 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2578 Add struct function argument and adjust.
2579 (find_func_aliases_for_call): Likewise.
2580 (find_func_aliases): Likewise.
2581 (find_func_clobbers): Likewise.
2582 (intra_create_variable_infos): Likewise.
2583 (compute_points_to_sets): Likewise.
2584 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2586 2014-04-15 Richard Biener <rguenther@suse.de>
2588 * tree.c (iterative_hash_expr): Use enum tree_code_class
2589 to store TREE_CODE_CLASS.
2590 (tree_block): Likewise.
2591 (tree_set_block): Likewise.
2592 * tree.h (fold_build_pointer_plus_loc): Use
2593 convert_to_ptrofftype_loc.
2595 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2598 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2601 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2603 * cfgloop.h (struct loop): Move force_vectorize down.
2604 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2605 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2606 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2607 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2608 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2609 * tree-core.h (enum annot_expr_kind): Add new kind values.
2610 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2611 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2613 * tree.def (ANNOTATE_EXPR): Tweak comment.
2615 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2617 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2620 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2622 * tree.h (TYPE_IDENTIFIER): Declare.
2623 * tree.c (subrange_type_for_debug_p): Use it.
2624 * godump.c (go_format_type): Likewise.
2625 * dwarf2out.c (is_cxx_auto, modified_type_die,
2626 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2627 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2629 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2632 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2634 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2636 * config/i386/i386.c (examine_argument): Return bool. Return true if
2637 parameter should be passed in memory.
2638 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2639 (construct_container): Update calls to examine_argument.
2640 (function_arg_advance_64): Ditto.
2641 (return_in_memory_32): Merge with ix86_return_in_memory.
2642 (return_in_memory_64): Ditto.
2643 (return_in_memory_ms_64): Ditto.
2645 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2647 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2648 * coverage.c (coverage_compute_profile_id): Handle externally visible
2651 2014-04-14 Martin Jambor <mjambor@suse.cz>
2653 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2654 DECL_DISREGARD_INLINE_LIMITS functions.
2656 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2659 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2661 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2664 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2665 optimize_insn_for_speed_p instead of
2666 optimize_function_for_speed_p.
2668 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2670 * doc/invoke.texi (free): Document AArch64.
2672 2014-04-14 Richard Biener <rguenther@suse.de>
2674 PR tree-optimization/60042
2675 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2676 (insert_into_preds_of_block): Do not prevent PHI insertion
2677 for REFERENCE exprs here ...
2678 (eliminate_dom_walker::before_dom_children): ... but prevent
2679 their use here under similar conditions when applied to the
2680 IL after PRE optimizations.
2682 2014-04-14 Richard Biener <rguenther@suse.de>
2684 * passes.def: Move early points-to after early SRA.
2686 2014-04-14 Richard Biener <rguenther@suse.de>
2688 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2689 check for which sign-changes we allow when forwarding
2690 a converted value into a switch.
2692 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2694 * stor-layout.c (place_field): Finalize non-constant offset for the
2697 2014-04-14 Richard Biener <rguenther@suse.de>
2699 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2701 (expand_switch_using_bit_tests_p): Likewise.
2702 (process_switch): Compute and pass on speed_p based on the
2704 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2705 optimize_bb_for_speed_p.
2707 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2709 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2710 * function.h (struct function): Rename has_force_vect_loops into
2711 has_force_vectorize_loops.
2712 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2713 (input_struct_function_base): Likewise.
2714 * lto-streamer-out.c (output_cfg): Likewise.
2715 (output_struct_function_base): Likewise.
2716 * omp-low.c (expand_omp_simd): Likewise.
2717 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2718 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2719 (version_loop_for_if_conversion): Likewise.
2720 (tree_if_conversion): Likewise.
2721 (main_tree_if_conversion): Likewise.
2722 (gate_tree_if_conversion): Likewise.
2723 * tree-inline.c (copy_loops): Likewise.
2724 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2725 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2726 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2727 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2728 * tree-vectorizer.c (vectorize_loops): Likewise.
2729 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2731 2014-04-14 Richard Biener <rguenther@suse.de>
2734 * lto-streamer-out.c (wrap_refs): New function.
2735 (lto_output): Wrap symbol references in global initializes in
2736 type-preserving MEM_REFs.
2738 2014-04-14 Christian Bruel <christian.bruel@st.com>
2740 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2742 2014-04-14 Christian Bruel <christian.bruel@st.com>
2744 * config/sh/sh.md (setmemqi): New expand pattern.
2745 * config/sh/sh.h (CLEAR_RATIO): Define.
2746 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2747 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2749 2014-04-14 Richard Biener <rguenther@suse.de>
2752 * fold-const.c (negate_expr_p): Don't negate directional rounding
2754 (fold_negate_expr): Likewise.
2756 2014-04-14 Richard Biener <rguenther@suse.de>
2758 PR tree-optimization/59817
2759 PR tree-optimization/60453
2760 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2761 recursion to catch all CHRECs in the scalar evolution and restrict
2762 the predicate for the remains appropriately.
2764 2014-04-12 Catherine Moore <clm@codesourcery.com>
2766 * config/mips/constraints.md: Add new register constraint "kb".
2767 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2768 (*movhi_internal): Likewise.
2769 (*movqi_internal): Likewise.
2770 * config/mips/mips.h (M16_STORE_REGS): New register class.
2771 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2772 (REG_CLASS_CONTENTS): Likewise.
2773 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2775 2014-04-11 Tobias Burnus <burnus@net-b.de>
2778 * doc/invoke.texi (-Wformat-signedness): Document it.
2779 (Wformat=2): Mention that this enables -Wformat-signedness.
2781 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2783 * common/config/epiphany/epiphany-common.c
2784 (epiphany_option_optimization_table): Enable section anchors by
2785 default at -O1 or higher.
2786 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2787 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2788 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2789 carries no extra cost.
2790 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2791 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2792 * config/epiphany/predicates.md (memclob_operand): New predicate.
2793 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2794 Use memclob_operand predicate and X constraint for operand 3.
2796 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2798 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2799 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2802 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2804 PR rtl-optimization/60651
2805 * mode-switching.c (optimize_mode_switching): Make sure to emit
2806 sets of a lower numbered entity before sets of a higher numbered
2807 entity to a mode of the same or lower priority.
2808 When creating a seginfo for a basic block that starts with a code
2809 label, move the insertion point past the code label.
2810 (new_seginfo): Document and enforce requirement that
2811 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2812 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2813 * doc/tm.texi: Regenerate.
2815 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2818 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2820 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2822 * BASE-VER: Set to 4.10.0.
2824 2014-04-11 Tobias Burnus <burnus@net-b.de>
2827 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2828 * doc/gcc.texi (Service): Update description in the @menu
2829 * doc/invoke.texi (Option Summary): Remove misplaced and
2832 2014-04-11 Steve Ellcey <sellcey@mips.com>
2833 Jakub Jelinek <jakub@redhat.com>
2836 * expr.c (convert_move): Use emit_store_flag_force instead of
2837 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2840 2014-04-11 Richard Biener <rguenther@suse.de>
2843 * varasm.c (assemble_alias): Avoid endless error reporting
2844 recursion by setting TREE_ASM_WRITTEN.
2846 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2848 * config/s390/s390.md: Add a splitter for NOT rtx.
2850 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2852 PR rtl-optimization/60663
2853 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2855 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2856 Jakub Jelinek <jakub@redhat.com>
2859 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2860 flag from decl_node to node.
2862 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2865 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2866 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2867 ameliorating the cases where it can be.
2869 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2872 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2874 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2875 (loadsync_<mode>): Change mode.
2876 (load_quadpti, store_quadpti): New.
2877 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2878 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2879 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2881 2014-04-09 Cong Hou <congh@google.com>
2884 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2887 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2889 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2890 instead of vnor to exploit possible fusion opportunity in the
2892 (altivec_expand_vec_perm_const_le): Likewise.
2894 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2896 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2897 (loadsync_<mode>): Change mode.
2898 (load_quadpti, store_quadpti): New.
2899 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2900 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2902 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2905 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2906 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2907 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2909 2014-04-08 Richard Biener <rguenther@suse.de>
2912 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2913 a 64bit widest int print double-int similar to on HWI64 hosts.
2915 2014-04-08 Richard Biener <rguenther@suse.de>
2917 PR tree-optimization/60785
2918 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2919 default defs properly.
2921 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2923 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2924 (Weffc++): Likewise.
2926 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2928 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2929 set completep to false rather than true.
2931 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2934 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2935 ARM_TARGET2_DWARF_FORMAT.
2937 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2940 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2941 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2944 2014-04-07 Richard Biener <rguenther@suse.de>
2946 PR tree-optimization/60766
2947 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2948 (may_eliminate_iv): Convert cand_value_at result to desired type.
2950 2014-04-07 Jason Merrill <jason@redhat.com>
2953 * common.opt (-fno-gnu-unique): Add.
2954 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2956 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2958 * haifa-sched.c: Fix outdated function reference and minor
2959 grammar errors in introductory comment.
2961 2014-04-07 Richard Biener <rguenther@suse.de>
2964 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2966 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2968 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2971 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2972 size accounting for thunks.
2973 (pa_asm_output_mi_thunk): Use final_start_function() and
2974 final_end_function() to output function start and end directives.
2976 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2978 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2979 device specific ISA/ feature information. Remove short_sp and
2980 errata_skip ds. Add avr_device_specific_features enum to have device
2982 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2983 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2984 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2985 updated device specific info.
2986 * config/avr/avr-mcus.def: Merge device specific details to
2987 dev_attribute field.
2988 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2990 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2991 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2992 assembler if RMW isa supported by current device.
2993 * config/avr/genmultilib.awk: Update as device info structure changed.
2994 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2996 2014-04-04 Cong Hou <congh@google.com>
2998 PR tree-optimization/60656
2999 * tree-vect-stmts.c (supportable_widening_operation):
3000 Fix a bug that elements in a vector with vect_used_by_reduction
3001 property are incorrectly reordered when the operation on it is not
3002 consistant with the one in reduction operation.
3004 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
3006 PR rtl-optimization/60155
3007 * gcse.c (record_set_data): New function.
3008 (single_set_gcse): New function.
3009 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
3010 (hoist_code): Likewise.
3011 (get_pressure_class_and_nregs): Likewise.
3013 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
3015 * explow.c (probe_stack_range): Emit a final optimization blockage.
3017 2014-04-04 Anthony Green <green@moxielogic.com>
3019 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
3022 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
3025 * lto-cgraph.c (input_overwrite_node): Check that partitioning
3026 flags are set only during streaming.
3027 * ipa.c (process_references, walk_polymorphic_call_targets,
3028 symtab_remove_unreachable_nodes): Drop bodies of always inline
3029 after early inlining.
3030 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
3032 2014-04-04 Jakub Jelinek <jakub@redhat.com>
3033 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3036 * dwarf2out.c (const_ok_for_output_1): Reject expressions
3039 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3042 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
3044 (cortex_a53_fdivd): Likewise.
3046 2014-04-04 Martin Jambor <mjambor@suse.cz>
3049 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
3051 * cgraph.c (clone_of_p): Also return true if thunks match.
3052 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
3053 cgraph_function_or_thunk_node and an obsolete comment.
3054 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
3056 (build_function_decl_skip_args): Likewise.
3057 (set_new_clone_decl_and_node_flags): New function.
3058 (duplicate_thunk_for_node): Likewise.
3059 (redirect_edge_duplicating_thunks): Likewise.
3060 (cgraph_clone_node): New parameter args_to_skip, pass it to
3061 redirect_edge_duplicating_thunks which is called instead of
3062 cgraph_redirect_edge_callee.
3063 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
3064 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
3066 2014-04-04 Jeff Law <law@redhat.com>
3069 * config/arm/predicates.md (const_int_I_operand): New predicate.
3070 (const_int_M_operand): Similarly.
3071 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
3073 (insv_t2, extv_reg, extzv_t2): Likewise.
3074 (load_multiple_with_writeback): Similarly for const_int_I_operand.
3075 (pop_multiple_with_writeback_and_return): Likewise.
3076 (vfp_pop_multiple_with_writeback): Likewise
3078 2014-04-04 Richard Biener <rguenther@suse.de>
3081 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
3082 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
3084 * gimplify.h (gimple_add_tmp_var_fn): Declare.
3085 * gimplify.c (gimple_add_tmp_var_fn): New function.
3086 * gimple-expr.h (create_tmp_reg_fn): Declare.
3087 * gimple-expr.c (create_tmp_reg_fn): New function.
3088 * gimple-low.c (record_vars_into): Don't change cfun.
3089 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
3090 code generation without cfun.
3092 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
3095 * Makefile.in (install-driver): Fix shell scripting.
3097 2014-04-03 Cong Hou <congh@google.com>
3099 PR tree-optimization/60505
3100 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
3101 threshold of number of iterations below which no vectorization
3103 * tree-vect-loop.c (new_loop_vec_info):
3104 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
3105 * tree-vect-loop.c (vect_analyze_loop_operations):
3106 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
3107 * tree-vect-loop.c (vect_transform_loop):
3108 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
3109 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
3110 of iterations of the loop and see if we should build the epilogue.
3112 2014-04-03 Richard Biener <rguenther@suse.de>
3114 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
3115 (streamer_tree_cache_create): Adjust.
3116 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
3117 to allow optional nodes array.
3118 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
3119 (streamer_tree_cache_append): Likewise.
3120 (streamer_tree_cache_create): Create nodes array optionally
3121 as specified by parameter.
3122 * lto-streamer-out.c (create_output_block): Avoid maintaining
3123 the node array in the writer cache.
3124 (DFS_write_tree): Remove assertion.
3125 (produce_asm_for_decls): Free the out decl state hash table early.
3126 * lto-streamer-in.c (lto_data_in_create): Adjust for
3127 streamer_tree_cache_create prototype change.
3129 2014-04-03 Richard Biener <rguenther@suse.de>
3131 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3132 set TREE_CHAIN to NULL_TREE.
3134 2014-04-03 Richard Biener <rguenther@suse.de>
3136 PR tree-optimization/60740
3137 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3138 over all GIMPLE_COND operands.
3140 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3142 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3143 (Weffc++): Remove Scott's numbering, merge lists and reference
3146 2014-04-03 Nick Clifton <nickc@redhat.com>
3148 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3151 2014-04-03 Martin Jambor <mjambor@suse.cz>
3153 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3154 mention gcc_unreachable before failing.
3155 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3158 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3161 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3162 inconsistent code and instead mark the context inconsistent.
3163 (possible_polymorphic_call_targets): For inconsistent contexts
3164 return empty complete list.
3166 2014-04-02 Anthony Green <green@moxielogic.com>
3168 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3169 (extendqisi2, extendhisi2): Define.
3170 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
3171 (WCHAR_TYPE): Change to unsigned int.
3173 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3175 PR tree-optimization/60733
3176 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
3177 insertion point for PHI candidates to be the end of the feeding
3178 block for the PHI argument.
3180 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
3182 PR rtl-optimization/60650
3183 * lra-constraints.c (process_alt_operands): Decrease reject for
3184 earlyclobber matching.
3186 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3188 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
3190 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3192 * config/spu/spu.c (pad_bb): Do not crash when the last
3193 insn is CODE_FOR_blockage.
3195 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3197 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
3198 lies outside the target mode.
3200 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3203 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3204 software floating point or no floating point registers, do not
3205 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3206 in GPRs that occurs after we tested for GPRs that would never be
3209 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3210 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3211 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3212 specifically allow DDmode, since that does not use the SPE SIMD
3215 2014-04-02 Richard Biener <rguenther@suse.de>
3218 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
3219 MODE_INTs. Properly use negv_optab.
3220 (expand_abs): Likewise.
3222 2014-04-02 Richard Biener <rguenther@suse.de>
3225 * Makefile.in (install-driver): Guard extra installs with special
3228 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
3230 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3233 2014-04-01 Richard Henderson <rth@redhat.com>
3236 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
3237 alternative enabled before register allocation.
3239 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
3241 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
3242 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
3244 (nios2_large_got_address): Remove unneeded 'sym' parameter.
3245 (nios2_got_address): Update nios2_large_got_address call site.
3246 (nios2_delegitimize_address): New function.
3247 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
3248 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
3249 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
3251 2014-04-01 Martin Husemann <martin@duskware.de>
3253 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
3256 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
3258 PR rtl-optimization/60604
3259 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
3260 check from register_operand.
3261 (register_operand): Redefine in terms of general_operand.
3262 (nonmemory_operand): Use register_operand for the non-constant cases.
3264 2014-04-01 Richard Biener <rguenther@suse.de>
3266 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
3268 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
3270 * doc/invoke.texi (mapp-regs): Clarify.
3272 2014-03-31 Ulrich Drepper <drepper@gmail.com>
3274 * config/i386/avx512fintrin.h (__v32hi): Define type.
3275 (__v64qi): Likewise.
3276 (_mm512_set1_epi8): Define.
3277 (_mm512_set1_epi16): Define.
3278 (_mm512_set4_epi32): Define.
3279 (_mm512_set4_epi64): Define.
3280 (_mm512_set4_pd): Define.
3281 (_mm512_set4_ps): Define.
3282 (_mm512_setr4_epi64): Define.
3283 (_mm512_setr4_epi32): Define.
3284 (_mm512_setr4_pd): Define.
3285 (_mm512_setr4_ps): Define.
3286 (_mm512_setzero_epi32): Define.
3288 2014-03-31 Martin Jambor <mjambor@suse.cz>
3291 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3292 callsite_arguments_match_p. Updated all callers. Also check types of
3293 corresponding formal parameters and actual arguments.
3294 (not_all_callers_have_enough_arguments_p) Renamed to
3295 some_callers_have_mismatched_arguments_p.
3297 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3299 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3301 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3304 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3307 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3309 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3310 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3312 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3313 Use FMAMODE_NOVF512 mode iterator.
3314 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3315 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3316 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3317 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3319 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3320 Use VF_128_256 mode iterator.
3321 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3324 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3326 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3327 static chain if needed.
3329 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3332 * lra-constraints.c (index_part_to_reg): New.
3333 (process_address): Use it.
3335 2014-03-27 Jeff Law <law@redhat.com>
3336 Jakub Jelinek <jakub@redhat.com>
3339 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3340 gen_rtx_{PLUS,MULT} to build up the address expression.
3342 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3343 creating non-canonical RTL.
3345 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3348 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3349 functions; reorganize to make cheap checks first.
3350 (inline_small_functions): Do not estimate growth when dumping;
3352 * ipa-inline.h (inline_summary): Add min_size.
3353 (growth_likely_positive): New function.
3354 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3355 (set_cond_stmt_execution_predicate): Cleanup.
3356 (estimate_edge_size_and_time): Compute min_size.
3357 (estimate_calls_size_and_time): Likewise.
3358 (estimate_node_size_and_time): Likewise.
3359 (inline_update_overall_summary): Update min_size.
3360 (do_estimate_edge_time): Likewise.
3361 (do_estimate_edge_size): Update.
3362 (do_estimate_edge_hints): Update.
3363 (growth_likely_positive): New function.
3365 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3368 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3369 also if addr has VOIDmode.
3371 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3373 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3374 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3376 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3377 instructions as well as AdvancedSIMD loads.
3379 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3381 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3382 Use crypto_aese type.
3383 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3384 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3385 crypto_aese, crypto_aesmc. Move to types.md.
3386 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3388 * config/arm/iterators.md (crypto_type): Likewise.
3390 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3392 * cgraph.c: Include expr.h and tree-dfa.h.
3393 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3396 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3399 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3400 regs from checking multi-reg pseudos.
3402 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3404 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3406 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3408 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3409 if it would clobber the stack pointer, even temporarily.
3411 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3413 * mode-switching.c: Make small adjustments to the top comment.
3415 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3417 * config/rs6000/constraints.md (wD constraint): New constraint to
3418 match the constant integer to get the top DImode/DFmode out of a
3419 vector in a VSX register.
3421 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3422 match the constant integer to get the top DImode/DFmode out of a
3423 vector in a VSX register.
3425 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3428 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3431 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3432 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3434 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3435 Optimize vec_extract of 64-bit values, where the value being
3436 extracted is in the top word, where we can use scalar
3437 instructions. Add direct move and store support. Combine the big
3438 endian/little endian vector select load support into a single insn.
3439 (vsx_extract_<mode>_internal1): Likewise.
3440 (vsx_extract_<mode>_internal2): Likewise.
3441 (vsx_extract_<mode>_load): Likewise.
3442 (vsx_extract_<mode>_store): Likewise.
3443 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3444 combined into vsx_extract_<mode>_load.
3445 (vsx_extract_<mode>_one_le): Likewise.
3447 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3448 define the top 64-bit vector element.
3450 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3453 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3454 Document vec_vbpermq builtin.
3457 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3458 enable use of xxsldwi and xxpermdi builtin functions.
3459 (vec_xxpermdi): Likewise.
3461 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3462 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3464 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3466 PR rtl-optimization/60650
3467 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3469 (find_spills_for): New.
3470 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3471 Spill all pseudos on the second iteration.
3473 2014-03-27 Marek Polacek <polacek@redhat.com>
3476 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3479 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3481 * config/s390/s390.c (s390_can_use_return_insn): Check for
3482 call-saved FPRs on 31 bit.
3484 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3487 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3488 if they need regimplification, just drop them instead of
3489 calling gimple_regimplify_operands on them.
3491 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3494 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3495 (aarch64_frame_pointer_required): Adjust logic.
3496 (aarch64_can_eliminate): Adjust logic.
3497 (aarch64_override_options_after_change): Adjust logic.
3499 2014-03-27 Dehao Chen <dehao@google.com>
3501 * ipa-inline.c (early_inliner): Update node's inline info.
3503 2014-03-26 Dehao Chen <dehao@google.com>
3505 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3506 compiler inserted conditional jumps for NAN float check.
3508 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3510 * ubsan.h (ubsan_create_data): Change second argument's type
3511 to const location_t *.
3512 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3514 (ubsan_create_data): Change second argument to const location_t *PLOC.
3515 Create Loc field whenever PLOC is non-NULL.
3516 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3517 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3521 * real.c (real_to_integer2): Change type of low to UHWI.
3523 2014-03-26 Tobias Burnus <burnus@net-b.de>
3525 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3526 (CILK_SELF_SPECS): New define.
3527 (driver_self_specs): Use it.
3529 2014-03-26 Richard Biener <rguenther@suse.de>
3531 * tree-pretty-print.c (percent_K_format): Implement special
3532 case for LTO and its stripped down BLOCK tree.
3534 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3537 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3539 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3540 one range is range_int_cst_p, but not both, at least optimize
3541 addition/subtraction of 0 and multiplication by 0 or 1.
3542 * gimple-fold.c (gimple_fold_call): Fold
3543 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3544 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3545 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3547 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3549 PR rtl-optimization/60452
3550 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3551 <case REG>: Return 1 for invalid offsets from the frame pointer.
3553 2014-03-26 Marek Polacek <polacek@redhat.com>
3556 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3559 2014-03-26 Marek Polacek <polacek@redhat.com>
3562 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3565 2014-03-26 Marek Polacek <polacek@redhat.com>
3568 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3569 multiplication in unsigned type.
3571 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3573 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3575 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3577 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3579 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3582 * cif-code.def (UNREACHABLE) New code.
3583 * ipa-inline.c (inline_small_functions): Skip edges to
3584 __builtlin_unreachable.
3585 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3586 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3587 predicate to __bulitin_unreachable.
3588 (set_cond_stmt_execution_predicate): Fix issue when
3589 invert_tree_comparison returns ERROR_MARK.
3590 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3591 propagate to inline clones.
3592 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3594 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3595 * cgraphclones.c (cgraph_clone_node): If call destination is already
3596 ureachable, do not redirect it back.
3597 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3600 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3602 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3603 Do not modify inline clones.
3605 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3607 * config/i386/i386.md (general_sext_operand): New mode attr.
3608 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3609 don't generate (sign_extend (const_int)).
3610 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3611 operands[2]. Use We constraint instead of <i> and
3612 <general_sext_operand> predicate instead of <general_operand>.
3613 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3614 * config/i386/constraints.md (We): New constraint.
3615 * config/i386/predicates.md (x86_64_sext_operand,
3616 sext_operand): New predicates.
3618 2014-03-25 Martin Jambor <mjambor@suse.cz>
3621 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3622 inconsistent devirtualizations to __builtin_unreachable.
3624 2014-03-25 Marek Polacek <polacek@redhat.com>
3627 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3629 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3631 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3632 order of elements for big-endian.
3634 2014-03-25 Richard Biener <rguenther@suse.de>
3637 * gimplify-me.c (gimple_regimplify_operands): Update the
3640 2014-03-25 Martin Jambor <mjambor@suse.cz>
3643 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3644 (lto_output_varpool_node): Likewise.
3645 (input_overwrite_node): Likewise.
3646 (input_varpool_node): Likewise.
3648 2014-03-25 Richard Biener <rguenther@suse.de>
3650 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3651 (run_gcc): Likewise.
3653 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3655 * combine.c (simplify_compare_const): Add MODE argument.
3656 Handle mode_width 0 as very large mode_width.
3657 (try_combine, simplify_comparison): Adjust callers.
3659 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3660 type to avoid signed integer overflow.
3661 * explow.c (plus_constant): Likewise.
3663 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3665 * doc/generic.texi: Correct typos.
3667 2014-03-24 Tobias Burnus <burnus@net-b.de>
3669 * doc/invoke.texi (-flto): Expand section about
3670 using static libraries with LTO.
3672 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3674 PR rtl-optimization/60501
3675 * optabs.def (addptr3_optab): New optab.
3676 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3677 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3678 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3680 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3682 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3684 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3686 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3689 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3690 (_mm256_undefined_ps): Define.
3691 (_mm256_undefined_pd): Define.
3692 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3693 (_mm_undefined_pd): Define.
3694 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3695 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3696 (_mm512_undefined_ps): Define.
3697 (_mm512_undefined_pd): Define.
3698 Use _mm*_undefined_*.
3699 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3701 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3703 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3704 (lshr_simd): DI mode added.
3705 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3706 (aarch64_ushr_simddi): Likewise.
3707 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3708 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3709 (vshrd_n_u64): Likewise.
3711 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3713 * Makefile.in (s-macro_list): Depend on cc1.
3715 2014-03-23 Teresa Johnson <tejohnson@google.com>
3717 * ipa-utils.c (ipa_print_order): Use specified dump file.
3719 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3721 PR rtl-optimization/60601
3722 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3724 * gcc.c (eval_spec_function): Initialize save_growing_value.
3726 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3729 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3730 code == MINUS_EXPR, never swap op0 with op1.
3732 * toplev.c (init_local_tick): Avoid signed integer multiplication
3734 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3735 shift by first operand's bitsize.
3737 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3740 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3742 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3743 TARGET_ISA_64BIT_P(x).
3745 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3747 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3748 pattern for vector nor instead of subtract from splat(-1).
3749 (altivec_expand_vec_perm_const_le): Likewise.
3751 2014-03-21 Richard Henderson <rth@twiddle.net>
3754 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3755 related insns after epilogue_completed.
3757 2014-03-21 Martin Jambor <mjambor@suse.cz>
3760 * cgraph.h (symtab_node): New flag body_removed.
3761 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3762 when removing bodies.
3763 * symtab.c (dump_symtab_base): Dump body_removed flag.
3764 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3765 had their bodies removed.
3767 2014-03-21 Martin Jambor <mjambor@suse.cz>
3770 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3773 2014-03-21 Richard Biener <rguenther@suse.de>
3775 PR tree-optimization/60577
3776 * tree-core.h (struct tree_base): Document nothrow_flag use
3777 in VAR_DECL_NONALIASED.
3778 * tree.h (VAR_DECL_NONALIASED): New.
3779 (may_be_aliased): Adjust.
3780 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3782 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3784 * expr.c (expand_expr_real_1): Remove outdated comment.
3786 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3789 * ira.c (adjust_cleared_regs): Call copy_rtx on
3790 *reg_equiv[REGNO (loc)].src_p before passing it to
3791 simplify_replace_fn_rtx.
3794 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3795 into CONST, put pic register as first operand of PLUS. Use
3796 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3798 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3800 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3802 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3804 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3805 around for store forwarding issue in the FPU on the UT699.
3806 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3807 loads and operations if -mfix-ut699 is specified.
3808 (divtf3_hq): Tweak attribute.
3809 (sqrttf2_hq): Likewise.
3811 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3813 * calls.c (store_one_arg): Remove incorrect const qualification on the
3814 type of the temporary.
3815 * cfgexpand.c (expand_return): Likewise.
3816 * expr.c (expand_constructor): Likewise.
3817 (expand_expr_real_1): Likewise.
3819 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3821 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3824 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3827 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3829 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3831 * config/arm/aarch-common-protos.h
3832 (alu_cost_table): Fix spelling of "extend".
3833 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3835 2014-03-19 Richard Biener <rguenther@suse.de>
3838 * tree-core.h (tree_type_common): Re-order pointer members
3839 to reduce recursion depth during GC walks.
3841 2014-03-19 Marek Polacek <polacek@redhat.com>
3844 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3845 before accessing it.
3847 2014-03-19 Richard Biener <rguenther@suse.de>
3850 * lto-streamer-in.c (input_function): In WPA stage do not drop
3853 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3855 PR tree-optimization/60559
3856 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3857 with build_zero_cst assignment.
3859 2014-03-18 Kai Tietz <ktietz@redhat.com>
3861 PR rtl-optimization/56356
3862 * sdbout.c (sdbout_parms): Verify that parms'
3863 incoming argument is valid.
3864 (sdbout_reg_parms): Likewise.
3866 2014-03-18 Richard Henderson <rth@redhat.com>
3869 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3870 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3871 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3873 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3875 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3876 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3877 Italicize plugin event names in description. Explain that
3878 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3879 Remind that no GCC functions should be called after PLUGIN_FINISH.
3880 Explain what pragmas with expansion are.
3882 2014-03-18 Martin Liska <mliska@suse.cz>
3884 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3885 gimple call statement is update.
3886 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3887 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3889 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3892 * ubsan.c (ubsan_instrument_unreachable): Call
3893 initialize_sanitizer_builtins.
3894 (ubsan_pass): Likewise.
3897 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3898 varpool_finalize_decl instead of rest_of_decl_compilation.
3900 2014-03-18 Richard Biener <rguenther@suse.de>
3902 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3903 by using bitmap_and_compl instead of bitmap_and_compl_into.
3904 (df_rd_transfer_function): Likewise.
3906 2014-03-18 Richard Biener <rguenther@suse.de>
3908 * doc/lto.texi (fresolution): Fix typo.
3910 2014-03-18 Richard Biener <rguenther@suse.de>
3912 * doc/invoke.texi (flto): Update for changes in 4.9.
3914 2014-03-18 Richard Biener <rguenther@suse.de>
3916 * doc/loop.texi: Remove section on the removed lambda framework.
3917 Update loop docs with recent changes in preserving loop structure.
3919 2014-03-18 Richard Biener <rguenther@suse.de>
3921 * doc/lto.texi (-fresolution): Document.
3923 2014-03-18 Richard Biener <rguenther@suse.de>
3925 * doc/contrib.texi: Adjust my name.
3927 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3930 * internal-fn.c: Include diagnostic-core.h.
3931 (expand_BUILTIN_EXPECT): New function.
3932 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3933 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3934 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3935 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3937 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3938 Revert 3 argument __builtin_expect code.
3939 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3940 * gimple-fold.c (gimple_fold_call): Likewise.
3941 * tree.h (fold_builtin_expect): New prototype.
3942 * builtins.c (build_builtin_expect_predicate): Add predictor
3943 argument, if non-NULL, create 3 argument __builtin_expect.
3944 (fold_builtin_expect): No longer static. Add ARG2 argument,
3945 pass it through to build_builtin_expect_predicate.
3946 (fold_builtin_2): Adjust caller.
3947 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3948 * internal-fn.def (BUILTIN_EXPECT): New.
3950 2014-03-18 Tobias Burnus <burnus@net-b.de>
3953 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3954 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3955 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3957 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3960 * predict.c (combine_predictions_for_bb): Fix up formatting.
3961 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3962 fill what it points to if non-NULL.
3963 (tree_predict_by_opcode): Adjust caller, use the predictor.
3964 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3966 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3968 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3969 proper constant for the store mode.
3971 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3973 * symtab.c (change_decl_assembler_name): Fix transparent alias
3976 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3978 * config/aarch64/aarch64.c: Correct the comments about the
3979 aarch64 stack layout.
3981 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3983 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3984 check for GF_OMP_FOR_KIND_FOR.
3986 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3988 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3989 ymm and zmm register names.
3991 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3994 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3995 note creation for the 2010-08-31 changes.
3997 2014-03-17 Marek Polacek <polacek@redhat.com>
4000 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
4001 as -fno-tree-loop-vectorize.
4002 (expand_omp_simd): Likewise.
4004 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
4006 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
4007 (eligible_for_call_delay): New prototype.
4008 * config/sparc/sparc.c (tls_call_delay): Rename into...
4009 (eligible_for_call_delay): ...this. Return false if the instruction
4010 cannot be put in the delay slot of a branch.
4011 (eligible_for_restore_insn): Simplify.
4012 (eligible_for_return_delay): Return false if the instruction cannot be
4013 put in the delay slot of a branch and simplify.
4014 (eligible_for_sibcall_delay): Return false if the instruction cannot be
4015 put in the delay slot of a branch.
4016 * config/sparc/sparc.md (fix_ut699): New attribute.
4017 (tls_call_delay): Delete.
4018 (in_call_delay): Reimplement.
4019 (eligible_for_sibcall_delay): Rename into...
4020 (in_sibcall_delay): ...this.
4021 (eligible_for_return_delay): Rename into...
4022 (in_return_delay): ...this.
4023 (in_branch_delay): Reimplement.
4024 (in_uncond_branch_delay): Delete.
4025 (in_annul_branch_delay): Delete.
4027 2014-03-14 Richard Henderson <rth@redhat.com>
4030 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
4031 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
4032 (*floathi<X87MODEF>2_i387_with_temp): Remove.
4033 (floathi splitters): Remove.
4034 (float<SWI48x>xf2): New pattern.
4035 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
4036 code that tried to handle DImode for 32-bit, but which was excluded
4037 by the pattern's condition. Drop allocation of stack temporary.
4038 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
4039 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
4040 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
4041 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
4042 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
4043 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
4044 (*float<SWI48><MODEF>2_sse_interunit): Remove.
4045 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
4046 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
4047 (*float<SWI48x><X87MODEF>2_i387): Remove.
4048 (all float _with_temp splitters): Remove.
4049 (*float<SWI48x><MODEF>2_i387): New pattern.
4050 (*float<SWI48><MODEF>2_sse): New pattern.
4051 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
4052 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
4054 2014-03-14 Jakub Jelinek <jakub@redhat.com>
4055 Marek Polacek <polacek@redhat.com>
4058 * common.opt (dump_base_name_prefixed): New Variable.
4059 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
4060 if x_dump_base_name_prefixed is already set, set it at the end.
4062 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
4064 PR rtl-optimization/60508
4065 * lra-constraints.c (get_reload_reg): Add new parameter
4067 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
4068 Pass the new parameter values.
4070 2014-03-14 Richard Biener <rguenther@suse.de>
4072 * common.opt: Revert unintented changes from r205065.
4075 2014-03-14 Richard Biener <rguenther@suse.de>
4078 * cfghooks.c (split_block): Properly adjust all loops the
4079 block was a latch of.
4081 2014-03-14 Martin Jambor <mjambor@suse.cz>
4084 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
4087 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
4090 * config/avr/avr.c (avr_set_current_function): Pass function name
4091 through default_strip_name_encoding before sanity checking instead
4092 of skipping the first char of the assembler name.
4094 2014-03-13 Richard Henderson <rth@redhat.com>
4097 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
4098 (ix86_force_to_memory, ix86_free_from_memory): Remove.
4099 * config/i386/i386-protos.h: Likewise.
4100 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
4101 in the expander instead of a splitter.
4102 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
4103 any possibility of requiring a memory.
4104 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
4105 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
4106 (fp branch splitters): Update for ix86_split_fp_branch.
4107 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
4108 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
4109 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
4110 (*fop_<MODEF>_2_i387): Remove f/r alternative.
4111 (*fop_<MODEF>_3_i387): Likewise.
4112 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
4113 (splitters for the fop_* register patterns): Remove.
4114 (fscalexf4_i387): Rename from *fscalexf4_i387.
4115 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
4117 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4119 PR tree-optimization/59779
4120 * tree-dfa.c (get_ref_base_and_extent): Use double_int
4121 type for bitsize and maxsize instead of HOST_WIDE_INT.
4123 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
4125 PR rtl-optimization/57320
4126 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
4127 the CFG after thread_prologue_and_epilogue_insns.
4129 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4131 PR rtl-optimization/57189
4132 * lra-constraints.c (process_alt_operands): Disfavor spilling
4135 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4137 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4139 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4141 PR tree-optimization/59025
4143 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4144 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4146 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4149 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4150 calls of avr_out_plus_1.
4152 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4154 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4155 BB's single pred and update the father loop's latch info later.
4157 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4159 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4163 (VEC_base): Likewise.
4164 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4165 registers, we need to swap double words in little endian mode.
4167 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4168 to be a container mode for 128-bit integer operations added in ISA
4169 2.07. Unlike TImode and PTImode, the preferred register set is
4170 the Altivec/VMX registers for the 128-bit operations.
4172 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
4174 (rs6000_split_128bit_ok_p): Likewise.
4176 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
4177 macros for creating ISA 2.07 normal and overloaded builtin
4178 functions with 3 arguments.
4179 (BU_P8V_OVERLOAD_3): Likewise.
4180 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
4181 for use as overloaded functions.
4182 (VPERM_1TI_UNS): Likewise.
4183 (VSEL_1TI): Likewise.
4184 (VSEL_1TI_UNS): Likewise.
4185 (ST_INTERNAL_1ti): Likewise.
4186 (LD_INTERNAL_1ti): Likewise.
4187 (XXSEL_1TI): Likewise.
4188 (XXSEL_1TI_UNS): Likewise.
4189 (VPERM_1TI): Likewise.
4190 (VPERM_1TI_UNS): Likewise.
4191 (XXPERMDI_1TI): Likewise.
4192 (SET_1TI): Likewise.
4193 (LXVD2X_V1TI): Likewise.
4194 (STXVD2X_V1TI): Likewise.
4195 (VEC_INIT_V1TI): Likewise.
4196 (VEC_SET_V1TI): Likewise.
4197 (VEC_EXT_V1TI): Likewise.
4198 (EQV_V1TI): Likewise.
4199 (NAND_V1TI): Likewise.
4200 (ORC_V1TI): Likewise.
4201 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
4202 added in ISA 2.07. Add both normal 'altivec' builtins, and the
4204 (VADDUQM): Likewise.
4205 (VSUBCUQ): Likewise.
4206 (VADDEUQM): Likewise.
4207 (VADDECUQ): Likewise.
4208 (VSUBEUQM): Likewise.
4209 (VSUBECUQ): Likewise.
4211 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
4212 __int128_t and __uint128_t types.
4213 (__uint128_type): Likewise.
4214 (altivec_categorize_keyword): Add support for vector __int128_t,
4215 vector __uint128_t, vector __int128, and vector unsigned __int128
4216 as a container type for TImode operations that need to be done in
4217 VSX/Altivec registers.
4218 (rs6000_macro_to_expand): Likewise.
4219 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
4220 to support 128-bit integer instructions vaddcuq, vadduqm,
4221 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
4222 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
4224 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
4225 for V1TImode, and set up preferences to use VSX/Altivec registers.
4226 Setup VSX reload handlers.
4227 (rs6000_debug_reg_global): Likewise.
4228 (rs6000_init_hard_regno_mode_ok): Likewise.
4229 (rs6000_preferred_simd_mode): Likewise.
4230 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
4231 (easy_altivec_constant): Likewise.
4232 (output_vec_const_move): Likewise.
4233 (rs6000_expand_vector_set): Convert V1TImode set and extract to
4235 (rs6000_expand_vector_extract): Likewise.
4236 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
4238 (rs6000_const_vec): Add support for V1TImode.
4239 (rs6000_emit_le_vsx_load): Swap double words when loading or
4240 storing TImode/V1TImode.
4241 (rs6000_emit_le_vsx_store): Likewise.
4242 (rs6000_emit_le_vsx_move): Likewise.
4243 (rs6000_emit_move): Add support for V1TImode.
4244 (altivec_expand_ld_builtin): Likewise.
4245 (altivec_expand_st_builtin): Likewise.
4246 (altivec_expand_vec_init_builtin): Likewise.
4247 (altivec_expand_builtin): Likewise.
4248 (rs6000_init_builtins): Add support for V1TImode type. Add
4249 support for ISA 2.07 128-bit integer builtins. Define type names
4250 for the VSX/Altivec vector types.
4251 (altivec_init_builtins): Add support for overloaded vector
4252 functions with V1TImode type.
4253 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
4254 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
4256 (rs6000_split_128bit_ok_p): Likewise.
4257 (rs6000_handle_altivec_attribute): Create V1TImode from vector
4258 __int128_t and vector __uint128_t.
4260 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
4261 and mode attributes.
4268 (VS_scalar): Likewise.
4269 (VS_double): Likewise.
4270 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
4272 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
4273 we support the ISA 2.07 128-bit integer arithmetic instructions.
4274 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
4275 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
4276 and TImode types for use with the builtin functions.
4277 (V1TI_type_node): Likewise.
4278 (unsigned_V1TI_type_node): Likewise.
4279 (intTI_type_internal_node): Likewise.
4280 (uintTI_type_internal_node): Likewise.
4282 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4283 128-bit builtin functions.
4284 (UNSPEC_VADDEUQM): Likewise.
4285 (UNSPEC_VADDECUQ): Likewise.
4286 (UNSPEC_VSUBCUQ): Likewise.
4287 (UNSPEC_VSUBEUQM): Likewise.
4288 (UNSPEC_VSUBECUQ): Likewise.
4289 (VM): Add V1TImode to vector mode iterators.
4291 (VI_unit): Likewise.
4292 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4293 (altivec_vaddcuq): Likewise.
4294 (altivec_vsubuqm): Likewise.
4295 (altivec_vsubcuq): Likewise.
4296 (altivec_vaddeuqm): Likewise.
4297 (altivec_vaddecuq): Likewise.
4298 (altivec_vsubeuqm): Likewise.
4299 (altivec_vsubecuq): Likewise.
4301 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4303 (BOOL_128): Likewise.
4304 (BOOL_REGS_OUTPUT): Likewise.
4305 (BOOL_REGS_OP1): Likewise.
4306 (BOOL_REGS_OP2): Likewise.
4307 (BOOL_REGS_UNARY): Likewise.
4308 (BOOL_REGS_AND_CR0): Likewise.
4310 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4311 128-bit integer builtin support.
4312 (vec_vadduqm): Likewise.
4313 (vec_vaddecuq): Likewise.
4314 (vec_vaddeuqm): Likewise.
4315 (vec_vsubecuq): Likewise.
4316 (vec_vsubeuqm): Likewise.
4317 (vec_vsubcuq): Likewise.
4318 (vec_vsubuqm): Likewise.
4320 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4321 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4322 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4323 128-bit integer add/subtract to ISA 2.07.
4325 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4327 * config/arc/arc.c (arc_predicate_delay_insns):
4328 Fix third argument passed to conditionalize_nonjump.
4330 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4332 * config/aarch64/aarch64-builtins.c
4333 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4334 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4335 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4336 instead of __builtin_lfloor.
4337 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4339 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4341 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4342 (tree_ssa_ifcombine_bb_1): New function.
4343 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4344 is an empty forwarder block to then_bb or vice versa and then_bb
4345 and else_bb are effectively swapped.
4347 2014-03-12 Christian Bruel <christian.bruel@st.com>
4350 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4351 REG_CFA_DEF_CFA note.
4352 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4353 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4355 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4357 PR tree-optimization/60454
4358 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4360 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4362 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4363 Do not define target_cpu_default2 to generic.
4364 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4365 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4366 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4368 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4369 Marc Glisse <marc.glisse@inria.fr>
4371 PR tree-optimization/60502
4372 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4373 instead of build_low_bits_mask.
4375 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4378 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4379 if there are multiple uses, but op doesn't live on E edge.
4380 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4381 clobber stmts before __builtin_unreachable.
4383 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4385 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4386 hard_frame_pointer_rtx.
4387 * cse.c (cse_insn): Remove volatile check.
4388 * cselib.c (cselib_process_insn): Likewise.
4389 * dse.c (scan_insn): Likewise.
4391 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4393 * config/arc/arc.c (conditionalize_nonjump): New function,
4395 (arc_ifcvt): ... this.
4396 (arc_predicate_delay_insns): Use it.
4398 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4400 * config/arc/predicates.md (extend_operand): During/after reload,
4401 allow const_int_operand.
4402 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4403 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4404 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4406 (umulsi3_highpart_i): Likewise.
4408 2014-03-11 Richard Biener <rguenther@suse.de>
4410 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4411 Add asserts to guard possible wrong-code bugs.
4413 2014-03-11 Richard Biener <rguenther@suse.de>
4415 PR tree-optimization/60429
4416 PR tree-optimization/60485
4417 * tree-ssa-structalias.c (set_union_with_increment): Properly
4418 take into account all fields that overlap the shifted vars.
4419 (do_sd_constraint): Likewise.
4420 (do_ds_constraint): Likewise.
4421 (get_constraint_for_ptr_offset): Likewise.
4423 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4425 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4426 (nios2_compute_frame_layout):
4427 Add calculation of cfun->machine->fp_save_offset.
4428 (nios2_expand_prologue): Correct setting of frame pointer register
4430 (nios2_expand_epilogue): Update recovery of stack pointer from
4431 frame pointer accordingly.
4432 (nios2_initial_elimination_offset): Update calculation of offset
4433 for eliminating to HARD_FRAME_POINTER_REGNUM.
4435 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4438 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4439 cgraph_get_create_node on VAR_DECLs.
4441 2014-03-10 Richard Biener <rguenther@suse.de>
4444 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4446 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4448 * config/vms/vms.opt (vms_float_format): New variable.
4450 2014-03-08 Tobias Burnus <burnus@net-b.de>
4452 * doc/invoke.texi (-fcilkplus): Update implementation status.
4454 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4455 Richard Biener <rguenther@suse.de>
4457 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4458 consistently accross all TUs.
4459 (run_gcc): Enable -fshort-double automatically at link at link-time
4460 and disallow override.
4462 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4465 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4466 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4467 if they can't be used.
4469 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4471 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4472 for Solaris 11/x86 ld.
4473 * configure: Regenerate.
4475 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4477 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4478 (LIB_TLS_SPEC): Save as ld_tls_libs.
4479 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4480 (HAVE_AS_IX86_TLSLDM): New test.
4481 * configure, config.in: Regenerate.
4482 * config/i386/i386.c (legitimize_tls_address): Fall back to
4483 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4484 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4485 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4486 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4488 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4490 * common.opt (fira-loop-pressure): Mark as optimization.
4492 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4494 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4495 an OpenMP mappable type.
4497 2014-03-06 Matthias Klose <doko@ubuntu.com>
4499 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4500 MULTILIB_OSDIRNAMES is not defined.
4502 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4503 Meador Inge <meadori@codesourcery.com>
4506 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4507 (arm_legitimize_address): Call legitimize_tls_address for any
4508 arm_tls_referenced_p expression, handle constant addend. Call it
4509 before testing for !TARGET_ARM.
4510 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4512 2014-03-06 Richard Biener <rguenther@suse.de>
4518 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4520 * tree-streamer.c (record_common_node): Assert we don't record
4521 nodes with type double.
4522 (preload_common_node): Skip type double, complex double and double
4523 pointer since it is now frontend dependent due to fshort-double option.
4525 2014-03-06 Richard Biener <rguenther@suse.de>
4527 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4528 or -fno-lto is specified and the linker has full plugin support.
4529 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4530 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4531 * lto-wrapper.c (merge_and_complain): Merge compile-time
4532 optimization levels.
4533 (run_gcc): And pass it through to the link options.
4535 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4539 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4541 * cselib.c (remove_useless_values): Skip to avoid quadratic
4542 behavior if the condition moved from...
4543 (cselib_process_insn): ... here holds.
4545 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4548 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4549 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4552 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4553 (TM_H): Add x86-tune.def.
4555 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4557 * config/aarch64/aarch64.c (generic_tunings):
4558 Use cortexa57_extra_costs.
4560 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4563 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4564 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4565 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4568 2014-03-04 Heiher <r@hev.cc>
4570 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4571 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4573 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4575 * config/i386/predicates.md (const2356_operand): Change to ...
4576 (const2367_operand): ... this.
4577 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4579 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4580 (*avx512pf_scatterpf<mode>sf): Ditto.
4581 (avx512pf_scatterpf<mode>df): Ditto.
4582 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4583 (*avx512pf_scatterpf<mode>df): Ditto.
4584 * config/i386/i386.c (ix86_expand_builtin): Update
4585 incorrect hint operand error message.
4587 2014-03-04 Richard Biener <rguenther@suse.de>
4589 * lto-section-in.c (lto_get_section_data): Fix const cast.
4591 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4593 * tree-streamer.c (record_common_node): Assert we don't record
4594 nodes with type double.
4595 (preload_common_node): Skip type double, complex double and double
4596 pointer since it is now frontend dependent due to fshort-double option.
4598 2014-03-04 Richard Biener <rguenther@suse.de>
4601 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4602 (lto_input_toplevel_asms): Likewise.
4603 * lto-section-in.c (lto_get_section_data): Instead do it here
4606 2014-03-04 Richard Biener <rguenther@suse.de>
4608 PR tree-optimization/60382
4609 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4610 dead PHIs a reduction.
4612 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4614 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4616 (_mm_prefetch): Move out of GCC target("sse") pragma.
4617 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4618 GCC target("prfchw") pragma.
4619 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4621 * config/i386/i386.c (ix86_option_override_internal): Enable
4622 -mprfchw with -mprefetchwt1.
4624 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4626 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4629 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4631 * opts.h (CL_PCH_IGNORE): Define.
4632 * targhooks.c (option_affects_pch_p):
4633 Return false for options that have CL_PCH_IGNORE set.
4634 * opt-functions.awk: Process PchIgnore.
4635 * doc/options.texi: Document PchIgnore.
4637 * config/arc/arc.opt (misize): Add PchIgnore property.
4639 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4641 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4642 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4643 constraint on constants to permit them being loaded into
4644 GENERAL_REGS or BASE_REGS.
4646 2014-03-03 Nick Clifton <nickc@redhat.com>
4648 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4649 anti-cacnonical alternatives.
4650 (negandhi3_real): New pattern.
4651 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4653 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4655 * config/avr/avr-mcus.def: Remove atxmega16x1.
4656 * config/avr/avr-tables.opt: Regenerate.
4657 * config/avr/t-multilib: Regenerate.
4658 * doc/avr-mmcu.texi: Regenerate.
4660 2014-03-03 Tobias Grosser <tobias@grosser.es>
4661 Mircea Namolaru <mircea.namolaru@inria.fr>
4663 PR tree-optimization/58028
4664 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4667 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4669 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4670 not handled by recognizers.
4672 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4675 * function.c (expand_function_end): Don't emit
4676 clobber_return_register sequence if clobber_after is a BARRIER.
4677 * cfgexpand.c (construct_exit_block): Append instructions before
4678 return_label to prev_bb.
4680 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4682 * config/rs6000/constraints.md: Document reserved use of "wc".
4684 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4687 * ipa.c (function_and_variable_visibility): When dissolving comdat
4688 group, also set all symbols to local.
4690 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4695 2013-12-14 Jan Hubicka <jh@suse.cz>
4697 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4699 2014-03-02 Jon Beniston <jon@beniston.com>
4705 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4706 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4707 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4708 (simple_return, *simple_return): New patterns
4709 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4710 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4712 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4714 * dwarf2out.c (gen_subprogram_die): Tidy.
4716 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4719 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4720 (*mov_t_msb_neg_negc): ... this new insn.
4722 2014-02-28 Jason Merrill <jason@redhat.com>
4725 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4728 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4731 * dwarf2out.c (decltype_auto_die): New static.
4732 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4733 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4734 (is_cxx_auto): Likewise.
4736 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4738 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4739 we are not using general regs only.
4741 2014-02-28 Richard Biener <rguenther@suse.de>
4744 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4745 previous fix and only allow to remove trivial pre-headers
4746 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4747 (remove_forwarder_block): Properly update the latch of a loop.
4749 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4752 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4753 (cselib_preserved_hash_table): New.
4754 (preserve_constants_and_equivs): Move preserved vals to it.
4755 (cselib_find_slot): Look it up first.
4756 (cselib_init): Initialize it.
4757 (cselib_finish): Release it.
4758 (dump_cselib_table): Dump it.
4760 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4763 * cselib.c (remove_useless_values): Skip to avoid quadratic
4764 behavior if the condition moved from...
4765 (cselib_process_insn): ... here holds.
4767 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4770 * var-tracking.c (vt_initialize): Apply the same condition to
4771 preserve the CFA base value.
4773 2014-02-28 Joey Ye <joey.ye@arm.com>
4776 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4777 if reload in progress or completed.
4779 2014-02-28 Tobias Burnus <burnus@net-b.de>
4782 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4785 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4787 * doc/tm.texi.in (Condition Code Status): Update documention for
4788 relative locations of cc0-setter and cc0-user.
4790 2014-02-27 Jeff Law <law@redhat.com>
4792 PR rtl-optimization/52714
4793 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4794 into two independent simple sets, if I3 is a jump, ensure the
4795 pattern we place into I3 is a (set (pc) ...).
4797 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4798 Jeff Law <law@redhat.com>
4800 PR rtl-optimization/49847
4801 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4802 are in different blocks.
4803 * doc/tm.texi (Condition Code Status): Update documention for
4804 relative locations of cc0-setter and cc0-user.
4806 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4809 * lra.c (lra_emit_add): Check SUBREG too.
4811 2014-02-27 Andreas Schwab <schwab@suse.de>
4813 * config/m68k/m68k.c (m68k_option_override): Disable
4814 -flive-range-shrinkage for classic m68k.
4815 (m68k_override_options_after_change): Likewise.
4817 2014-02-27 Marek Polacek <polacek@redhat.com>
4820 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4821 -Wmaybe-uninitialized.
4823 2014-02-27 Alan Modra <amodra@gmail.com>
4826 * reload1.c (emit_input_reload_insns): When reload_override_in,
4827 set old to rl->in_reg when rl->in_reg is a subreg.
4829 2014-02-26 Richard Biener <rguenther@suse.de>
4832 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4834 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4836 * common/config/i386/predicates.md (const1256_operand): Remove.
4837 (const2356_operand): New.
4838 (const_1_to_2_operand): Remove.
4839 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4840 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4841 (*avx512pf_gatherpf<mode>sf): Ditto.
4842 (avx512pf_gatherpf<mode>df): Ditto.
4843 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4844 (*avx512pf_gatherpf<mode>df): Ditto.
4845 (avx512pf_scatterpf<mode>sf): Ditto.
4846 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4847 (*avx512pf_scatterpf<mode>sf): Ditto.
4848 (avx512pf_scatterpf<mode>df): Ditto.
4849 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4850 (*avx512pf_scatterpf<mode>df): Ditto.
4851 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4853 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4855 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4856 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4857 (_mm512_mask_testn_epi64_mask): Move to ...
4858 * config/i386/avx512cdintrin.h: Here.
4859 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4860 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4861 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4862 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4863 TARGET_AVX512F from TARGET_AVX512CD.
4865 2014-02-26 Richard Biener <rguenther@suse.de>
4868 * ipa.c (walk_polymorphic_call_targets): Properly guard
4869 call to inline_update_overall_summary.
4871 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4874 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4875 and latches only if requested. Fix latch if it is removed.
4876 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4877 LOOPS_HAVE_PREHEADERS.
4879 2014-02-25 Andrew Pinski <apinski@cavium.com>
4881 * builtins.c (expand_builtin_thread_pointer): Create a new target
4882 when the target is NULL.
4884 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4886 PR rtl-optimization/60317
4887 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4888 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4889 * lra-assigns.c: Include params.h.
4890 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4891 other reload pseudos considerations.
4893 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4895 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4896 to use canonical form for nor<mode>3.
4898 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4901 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4904 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4906 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4907 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4908 (ix86_handle_option): Handle OPT_mprefetchwt1.
4909 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4910 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4912 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4913 OPTION_MASK_ISA_PREFETCHWT1.
4914 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4915 (PTA_PREFETCHWT1): New.
4916 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4917 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4918 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4919 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4920 (*prefetch_avx512pf_<mode>_: Change into ...
4921 (*prefetch_prefetchwt1_<mode>: This.
4922 * config/i386/i386.opt (mprefetchwt1): New.
4923 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4924 (_mm_prefetch): Handle intent to write.
4925 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4927 2014-02-25 Richard Biener <rguenther@suse.de>
4930 * emit-rtl.c (mem_attrs_htab): Remove.
4931 (mem_attrs_htab_hash): Likewise.
4932 (mem_attrs_htab_eq): Likewise.
4933 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4934 (init_emit_once): Do not allocate mem_attrs_htab.
4936 2014-02-25 Richard Biener <rguenther@suse.de>
4939 * lto-opts.c (lto_write_options): Output non-explicit conservative
4940 -fwrapv, -fno-trapv and -fno-strict-overflow.
4941 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4942 (run_gcc): And pass them through.
4944 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4946 * sel-sched.c (calculate_new_fences): New parameter ptime.
4947 Calculate it as a maximum over all fence cycles.
4948 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4949 Print the final schedule timing when sched_verbose.
4951 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4953 PR rtl-optimization/60292
4954 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4955 bit fot the fence instruction.
4957 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4959 * calls.h: Fix typo in comment.
4961 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4963 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4964 adjusting offsetable addresses.
4966 2014-02-24 Guozhi Wei <carrot@google.com>
4968 * sparseset.h (sparseset_pop): Fix the wrong index.
4970 2014-02-24 Walter Lee <walt@tilera.com>
4972 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4973 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4975 * common/config/tilegx/tilegx-common.c
4976 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4977 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4979 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4980 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4981 (tilegx_gimplify_va_arg_expr): Handle big endian.
4982 (tilegx_expand_unaligned_load): Ditto.
4983 (tilegx_expand_unaligned_store): Ditto.
4984 (TARGET_RETURN_IN_MSB): New.
4985 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4986 (TARGET_ENDIAN_DEFAULT): New.
4987 (TARGET_BIG_ENDIAN): Handle big endian.
4988 (BYTES_BIG_ENDIAN): Ditto.
4989 (WORDS_BIG_ENDIAN): Ditto.
4990 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4993 * config/tilegx/tilegx.md (extv): Handle big endian.
4995 (insn_st<n>): Ditto.
4996 (insn_st<n>_add<bitsuffix>): Ditto.
4997 (insn_stnt<n>): Ditto.
4998 (insn_stnt<n>_add<bitsuffix>):Ditto.
4999 (vec_interleave_highv8qi): Handle big endian.
5000 (vec_interleave_highv8qi_be): New.
5001 (vec_interleave_highv8qi_le): New.
5002 (insn_v1int_h): Handle big endian.
5003 (vec_interleave_lowv8qi): Handle big endian.
5004 (vec_interleave_lowv8qi_be): New.
5005 (vec_interleave_lowv8qi_le): New.
5006 (insn_v1int_l): Handle big endian.
5007 (vec_interleave_highv4hi): Handle big endian.
5008 (vec_interleave_highv4hi_be): New.
5009 (vec_interleave_highv4hi_le): New.
5010 (insn_v2int_h): Handle big endian.
5011 (vec_interleave_lowv4hi): Handle big endian.
5012 (vec_interleave_lowv4hi_be): New.
5013 (vec_interleave_lowv4hi_le): New.
5014 (insn_v2int_l): Handle big endian.
5015 (vec_interleave_highv2si): Handle big endian.
5016 (vec_interleave_highv2si_be): New.
5017 (vec_interleave_highv2si_le): New.
5018 (insn_v4int_h): Handle big endian.
5019 (vec_interleave_lowv2si): Handle big endian.
5020 (vec_interleave_lowv2si_be): New.
5021 (vec_interleave_lowv2si_le): New.
5022 (insn_v4int_l): Handle big endian.
5023 * config/tilegx/tilegx.opt (mbig-endian): New option.
5024 (mlittle-endian): New option.
5025 * doc/install.texi: Document tilegxbe-linux.
5026 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
5028 2014-02-24 Martin Jambor <mjambor@suse.cz>
5031 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
5032 there are no parameter descriptors.
5034 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
5036 PR rtl-optimization/60268
5037 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
5038 initialization to ...
5039 (sched_rgn_init): ... here.
5040 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
5042 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5044 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
5047 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
5049 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
5052 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5054 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
5055 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
5057 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
5059 * config/microblaze/predicates.md: Add cmp_op predicate.
5060 * config/microblaze/microblaze.md: Add branch_compare instruction
5061 which uses cmp_op predicate and emits cmp insn before branch.
5062 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
5063 to microblaze_expand_conditional_branch and consolidate logic.
5064 (microblaze_expand_conditional_branch): emit branch_compare
5065 insn instead of handling cmp op separate from branch insn.
5067 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5069 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
5072 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5074 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
5075 define_insn with define_expand and new define_insn
5076 *altivec_lve<VI_char>x_internal.
5077 (altivec_stve<VI_char>x): Replace define_insn with define_expand
5078 and new define_insn *altivec_stve<VI_char>x_internal.
5079 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
5081 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
5083 (altivec_expand_stvex_be): New function.
5085 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
5087 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
5088 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
5089 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
5090 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
5092 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
5095 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
5096 instead of emit_move_insn.
5098 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5100 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
5102 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
5103 gen_altivec_vsumsws.
5105 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5107 * config/rs6000/altivec.md (altivec_lvxl): Rename as
5108 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
5109 (altivec_lvxl_<mode>): New define_expand incorporating
5110 -maltivec=be semantics where needed.
5111 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
5112 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
5113 semantics where needed.
5114 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
5115 (altivec_stvx_<mode>): New define_expand incorporating
5116 -maltivec=be semantics where needed.
5117 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
5118 VM2 iterator instead of V4SI.
5119 (altivec_stvxl_<mode>): New define_expand incorporating
5120 -maltivec=be semantics where needed.
5121 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
5122 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
5123 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
5124 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
5125 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
5126 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
5127 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5128 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5129 ALTIVEC_BUILTIN_STVXL.
5130 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5131 (altivec_expand_stvx_be): Likewise.
5132 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5133 (altivec_expand_lvx_be): Likewise.
5134 (altivec_expand_stvx_be): Likewise.
5135 (altivec_expand_builtin): Add cases for
5136 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5137 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5138 (altivec_init_builtins): Add definitions for
5139 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5140 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5142 2014-02-21 Catherine Moore <clm@codesourcery.com>
5144 * doc/invoke.texi (mvirt, mno-virt): Document.
5145 * config/mips/mips.opt (mvirt): New option.
5146 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5148 2014-02-21 Richard Biener <rguenther@suse.de>
5150 PR tree-optimization/60276
5151 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5152 (STMT_VINFO_MIN_NEG_DIST): New macro.
5153 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5154 STMT_VINFO_MIN_NEG_DIST.
5155 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5156 made for negative dependence distances still hold.
5158 2014-02-21 Richard Biener <rguenther@suse.de>
5161 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5162 DECL_INITIAL for globals not in the current function context.
5164 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5166 PR tree-optimization/56490
5167 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5168 * tree-ssa-uninit.c: Include params.h.
5169 (compute_control_dep_chain): Add num_calls argument, return false
5170 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
5171 num_calls to recursive call.
5172 (find_predicates): Change dep_chain into normal array,
5173 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
5174 variable and adjust compute_control_dep_chain caller.
5175 (find_def_preds): Likewise.
5177 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
5179 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
5180 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
5182 2014-02-21 Nick Clifton <nickc@redhat.com>
5184 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
5185 (pushhi1): Likewise.
5186 (popqi1): Add mode to pre_dec.
5189 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5191 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
5192 mode for mask of V8SFmode permutation.
5194 2014-02-20 Richard Henderson <rth@redhat.com>
5197 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
5198 a new pseudo for OLDVAL.
5200 2014-02-20 Jakub Jelinek <jakub@redhat.com>
5203 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
5204 gen_reg_rtx if d->testing_p.
5205 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
5206 if d->testing_p and we will certainly return true.
5207 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
5210 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
5212 * emit-rtl.c (gen_reg_rtx): Assert that
5213 crtl->emit.regno_pointer_align_length is non-zero.
5215 2014-02-20 Richard Henderson <rth@redhat.com>
5218 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
5219 on failure the store back into EXPECT.
5221 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
5222 Sandra Loosemore <sandra@codesourcery.com>
5224 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
5225 * config/nios2/nios2.c (nios2_function_profiler): Add
5226 -fPIC (flag_pic == 2) support.
5227 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
5228 (nios2_large_offset_p): New function.
5229 (nios2_unspec_reloc_p): Move up position, update to use
5230 nios2_large_offset_p.
5231 (nios2_unspec_address): Remove function.
5232 (nios2_unspec_offset): New function.
5233 (nios2_large_got_address): New function.
5234 (nios2_got_address): Add large offset support.
5235 (nios2_legitimize_tls_address): Update usage of removed and new
5237 (nios2_symbol_binds_local_p): New function.
5238 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
5239 (nios2_legitimize_address): Update to use nios2_large_offset_p.
5240 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
5241 (nios2_print_operand): Merge H/L processing, add hiadj/lo
5242 processing for (const (unspec ...)).
5243 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
5245 2014-02-20 Richard Biener <rguenther@suse.de>
5247 * tree-cfg.c (replace_uses_by): Mark altered BBs before
5248 doing the substitution.
5249 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
5251 2014-02-20 Martin Jambor <mjambor@suse.cz>
5254 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
5255 info when checking whether lattices are bottom.
5257 2014-02-20 Richard Biener <rguenther@suse.de>
5260 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
5263 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
5266 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
5267 parameter specifying the scaling.
5268 (inline_call): Update.
5269 (want_inline_recursively): Guard division by zero.
5270 (recursive_inlining): Update.
5271 * ipa-inline.h (clone_inlined_nodes): Update.
5273 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5276 * config/i386/i386.c (classify_argument): Pass structures of size
5277 64 bytes or less in register.
5279 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5280 Kirill Yukhin <kirill.yukhin@intel.com>
5282 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5283 (_mm_rcp28_round_ss): Ditto.
5284 (_mm_rsqrt28_round_sd): Ditto.
5285 (_mm_rsqrt28_round_ss): Ditto.
5286 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5287 (_mm_rcp14_round_ss): Ditto.
5288 (_mm_rsqrt14_round_sd): Ditto.
5289 (_mm_rsqrt14_round_ss): Ditto.
5290 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5291 the first input operand, get rid of match_dup.
5292 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5294 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5296 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5297 operand as the first input operand, set type attribute.
5298 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5300 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5301 operand as the first input operand, set type attribute.
5303 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5305 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5308 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5311 * config/i386/i386.c (construct_container): Remove TFmode check
5312 for X86_64_INTEGER_CLASS.
5314 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5317 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5318 only when -Wpsabi is enabled.
5320 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5323 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5324 passing arrays in registers are the same as for structs, so remove the
5325 special case for them.
5327 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5329 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5330 destination type, extract only the valid bits if the source type is not
5331 integral and has a different mode.
5333 2014-02-19 Richard Biener <rguenther@suse.de>
5336 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5339 2014-02-19 Richard Biener <rguenther@suse.de>
5342 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5343 (ipa_modify_call_arguments): Emit an argument load explicitely and
5344 preserve virtual SSA form there and for the replacement call.
5345 Do not update SSA form nor free dominance info.
5347 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5349 * ipa.c (function_and_variable_visibility): Also clear WEAK
5350 flag when disolving COMDAT_GROUP.
5352 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5354 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5355 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5357 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5358 do more sanity checks.
5359 (detect_type_change): Return true when giving up early.
5360 (compute_complex_assign_jump_func): Fix type parameter of
5361 ipa_set_ancestor_jf.
5362 (compute_complex_ancestor_jump_func): Likewise.
5363 (update_jump_functions_after_inlining): Fix updating of
5365 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5367 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5369 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5370 inline clones when edge disappears.
5372 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5375 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5376 Split 64-bit moves into 2 patterns. Do not allow the use of
5377 direct move for TDmode in little endian, since the decimal value
5378 has little endian bytes within a word, but the 64-bit pieces are
5379 ordered in a big endian fashion, and normal subreg's of TDmode are
5381 (mov<mode>_64bit_dm): Likewise.
5382 (movtd_64bit_nodm): Likewise.
5384 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5386 PR tree-optimization/60174
5387 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5388 statement of an SSA_NAME that occurs in an abnormal PHI node.
5390 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5393 * final.c (SEEN_BB): Remove.
5394 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5395 (final_scan_insn): Don't force_source_line on second
5396 NOTE_INSN_BASIC_BLOCK.
5398 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5401 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5402 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5403 (type_natural_mode): Warn ABI change when %zmm register is not
5404 available for AVX512F vector value passing.
5406 2014-02-18 Kai Tietz <ktietz@redhat.com>
5409 * config/i386/i386.c (ix86_expand_prologue): Use value in
5410 rax register as displacement when restoring %r10 or %rax.
5411 Fix wrong offset when restoring both registers.
5413 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5415 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5416 assertion with conditional return.
5418 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5419 Uros Bizjak <ubizjak@gmail.com>
5422 * config/i386/driver-i386.c (host_detect_local_cpu): If
5423 YMM state is not saved by the OS, also clear has_f16c. Move
5424 CPUID 0x80000001 handling before YMM state saving checking.
5426 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5428 PR rtl-optimization/58960
5429 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5430 factored out from ...
5431 (sched_init): ... here.
5432 (free_global_sched_pressure_data): New, factored out from ...
5433 (sched_finish): ... here.
5434 * sched-int.h (free_global_sched_pressure_data): Declare.
5435 * sched-rgn.c (nr_regions_initial): New static global.
5436 (haifa_find_rgns): Initialize it.
5437 (schedule_region): Disable sched-pressure for the newly
5440 2014-02-17 Richard Biener <rguenther@suse.de>
5442 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5443 release SSA defs of pattern stmts.
5445 2014-02-17 Richard Biener <rguenther@suse.de>
5447 * tree-inline.c (expand_call_inline): Release the virtual
5448 operand defined by the call we are about to inline.
5450 2014-02-17 Richard Biener <rguenther@suse.de>
5452 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5454 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5455 Ilya Tocar <ilya.tocar@intel.com>
5457 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5458 arguments order in builtin.
5459 (_mm512_permutexvar_epi64): Ditto.
5460 (_mm512_mask_permutexvar_epi64): Ditto
5461 (_mm512_maskz_permutexvar_epi32): Ditto
5462 (_mm512_permutexvar_epi32): Ditto
5463 (_mm512_mask_permutexvar_epi32): Ditto
5465 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5467 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5468 (p8_vmrgow): Likewise.
5470 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5472 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5475 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5478 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5479 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5480 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5481 using direct move instructions on ISA 2.07. Also adjust
5482 instruction length for 64-bit.
5483 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5484 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5486 2014-02-15 Alan Modra <amodra@gmail.com>
5490 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5491 find_replacement on parts of insn rtl that might be reloaded.
5493 2014-02-15 Richard Biener <rguenther@suse.de>
5495 PR tree-optimization/60183
5496 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5497 (tree_ssa_phiprop): Calculate and free post-dominators.
5499 2014-02-14 Jeff Law <law@redhat.com>
5501 PR rtl-optimization/60131
5502 * ree.c (get_extended_src_reg): New function.
5503 (combine_reaching_defs): Use it rather than assuming location of REG.
5504 (find_and_remove_re): Verify first operand of extension is
5505 a REG before adding the insns to the copy list.
5507 2014-02-14 Roland McGrath <mcgrathr@google.com>
5509 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5510 * configure: Regenerated.
5511 * config.in: Regenerated.
5512 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5513 instead of ASM_SHORT.
5515 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5516 Richard Earnshaw <rearnsha@arm.com>
5518 PR rtl-optimization/59535
5519 * lra-constraints.c (process_alt_operands): Encourage alternative
5520 when unassigned pseudo class is superset of the alternative class.
5521 (inherit_reload_reg): Don't inherit when optimizing for code size.
5522 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5523 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5524 modes not less than 4 for Thumb1.
5526 2014-02-14 Kyle McMartin <kyle@redhat.com>
5529 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5531 2014-02-14 Richard Biener <rguenther@suse.de>
5533 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5534 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5535 Do not strip INDIRECT_REFs.
5537 2014-02-14 Richard Biener <rguenther@suse.de>
5540 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5541 DECL_FUNCTION_SPECIFIC_TARGET.
5542 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5543 * tree-streamer-out.c (pack_ts_target_option): Remove.
5544 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5545 (write_ts_function_decl_tree_pointers): Do not stream
5546 DECL_FUNCTION_SPECIFIC_TARGET.
5547 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5548 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5549 (lto_input_ts_function_decl_tree_pointers): Do not stream
5550 DECL_FUNCTION_SPECIFIC_TARGET.
5552 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5554 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5555 (get_initial_def_for_induction, vectorizable_induction): Ignore
5556 debug stmts when looking for exit_phi.
5557 (vectorizable_live_operation): Fix up condition.
5559 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5561 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5562 nreverse() because it changes the content of original tree list.
5564 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5566 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5567 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5569 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5571 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5572 GNU coding standards.
5574 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5577 * dwarf2out.c (gen_subprogram_die): Don't call
5578 add_calling_convention_attribute if subr_die is old_die.
5580 2014-02-13 Sharad Singhai <singhai@google.com>
5582 * doc/optinfo.texi: Fix order of nodes.
5584 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5586 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5587 operands[2], not operands[3].
5589 2014-02-13 Richard Biener <rguenther@suse.de>
5592 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5593 mention the possibility of an in-tree build.
5594 (CLooG): Update recommended version to 0.18.1, mention the
5595 possibility of an in-tree build and clarify that the ISL
5596 bundled with CLooG does not work.
5598 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5601 * expr.c (compress_float_constant): If x is a hard register,
5602 extend into a pseudo and then move to x.
5604 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5606 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5607 caused by bad second argument to warning_at() with -mhotpatch and
5608 nested functions (e.g. with gfortran).
5610 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5612 * opts.c (option_name): Remove "enabled by default" rider.
5614 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5616 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5618 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5619 Uros Bizjak <ubizjak@gmail.com>
5622 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5623 * configure: Regenerated.
5625 2014-02-12 Richard Biener <rguenther@suse.de>
5627 * vec.c (vec_prefix::calculate_allocation): Move as
5628 inline variant to vec.h.
5629 (vec_prefix::calculate_allocation_1): New out-of-line version.
5630 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5631 (vec_prefix::m_has_auto_buf): Rename to ...
5632 (vec_prefix::m_using_auto_storage): ... this.
5633 (vec_prefix::calculate_allocation): Inline the easy cases
5634 and dispatch to calculate_allocation_1 which doesn't need the
5636 (va_heap::reserve): Use gcc_checking_assert.
5637 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5638 m_using_auto_storage.
5639 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5641 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5642 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5643 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5645 2014-02-12 Richard Biener <rguenther@suse.de>
5647 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5648 when we found a dependence.
5650 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5652 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5654 (maybe_fold_stmt): ... into this new function.
5655 * omp-low.c (lower_omp): Update comment.
5657 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5660 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5663 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5665 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5666 identifiers in comments.
5667 (cortexa53_extra_costs): Likewise.
5668 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5669 (cortexa7_extra_costs): Likewise.
5670 (cortexa12_extra_costs): Likewise.
5671 (cortexa15_extra_costs): Likewise.
5672 (v7m_extra_costs): Likewise.
5674 2014-02-12 Richard Biener <rguenther@suse.de>
5677 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5678 of posix_memalign being successful.
5679 (lower_stmt): Restrict lowering of posix_memalign to when
5680 -ftree-bit-ccp is enabled.
5682 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5684 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5686 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5688 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5690 PR rtl-optimization/60116
5691 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5692 other_insn once the combination has been validated.
5694 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5697 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5699 * ipa-devirt.c: Include demangle.h
5700 (odr_violation_reported): New static variable.
5701 (add_type_duplicate): Update odr_violations.
5702 (maybe_record_node): Add completep parameter; update it.
5703 (record_target_from_binfo): Add COMPLETEP parameter;
5704 update it as needed.
5705 (possible_polymorphic_call_targets_1): Likewise.
5706 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5707 rename FINAL to COMPLETE.
5708 (record_targets_from_bases): Sanity check we found the binfo;
5709 fix COMPLETEP updating.
5710 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5711 parameter, fix computing of COMPLETEP.
5712 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5713 at LTO time do demangling.
5714 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5715 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5717 (gimple_get_virt_method_for_binfo): Likewise.
5718 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5719 gimple_get_virt_method_for_vtable): Update prototypes.
5721 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5724 * genautomata.c (add_presence_absence): Fix typo with
5725 {final_}presence_list.
5727 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5730 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5731 for VSX/Altivec vectors that land in GPR registers.
5733 2014-02-11 Richard Henderson <rth@redhat.com>
5734 Jakub Jelinek <jakub@redhat.com>
5737 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5738 around drhs if type conversion to lacc->type is not useless.
5740 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5742 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5744 (cortex-a57.cortex-a53): Likewise.
5745 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5747 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5749 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5752 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5754 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5755 add_options_for_arm_vfp3.
5757 2014-02-11 Jeff Law <law@redhat.com>
5760 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5761 object with an undesirable mode.
5763 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5766 * config/i386/sol2-9.h: New file.
5767 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5768 *-*-solaris2.9*): Use it.
5770 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5772 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5773 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5775 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5777 * config/microblaze/microblaze.c: Extend mcpu version format
5779 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5781 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5783 2014-02-10 Richard Henderson <rth@redhat.com>
5786 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5787 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5788 ms-abi vs -mno-accumulate-outgoing-args.
5789 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5790 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5793 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5796 * cfgexpand.c (expand_asm_operands): Attach source location to
5797 ASM_INPUT rtx objects.
5798 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5800 2014-02-10 Nick Clifton <nickc@redhat.com>
5802 * config/mn10300/mn10300.c (popcount): New function.
5803 (mn10300_expand_prologue): Include saved registers in stack usage
5806 2014-02-10 Jeff Law <law@redhat.com>
5809 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5810 when changing the SET_DEST of a prior insn to avoid an input reload.
5812 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5814 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5815 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5816 -mcall-openbsd, or -mcall-linux.
5817 (CC1_ENDIAN_BIG_SPEC): Remove.
5818 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5819 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5820 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5821 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5822 and %cc1_endian_default.
5823 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5825 2014-02-10 Richard Biener <rguenther@suse.de>
5827 PR tree-optimization/60115
5828 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5829 MEM_REF handling. Properly verify that the accesses are not
5830 out of the objects bound.
5832 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5834 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5837 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5839 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5840 proper constants and fix formatting.
5841 (possible_polymorphic_call_targets): Fix formatting.
5843 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5844 Ilya Tocar <ilya.tocar@intel.com>
5846 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5847 (_mm512_loadu_epi32): Renamed into...
5848 (_mm512_loadu_si512): This.
5849 (_mm512_storeu_epi32): Renamed into...
5850 (_mm512_storeu_si512): This.
5851 (_mm512_maskz_ceil_ps): Removed.
5852 (_mm512_maskz_ceil_pd): Ditto.
5853 (_mm512_maskz_floor_ps): Ditto.
5854 (_mm512_maskz_floor_pd): Ditto.
5855 (_mm512_floor_round_ps): Ditto.
5856 (_mm512_floor_round_pd): Ditto.
5857 (_mm512_ceil_round_ps): Ditto.
5858 (_mm512_ceil_round_pd): Ditto.
5859 (_mm512_mask_floor_round_ps): Ditto.
5860 (_mm512_mask_floor_round_pd): Ditto.
5861 (_mm512_mask_ceil_round_ps): Ditto.
5862 (_mm512_mask_ceil_round_pd): Ditto.
5863 (_mm512_maskz_floor_round_ps): Ditto.
5864 (_mm512_maskz_floor_round_pd): Ditto.
5865 (_mm512_maskz_ceil_round_ps): Ditto.
5866 (_mm512_maskz_ceil_round_pd): Ditto.
5867 (_mm512_expand_pd): Ditto.
5868 (_mm512_expand_ps): Ditto.
5869 * config/i386/i386.c (ix86_builtins): Remove
5870 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5871 (bdesc_args): Ditto.
5872 * config/i386/predicates.md (const1256_operand): New.
5873 (const_1_to_2_operand): Ditto.
5874 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5875 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5876 (*avx512pf_gatherpf<mode>sf): Ditto.
5877 (avx512pf_gatherpf<mode>df): Ditto.
5878 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5879 (*avx512pf_gatherpf<mode>df): Ditto.
5880 (avx512pf_scatterpf<mode>sf): Ditto.
5881 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5882 (*avx512pf_scatterpf<mode>sf): Ditto.
5883 (avx512pf_scatterpf<mode>df): Ditto.
5884 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5885 (*avx512pf_scatterpf<mode>df): Ditto.
5886 (avx512f_expand<mode>): Removed.
5887 (<shift_insn><mode>3<mask_name>): Change predicate type.
5889 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5891 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5892 not at the end of datarefs vector use ordered_remove to avoid
5893 reordering datarefs vector.
5896 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5897 mark local addressable non-static vars as GOVD_PRIVATE
5898 instead of GOVD_LOCAL.
5899 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5900 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5904 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5905 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5906 assume_aligned or alloc_align attributes.
5907 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5908 arguments. Handle also assume_aligned and alloc_align attributes.
5909 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5910 calls to functions with assume_aligned or alloc_align attributes.
5911 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5913 2014-02-08 Terry Guo <terry.guo@arm.com>
5915 * doc/invoke.texi: Document ARM -march=armv7e-m.
5917 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5919 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5920 flag on __cilkrts_rethrow builtin.
5923 * ipa-cp.c (determine_versionability): Fail at -O0
5924 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5925 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5928 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5931 * tree-inline.c (copy_forbidden): Fail for
5932 __attribute__((optimize (0))) functions.
5934 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5936 * varpool.c: Include pointer-set.h.
5937 (varpool_remove_unreferenced_decls): Variables in other partitions
5938 will not be output; be however careful to not lose information
5941 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5943 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5944 lookup in the vtable constructor.
5946 2014-02-07 Jeff Law <law@redhat.com>
5949 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5950 define_insn_and_split.
5952 * ipa-inline.c (inline_small_functions): Fix typos.
5954 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5956 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5957 (s390_can_use_return_insn): Declare.
5958 * config/s390/s390.h (EPILOGUE_USES): Define.
5959 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5961 (s390_chunkify_start): Handle return JUMP_LABELs.
5962 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5963 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5964 (s390_can_use_return_insn): New functions.
5965 (s390_fix_long_loop_prediction): Handle conditional returns.
5966 (TARGET_SET_UP_BY_PROLOGUE): Define.
5967 * config/s390/s390.md (ANY_RETURN): New code iterator.
5968 (*creturn, *csimple_return, return, simple_return): New patterns.
5970 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5972 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5973 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5974 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5975 REG_CFA_RESTORE list when deciding not to restore a register.
5977 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5979 * config/s390/s390.c: Include tree-pass.h and context.h.
5980 (s390_early_mach): New function, split out from...
5981 (s390_emit_prologue): ...here.
5982 (pass_data_s390_early_mach): New pass structure.
5983 (pass_s390_early_mach): New class.
5984 (s390_option_override): Create and register early_mach pass.
5985 Move to end of file.
5987 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5989 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5990 to match for the exit block.
5992 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5994 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5995 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5996 Reject misaligned operands.
5998 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6000 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
6002 2014-02-07 Richard Biener <rguenther@suse.de>
6005 * gimple-low.c (lower_builtin_posix_memalign): New function.
6006 (lower_stmt): Call it to lower posix_memalign in a way
6007 to make alignment info accessible.
6009 2014-02-07 Jakub Jelinek <jakub@redhat.com>
6012 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
6013 __builtin_setjmp_receiver.
6015 2014-02-07 Richard Biener <rguenther@suse.de>
6018 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
6019 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
6020 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6021 Handle BUILT_IN_POSIX_MEMALIGN.
6022 (find_func_clobbers): Likewise.
6023 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
6024 (call_may_clobber_ref_p_1): Likewise.
6026 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6029 * ipa-devirt.c (record_target_from_binfo): Remove overactive
6032 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6035 * lto-cgraph.c (lto_output_node): Use
6036 symtab_get_symbol_partitioning_class.
6037 (lto_output_varpool_node): likewise.
6038 (symtab_get_symbol_partitioning_class): Move here from
6040 * cgraph.h (symbol_partitioning_class): Likewise.
6041 (symtab_get_symbol_partitioning_class): Declare.
6043 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6045 * ggc.h (ggc_internal_cleared_alloc): New macro.
6046 * vec.h (vec_safe_copy): Handle memory stats.
6047 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
6048 * target-globals.c (save_target_globals): Likewise.
6050 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
6053 * expr.c (emit_move_resolve_push): Export; be bit more selective
6054 on when to clear alias set.
6055 * expr.h (emit_move_resolve_push): Declare.
6056 * function.h (struct function): Add tail_call_marked.
6057 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
6058 * config/i386/i386-protos.h (ix86_expand_push): Remove.
6059 * config/i386/i386.md (TImode move expander): De not call
6061 (FP push expanders): Preserve memory attributes.
6062 * config/i386/sse.md (push<mode>1): Remove.
6063 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
6064 (ix86_expand_push): Remove.
6065 * config/i386/mmx.md (push<mode>1): Remove.
6067 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6069 PR rtl-optimization/60030
6070 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
6071 lopart with paradoxical subreg before shifting it up by hprec.
6073 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6075 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
6076 Remove extra newline at end of file.
6077 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
6078 (arm_issue_rate): Handle cortexa57.
6079 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
6080 (cortex-a57.cortex-a53): Likewise.
6082 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6085 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
6086 don't record in REG_FRAME_RELATED_EXPR registers not set in that
6088 (arm_expand_prologue): Adjust all callers.
6089 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
6090 info, registers also at the lowest numbered registers side. Use
6091 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
6095 * var-tracking.c (adjust_mems): Before adding a SET to
6096 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
6098 2014-02-06 Alan Modra <amodra@gmail.com>
6101 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
6102 change SDmode to DDmode when lra_in_progress.
6104 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6107 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
6108 free_data_ref on the dr first, and before goto again also set dr
6109 to the next dr. For simd_lane_access, free old datarefs[i] before
6110 overwriting it. For get_vectype_for_scalar_type failure, don't
6111 free_data_ref if simd_lane_access.
6113 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
6116 * tree.h (opts_for_fn): New inline function.
6117 (opt_for_fn): Define.
6118 * config/i386/i386.c (ix86_function_regparm): Use
6119 opt_for_fn (decl, optimize) instead of optimize.
6121 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
6123 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
6124 for SYMBOL_REF in large memory model.
6126 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6128 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6130 (cortex-a57): Likewise.
6131 (cortex-a57.cortex-a53): Likewise.
6133 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6134 Kugan Vivekanandarajah <kuganv@linaro.org>
6136 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6138 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6140 2014-02-06 Richard Biener <rguenther@suse.de>
6142 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6143 set_loop_copy and initialize_original_copy_tables.
6145 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6147 * config/aarch64/aarch64-simd.md
6148 (aarch64_ashr_simddi): Change QI to SI.
6150 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6151 Jakub Jelinek <jakub@redhat.com>
6154 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6157 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6159 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6160 CODE_FOR_altivec_vpku[hw]um to
6161 CODE_FOR_altivec_vpku[hw]um_direct.
6162 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6163 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6164 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6165 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6167 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6169 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6170 generation for -maltivec=be.
6171 (altivec_vsumsws): Simplify redundant test.
6173 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6175 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
6176 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
6177 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
6178 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
6179 gen_altivec_vpkuwum.
6180 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
6182 (altivec_vpks<VI_char>ss): Likewise.
6183 (altivec_vpks<VI_char>us): Likewise.
6184 (altivec_vpku<VI_char>us): Likewise.
6185 (altivec_vpku<VI_char>um): Likewise.
6186 (altivec_vpku<VI_char>um_direct): New (copy of
6187 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
6189 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
6190 target is little endian and -maltivec=be is not specified.
6191 (*altivec_vupkhs<VU_char>_direct): New (copy of
6192 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
6193 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
6194 target is little endian and -maltivec=be is not specified.
6195 (*altivec_vupkls<VU_char>_direct): New (copy of
6196 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
6197 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
6198 little endian and -maltivec=be is not specified.
6199 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
6200 little endian and -maltivec=be is not specified.
6202 2014-02-05 Richard Henderson <rth@redhat.com>
6205 * combine-stack-adj.c: Revert r206943.
6206 * sched-int.h (struct deps_desc): Add last_args_size.
6207 * sched-deps.c (init_deps): Initialize it.
6208 (sched_analyze_insn): Add OUTPUT dependencies between insns that
6209 contain REG_ARGS_SIZE notes.
6211 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6213 * lto-cgraph.c (asm_nodes_output): Make global.
6214 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
6215 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
6216 (driver_handle_option): Handle OPT_fwpa.
6218 2014-02-05 Jakub Jelinek <jakub@redhat.com>
6221 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
6222 a comment typo and formatting issue. If odr_hash hasn't been
6223 created, return vNULL and set *completep to false.
6226 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
6227 bb with no successors.
6229 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
6232 * doc/invoke.texi (-march): Clarify documentation for ARM.
6236 2014-02-05 Richard Biener <rguenther@suse.de>
6238 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
6239 when not vectorizing because of too many alias checks.
6240 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
6241 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
6243 2014-02-05 Nick Clifton <nickc@redhat.com>
6245 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
6246 accept extended registers in any mode when compiling for the MN10300.
6248 2014-02-05 Yury Gribov <y.gribov@samsung.com>
6250 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
6251 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
6252 sanitization attributes.
6253 (can_inline_edge_p): Likewise.
6254 (sanitize_attrs_match_for_inline_p): New function.
6256 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6258 * ipa-prop.c (detect_type_change): Shor circuit testing of
6259 type changes on THIS pointer.
6261 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
6264 * config/pa/pa.c (legitimize_tls_address): Return original address
6265 if not passed a SYMBOL_REF rtx.
6266 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
6268 (pa_emit_move_sequence): Simplify TLS source operands.
6269 (pa_legitimate_constant_p): Reject all TLS constants.
6270 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
6271 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
6273 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6275 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
6276 groups when we know they are controlled by LTO.
6277 * varasm.c (default_binds_local_p_1): If object is in other partition,
6278 it will be resolved locally.
6280 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6282 * config/host-linux.c (linux_gt_pch_use_address): Don't
6283 use SSIZE_MAX because it is not always defined.
6285 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6288 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6289 threshold for pseudo splitting.
6290 (update_ebb_live_info): Process call argument hard registers and
6291 hard registers from insn definition too.
6292 (max_small_class_regs_num): New constant.
6293 (inherit_in_ebb): Update live hard regs through EBBs. Update
6294 reloads_num only for small register classes. Don't split for
6297 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6300 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6303 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6305 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6308 2014-02-04 Marek Polacek <polacek@redhat.com>
6310 * gdbinit.in (pel): Define.
6312 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6314 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6317 2014-02-04 Richard Biener <rguenther@suse.de>
6320 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6321 in function context local.
6322 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6323 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6324 similar to LTO_imported_decl_ref.
6326 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6328 PR tree-optimization/60002
6329 * cgraphclones.c (build_function_decl_skip_args): Clear
6332 PR tree-optimization/60023
6333 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6334 false to gsi_replace.
6335 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6336 has been in some EH region and vec_stmt could throw, add
6337 vec_stmt into the same EH region.
6338 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6339 has no lhs, ignore it.
6340 * internal-fn.c (expand_MASK_LOAD): Likewise.
6343 * tree-inline.c (copy_forbidden): Fail for
6344 __attribute__((optimize (0))) functions.
6347 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6348 is set, copy one less argument.
6349 (expand_simd_clones): Don't subtract clone_info->inbranch
6350 from simd_clone_struct_alloc argument.
6352 PR rtl-optimization/57915
6353 * recog.c (simplify_while_replacing): If all unary/binary/relational
6354 operation arguments are constant, attempt to simplify those.
6357 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6358 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6360 2014-02-04 Richard Biener <rguenther@suse.de>
6362 PR tree-optimization/60012
6363 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6364 TBAA disambiguation to all DDRs.
6366 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6369 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6370 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6372 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6375 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6377 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6379 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6380 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6382 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6385 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6386 to figure out targets of polymorphic calls with known decl.
6387 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6388 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6389 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6390 (get_polymorphic_call_info): ... here.
6391 (get_polymorphic_call_info_from_invariant): New function.
6393 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6395 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6396 lookup via vtable pointer; check for type consistency
6397 and turn inconsitent facts into UNREACHABLE.
6398 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6399 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6400 type inconsistent querries; return UNREACHABLE instead.
6402 2014-02-03 Richard Henderson <rth@twiddle.net>
6405 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6406 already processed this node.
6407 (normalize_one_pred_1): Pass along mark_set.
6408 (normalize_one_pred): Create and destroy a pointer_set_t.
6409 (normalize_one_pred_chain): Likewise.
6411 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6413 PR gcov-profile/58602
6414 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6416 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6419 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6420 -fno-devirtualize; try to devirtualize by the knowledge of
6421 virtual table pointer given by aggregate propagation.
6422 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6423 (ipa_print_node_jump_functions): Dump also offset that
6424 is relevant for polymorphic calls.
6425 (determine_known_aggregate_parts): Add arg_type parameter; use it
6426 instead of determining the type from pointer type.
6427 (ipa_compute_jump_functions_for_edge): Update call of
6428 determine_known_aggregate_parts.
6429 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6430 (gimple_get_virt_method_for_binfo): ... here; simplify using
6431 vtable_pointer_value_to_vtable.
6432 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6433 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6434 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6435 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6437 (vtable_pointer_value_to_binfo): ... here.
6438 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6440 2014-02-03 Teresa Johnson <tejohnson@google.com>
6442 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6443 redef of outer loop index variable.
6445 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6449 * doc/extend.texi (Function Attributes): Typo.
6451 2014-02-03 Cong Hou <congh@google.com>
6453 PR tree-optimization/60000
6454 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6455 if the vectorized statement is a store. A store statement can only
6456 appear at the end of pattern statements.
6458 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6460 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6461 (ix86_option_override_internal): Default long double to 64-bit for
6462 32-bit Bionic and to 128-bit for 64-bit Bionic.
6464 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6465 TARGET_LONG_DOUBLE_128 is true.
6466 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6468 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6469 (mlong-double-64): Negate -mlong-double-128.
6470 (mlong-double-128): New option.
6472 * config/i386/i386-c.c (ix86_target_macros): Define
6473 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6475 * doc/invoke.texi: Document -mlong-double-128.
6477 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6479 PR rtl-optimization/60024
6480 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6482 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6484 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6486 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6488 PR rtl-optimization/57662
6489 * sel-sched.c (code_motion_path_driver): Do not mark already not
6490 existing blocks in the visiting bitmap.
6492 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6494 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6495 on the insn being emitted.
6497 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6498 Will Deacon <will.deacon@arm.com>
6500 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6502 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6504 * config/arm/arm-tables.opt: Regenerate.
6506 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6508 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6509 for vector types other than V16QImode.
6510 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6511 define_expand, and call altivec_expand_vec_perm_le when producing
6512 code with little endian element order.
6513 (*altivec_vperm_<mode>_internal): New insn having previous
6514 behavior of altivec_vperm_<mode>.
6515 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6516 altivec_expand_vec_perm_le when producing code with little endian
6518 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6519 behavior of altivec_vperm_<mode>_uns.
6521 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6523 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6524 (altivec_vsumsws): Add handling for -maltivec=be with a little
6526 (altivec_vsumsws_direct): New.
6527 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6528 gen_altivec_vsumsws.
6530 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6532 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6533 vtable_pointer_value_to_binfo): New functions.
6534 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6535 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6537 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6539 * config/nios2/nios2.md (load_got_register): Initialize GOT
6540 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6541 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6543 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6545 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6546 preserverd by passthrough, do not propagate the type.
6548 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6550 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6551 (mips_atomic_assign_expand_fenv): New function.
6552 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6554 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6556 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6557 (__builtin_mips_set_fcsr): Likewise.
6558 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6559 MIPS_USI_FTYPE_VOID.
6560 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6561 (mips16_expand_set_fcsr): Likewise.
6562 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6563 (mips16_set_fcsr_stub): Likewise.
6564 (mips16_get_fcsr_one_only_stub): New class.
6565 (mips16_set_fcsr_one_only_stub): Likewise.
6566 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6567 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6568 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6569 (hard_float): New availability predicate.
6570 (mips_builtins): Add get_fcsr and set_fcsr.
6571 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6572 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6573 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6574 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6575 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6578 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6580 * config/mips/mips.c (mips_one_only_stub): New class.
6581 (mips_need_mips16_rdhwr_p): Replace with...
6582 (mips16_rdhwr_stub): ...this new variable.
6583 (mips16_stub_call_address): New function.
6584 (mips16_rdhwr_one_only_stub): New class.
6585 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6586 (mips_output_mips16_rdhwr): Delete.
6587 (mips_finish_stub): New function.
6588 (mips_code_end): Use it to handle rdhwr stubs.
6590 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6593 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6594 when calculating size of integer atomic types.
6596 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6598 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6600 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6602 PR tree-optimization/60003
6603 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6604 * profile.c (branch_prob): Use gimple_call_builtin_p
6605 to check for BUILT_IN_SETJMP_RECEIVER.
6606 * tree-inline.c (copy_bb): Call notice_special_calls.
6608 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6611 * lra-constraints.c (process_alt_operands): Update reload_sum only
6614 2014-01-31 Richard Henderson <rth@redhat.com>
6617 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6618 until after else_eh is processed.
6620 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6622 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6623 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6624 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6625 in smmintrin.h, remove them.
6626 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6627 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6628 * config/i386/i386.md (ROUND_SAE): Fix value.
6629 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6630 (const48_operand): New.
6631 * config/i386/subst.md (round), (round_expand): Use
6632 const_4_or_8_to_11_operand.
6633 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6635 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6637 * config/i386/constraints.md (Yk): Swap meaning with k.
6638 * config/i386/i386.md (movhi_internal): Change Yk to k.
6639 (movqi_internal): Ditto.
6640 (*k<logic><mode>): Ditto.
6643 (kandn<mode>): Ditto.
6644 (*<code>hi_1): Ditto.
6645 (*<code>qi_1): Ditto.
6646 (kxnor<mode>): Ditto.
6647 (kortestzhi): Ditto.
6648 (kortestchi): Ditto.
6650 (*one_cmplhi2_1): Ditto.
6651 (*one_cmplqi2_1): Ditto.
6652 * config/i386/sse.md (): Change k to Yk.
6653 (avx512f_load<mode>_mask): Ditto.
6654 (avx512f_blendm<mode>): Ditto.
6655 (avx512f_store<mode>_mask): Ditto.
6656 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6657 (avx512f_storedqu<mode>_mask): Ditto.
6658 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6660 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6661 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6662 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6663 (avx512f_maskcmp<mode>3): Ditto.
6664 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6665 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6666 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6667 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6668 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6669 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6670 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6671 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6672 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6673 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6674 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6675 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6676 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6677 (vec_extract_lo_<mode>_maskm): Ditto.
6678 (vec_extract_hi_<mode>_maskm): Ditto.
6679 (avx512f_vternlog<mode>_mask): Ditto.
6680 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6681 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6682 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6683 (avx512f_<code>v8div16qi2_mask): Ditto.
6684 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6685 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6686 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6687 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6688 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6689 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6690 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6691 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6692 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6693 (avx512cd_maskb_vec_dupv8di): Ditto.
6694 (avx512cd_maskw_vec_dupv16si): Ditto.
6695 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6696 (avx512f_vpermi2var<mode>3_mask): Ditto.
6697 (avx512f_vpermi2var<mode>3_mask): Ditto.
6698 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6699 (*avx512f_gathersi<mode>): Ditto.
6700 (*avx512f_gathersi<mode>_2): Ditto.
6701 (*avx512f_gatherdi<mode>): Ditto.
6702 (*avx512f_gatherdi<mode>_2): Ditto.
6703 (*avx512f_scattersi<mode>): Ditto.
6704 (*avx512f_scatterdi<mode>): Ditto.
6705 (avx512f_compress<mode>_mask): Ditto.
6706 (avx512f_compressstore<mode>_mask): Ditto.
6707 (avx512f_expand<mode>_mask): Ditto.
6708 * config/i386/subst.md (mask): Change k to Yk.
6709 (mask_scalar_merge): Ditto.
6712 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6714 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6716 2014-01-31 Richard Biener <rguenther@suse.de>
6719 * builtins.c (fold_builtin_memory_op): Make sure to not
6720 use a floating-point mode or a boolean or enumeral type for
6723 2014-01-30 DJ Delorie <dj@redhat.com>
6725 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6726 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6727 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6728 whenever main() has an epilogue.
6730 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6732 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6733 unused variable "field".
6734 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6735 (vsx_mergeh_<mode>): Likewise.
6736 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6737 (altivec_vmrghh): Likewise.
6738 (altivec_vmrghw): Likewise.
6739 (altivec_vmrglb): Likewise.
6740 (altivec_vmrglh): Likewise.
6741 (altivec_vmrglw): Likewise.
6742 (altivec_vspltb): Add missing uses.
6743 (altivec_vsplth): Likewise.
6744 (altivec_vspltw): Likewise.
6745 (altivec_vspltsf): Likewise.
6747 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6750 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6751 frame related instructions.
6753 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6755 PR rtl-optimization/59959
6756 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6757 any reload of register whose subreg is invalid.
6759 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6761 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6762 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6763 Add missing return type - void.
6765 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6767 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6768 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6769 remove element index adjustment for endian (now handled in vsx.md
6771 (altivec_expand_vec_perm_const): Use
6772 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6773 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6774 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6775 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6776 define_expand and a new define_insn *altivec_vspltb_internal;
6777 adjust for -maltivec=be on a little endian target.
6778 (altivec_vspltb_direct): New.
6779 (altivec_vsplth): Divide into a define_expand and a new
6780 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6781 little endian target.
6782 (altivec_vsplth_direct): New.
6783 (altivec_vspltw): Divide into a define_expand and a new
6784 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6785 little endian target.
6786 (altivec_vspltw_direct): New.
6787 (altivec_vspltsf): Divide into a define_expand and a new
6788 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6789 a little endian target.
6791 2014-01-30 Richard Biener <rguenther@suse.de>
6793 PR tree-optimization/59993
6794 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6795 can propagate form the earlier stmt and avoid the transform
6796 when the intermediate result is needed.
6798 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6800 * README.Portability: Fix typo.
6802 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6804 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6805 comparison_operator with ordered_comparison_operator.
6807 2014-01-30 Nick Clifton <nickc@redhat.com>
6809 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6810 Rename to mn10300_store_multiple_regs.
6811 * config/mn10300/mn10300.c: Likewise.
6812 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6813 store_multiple_regs.
6814 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6815 Call mn10300_store_multiple_regs.
6817 2014-01-30 Nick Clifton <nickc@redhat.com>
6818 DJ Delorie <dj@redhat.com>
6820 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6821 %fp 2 to keep registers after it properly word-aligned.
6822 (rl78_alloc_physical_registers_umul): Handle the case where both
6823 input operands are the same.
6825 2014-01-30 Richard Biener <rguenther@suse.de>
6827 PR tree-optimization/59903
6828 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6831 2014-01-30 Jason Merrill <jason@redhat.com>
6834 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6837 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6839 2014-01-30 Richard Biener <rguenther@suse.de>
6841 PR tree-optimization/59951
6842 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6844 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6847 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6848 SFmode to DFmode case.
6850 2014-01-29 DJ Delorie <dj@redhat.com>
6852 * config/msp430/msp430.opt (-minrt): New.
6853 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6855 (ENDFILE_SPEC): Likewise.
6857 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6859 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6860 (estimate_function_body_sizes): Use it.
6862 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6865 * dwarf2out.c (is_cxx_auto): New.
6866 (is_base_type): Use it.
6867 (gen_type_die_with_usage): Likewise.
6869 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6871 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6872 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6873 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6874 -maltivec=be with LE targets.
6875 (vsx_mergeh_<mode>): Likewise.
6876 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6877 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6878 (altivec_vmrghb): Replace with define_expand and new
6879 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6880 (altivec_vmrghb_direct): New define_insn.
6881 (altivec_vmrghh): Replace with define_expand and new
6882 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6883 (altivec_vmrghh_direct): New define_insn.
6884 (altivec_vmrghw): Replace with define_expand and new
6885 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6886 (altivec_vmrghw_direct): New define_insn.
6887 (*altivec_vmrghsf): Adjust for endianness.
6888 (altivec_vmrglb): Replace with define_expand and new
6889 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6890 (altivec_vmrglb_direct): New define_insn.
6891 (altivec_vmrglh): Replace with define_expand and new
6892 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6893 (altivec_vmrglh_direct): New define_insn.
6894 (altivec_vmrglw): Replace with define_expand and new
6895 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6896 (altivec_vmrglw_direct): New define_insn.
6897 (*altivec_vmrglsf): Adjust for endianness.
6898 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6899 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6900 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6901 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6902 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6903 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6904 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6905 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6907 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6909 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6910 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6913 2014-01-29 Richard Biener <rguenther@suse.de>
6915 PR tree-optimization/58742
6916 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6917 associate_pointerplus_align.
6918 (associate_pointerplus_diff): New function.
6919 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6920 and associate_pointerplus_diff.
6922 2014-01-29 Richard Biener <rguenther@suse.de>
6924 * lto-streamer.h (LTO_major_version): Bump to 3.
6925 (LTO_minor_version): Reset to 0.
6927 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6929 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6930 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6931 (arm_file_start): Generate correct asm header for armv7ve.
6932 * config/arm/bpabi.h: Add multilib support for armv7ve.
6933 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6934 and cortex-a15 to armv7ve.
6935 * config/arm/t-aprofile: Add multilib support for armv7ve.
6936 * doc/invoke.texi: Document -march=armv7ve.
6938 2014-01-29 Richard Biener <rguenther@suse.de>
6940 PR tree-optimization/58742
6941 * tree-ssa-forwprop.c (associate_plusminus): Return true
6942 if we changed sth, defer EH cleanup to ...
6943 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6944 (simplify_mult): New function.
6946 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6949 PR tree-optimization/59920
6950 * tree.c (build_common_builtin_nodes): Remove
6951 __builtin_setjmp_dispatcher initialization.
6952 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6953 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6954 instead of gsi_after_labels + manually skipping debug stmts.
6955 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6956 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6957 * tree-inline.c (copy_edges_for_bb): Remove
6958 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6959 argument. Ignore computed_goto_p stmts. Don't call
6960 make_abnormal_goto_edges. If a call might need abnormal edges
6961 for non-local gotos, see if it already has an edge to
6962 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6963 with true argument, don't do anything then, otherwise add
6964 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6965 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6967 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6968 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6969 (lower_stmt): Don't set data->calls_builtin_setjmp.
6970 (lower_builtin_setjmp): Adjust comment.
6971 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6972 * tree-cfg.c (found_computed_goto): Remove.
6973 (factor_computed_gotos): Remove.
6974 (make_goto_expr_edges): Return bool, true for computed gotos.
6975 Don't call make_abnormal_goto_edges.
6976 (build_gimple_cfg): Don't set found_computed_goto, don't call
6977 factor_computed_gotos.
6978 (computed_goto_p): No longer static.
6979 (make_blocks): Don't set found_computed_goto.
6980 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6981 (make_edges): If make_goto_expr_edges returns true, push bb
6982 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6983 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6984 vector. Record mapping between bbs and OpenMP regions if there
6985 are any, adjust make_gimple_omp_edges caller. Call
6986 handle_abnormal_edges.
6987 (make_abnormal_goto_edges): Remove.
6988 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6989 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6990 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6991 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6992 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6993 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6994 filling *region also set *region_idx to (*region)->entry->index.
6997 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6998 For REGs set ORIGINAL_REGNO.
7000 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
7002 * doc/md.texi: Mention that a target shouldn't implement
7003 vec_widen_(s|u)mul_even/odd pair if it is less efficient
7006 2014-01-29 Jakub Jelinek <jakub@redhat.com>
7008 PR tree-optimization/59594
7009 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
7010 a copy of the datarefs vector rather than the vector itself.
7012 2014-01-28 Jason Merrill <jason@redhat.com>
7015 * dwarf2out.c (auto_die): New static.
7016 (gen_type_die_with_usage): Handle C++1y 'auto'.
7017 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
7020 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
7023 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
7024 (SPEC_X32): Likewise.
7025 (SPEC_64): Likewise.
7026 * config/i386/i386.c (ix86_option_override_internal): Turn off
7027 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
7029 (x86_file_start): Output .code16gcc for TARGET_16BIT.
7030 * config/i386/i386.h (TARGET_16BIT): New macro.
7031 (TARGET_16BIT_P): Likewise.
7032 * config/i386/i386.opt: Add m16.
7033 * doc/invoke.texi: Document -m16.
7035 2014-01-28 Jakub Jelinek <jakub@redhat.com>
7037 PR preprocessor/59935
7038 * input.c (location_get_source_line): Bail out on when line number
7039 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
7041 2014-01-28 Richard Biener <rguenther@suse.de>
7043 PR tree-optimization/58742
7044 * tree-ssa-forwprop.c (associate_plusminus): Handle
7045 pointer subtraction of the form (T)(P + A) - (T)P.
7047 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7049 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
7052 2014-01-28 Richard Biener <rguenther@suse.de>
7055 2014-01-28 Richard Biener <rguenther@suse.de>
7057 PR rtl-optimization/45364
7058 PR rtl-optimization/59890
7059 * var-tracking.c (local_get_addr_clear_given_value): Handle
7060 already cleared slot.
7061 (val_reset): Handle not allocated local_get_addr_cache.
7062 (vt_find_locations): Use post-order on the inverted CFG.
7064 2014-01-28 Richard Biener <rguenther@suse.de>
7066 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
7068 2014-01-28 Richard Biener <rguenther@suse.de>
7070 PR rtl-optimization/45364
7071 PR rtl-optimization/59890
7072 * var-tracking.c (local_get_addr_clear_given_value): Handle
7073 already cleared slot.
7074 (val_reset): Handle not allocated local_get_addr_cache.
7075 (vt_find_locations): Use post-order on the inverted CFG.
7077 2014-01-28 Alan Modra <amodra@gmail.com>
7079 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
7080 * configure.ac <recursive call for build != host>: Define
7081 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
7082 and LD_FOR_BUILD too.
7083 * configure: Regenerate.
7085 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
7087 * config/i386/i386.c (get_builtin_code_for_version): Separate
7088 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
7089 Broadwell from Haswell.
7091 2014-01-27 Steve Ellcey <sellcey@mips.com>
7093 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
7094 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
7095 * config/mips/mips.c (mips_option_override): Change setting
7097 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
7098 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
7099 Change from Mask to Var.
7101 2014-01-27 Jeff Law <law@redhat.com>
7103 * ipa-inline.c (inline_small_functions): Fix typo.
7105 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
7107 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
7108 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
7109 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
7110 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
7111 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
7112 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
7113 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
7114 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
7115 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
7116 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
7117 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
7118 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
7119 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
7120 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
7121 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
7122 (_mm512_storeu_epi64): Ditto.
7123 (_mm512_cmpge_epi32_mask): Ditto.
7124 (_mm512_cmpge_epu32_mask): Ditto.
7125 (_mm512_cmpge_epi64_mask): Ditto.
7126 (_mm512_cmpge_epu64_mask): Ditto.
7127 (_mm512_cmple_epi32_mask): Ditto.
7128 (_mm512_cmple_epu32_mask): Ditto.
7129 (_mm512_cmple_epi64_mask): Ditto.
7130 (_mm512_cmple_epu64_mask): Ditto.
7131 (_mm512_cmplt_epi32_mask): Ditto.
7132 (_mm512_cmplt_epu32_mask): Ditto.
7133 (_mm512_cmplt_epi64_mask): Ditto.
7134 (_mm512_cmplt_epu64_mask): Ditto.
7135 (_mm512_cmpneq_epi32_mask): Ditto.
7136 (_mm512_cmpneq_epu32_mask): Ditto.
7137 (_mm512_cmpneq_epi64_mask): Ditto.
7138 (_mm512_cmpneq_epu64_mask): Ditto.
7139 (_mm512_expand_pd): Ditto.
7140 (_mm512_expand_ps): Ditto.
7141 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7142 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7143 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7144 * config/i386/i386.c (ix86_builtins): Add
7145 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7146 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7147 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7148 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7149 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7150 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7151 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7152 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7153 IX86_BUILTIN_PMOVUSQW512_MEM.
7154 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7155 __builtin_ia32_pmovsqd512mem_mask,
7156 __builtin_ia32_pmovqd512mem_mask,
7157 __builtin_ia32_pmovusqw512mem_mask,
7158 __builtin_ia32_pmovsqw512mem_mask,
7159 __builtin_ia32_pmovqw512mem_mask,
7160 __builtin_ia32_pmovusdw512mem_mask,
7161 __builtin_ia32_pmovsdw512mem_mask,
7162 __builtin_ia32_pmovdw512mem_mask,
7163 __builtin_ia32_pmovqb512mem_mask,
7164 __builtin_ia32_pmovusqb512mem_mask,
7165 __builtin_ia32_pmovsqb512mem_mask,
7166 __builtin_ia32_pmovusdb512mem_mask,
7167 __builtin_ia32_pmovsdb512mem_mask,
7168 __builtin_ia32_pmovdb512mem_mask.
7169 (bdesc_args): Add __builtin_ia32_expanddf512,
7170 __builtin_ia32_expandsf512.
7171 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
7172 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
7173 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
7174 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
7175 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
7176 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
7177 (avx512f_<code>v8div16qi2_mask_store): This.
7178 (avx512f_expand<mode>): New.
7180 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
7182 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
7184 (_mm512_mask_prefetch_i64gather_pd): Ditto.
7185 (_mm512_prefetch_i32scatter_pd): Ditto.
7186 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
7187 (_mm512_prefetch_i64scatter_pd): Ditto.
7188 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
7189 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
7190 (_mm512_mask_prefetch_i64gather_ps): Ditto.
7191 (_mm512_prefetch_i32scatter_ps): Ditto.
7192 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
7193 (_mm512_prefetch_i64scatter_ps): Ditto.
7194 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
7195 * config/i386/i386-builtin-types.def: Define
7196 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
7197 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
7198 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
7199 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
7200 IX86_BUILTIN_SCATTERPFQPD.
7201 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
7202 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
7203 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
7204 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
7205 __builtin_ia32_scatterpfqps.
7206 (ix86_expand_builtin): Expand new built-ins.
7207 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
7208 fix memory access data type.
7209 (*avx512pf_gatherpf<mode>_mask): Ditto.
7210 (*avx512pf_gatherpf<mode>): Ditto.
7211 (avx512pf_scatterpf<mode>): Ditto.
7212 (*avx512pf_scatterpf<mode>_mask): Ditto.
7213 (*avx512pf_scatterpf<mode>): Ditto.
7214 (GATHER_SCATTER_SF_MEM_MODE): New.
7215 (avx512pf_gatherpf<mode>df): Ditto.
7216 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7217 (*avx512pf_scatterpf<mode>df): Ditto.
7219 2014-01-27 Jakub Jelinek <jakub@redhat.com>
7222 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
7223 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
7226 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7228 * common/config/arm/arm-common.c
7229 (arm_rewrite_mcpu): Handle multiple names.
7231 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7233 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7235 * gimple-builder.h (create_gimple_tmp): Delete.
7237 2014-01-27 Christian Bruel <christian.bruel@st.com>
7239 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
7242 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
7244 * config/pa/pa.md (call): Generate indirect long calls to non-local
7245 functions when outputing 32-bit code.
7246 (call_value): Likewise except for special call to buggy powf function.
7248 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
7249 portable runtime and PIC indirect calls.
7250 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
7251 and PIC call sequences. Use ldo instead of blr to set return register
7252 in PIC call sequence.
7254 2014-01-25 Walter Lee <walt@tilera.com>
7256 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
7257 avoid clobbering a live register.
7259 2014-01-25 Walter Lee <walt@tilera.com>
7261 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
7262 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
7263 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
7264 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
7266 2014-01-25 Walter Lee <walt@tilera.com>
7268 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
7269 arguments on even registers.
7270 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
7272 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
7273 (BIGGEST_ALIGNMENT): Ditto.
7274 (BIGGEST_FIELD_ALIGNMENT): Ditto.
7276 2014-01-25 Walter Lee <walt@tilera.com>
7278 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
7279 insns before bundling.
7280 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
7282 2014-01-25 Walter Lee <walt@tilera.com>
7284 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7285 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7286 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7288 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7290 * config/mips/constraints.md (kl): Delete.
7291 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7292 define expands, using...
7293 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7294 instructions for MIPS16.
7295 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7296 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7298 2014-01-25 Walter Lee <walt@tilera.com>
7300 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7304 2014-01-25 Walter Lee <walt@tilera.com>
7306 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7307 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7309 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7311 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7314 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7316 * print-rtl.c (in_call_function_usage): New var.
7317 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7318 EXPR_LIST mode as mode and not as reg note name.
7321 * cfgloopmanip.c (copy_loop_info): If
7322 loop->warned_aggressive_loop_optimizations, make sure
7323 the flag is set in target loop too.
7325 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7327 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7329 * builtins.def: Likewise.
7330 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7331 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7332 * ira.c (ira_setup_eliminable_regset): Likewise.
7333 * omp-low.c (gate_expand_omp): Likewise.
7334 (execute_lower_omp): Likewise.
7335 (diagnose_sb_0): Likewise.
7336 (gate_diagnose_omp_blocks): Likewise.
7337 (simd_clone_clauses_extract): Likewise.
7340 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7342 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7343 correction for little endian...
7344 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7347 2014-01-24 Jeff Law <law@redhat.com>
7349 PR tree-optimization/59919
7350 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7351 for non-returning calls.
7353 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7355 * common/config/aarch64/aarch64-common.c
7356 (aarch64_rewrite_mcpu): Handle multiple names.
7357 * config/aarch64/aarch64.h
7358 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7360 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7362 * input.c (add_file_to_cache_tab): Handle the case where fopen
7365 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7368 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7369 from push operand if code of push isn't PRE_DEC.
7371 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7374 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7375 -mquad-memory-atomic. Update -mquad-memory documentation to say
7376 it is only used for non-atomic loads/stores.
7378 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7379 -mquad-memory or -mquad-memory-atomic switches.
7381 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7382 -mquad-memory-atomic to ISA 2.07 support.
7384 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7385 to separate support of normal quad word memory operations (ldq, stq)
7386 from the atomic quad word memory operations.
7388 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7389 support to separate non-atomic quad word operations from atomic
7390 quad word operations. Disable non-atomic quad word operations in
7391 little endian mode so that we don't have to swap words after the
7392 load and before the store.
7393 (quad_load_store_p): Add comment about atomic quad word support.
7394 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7395 options printed with -mdebug=reg.
7397 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7398 -mquad-memory-atomic as the test for whether we have quad word
7399 atomic instructions.
7400 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7401 or -mp8-vector are used, allow byte/half-word atomic operations.
7403 * config/rs6000/sync.md (load_lockedti): Insure that the address
7404 is a proper indexed or indirect address for the lqarx instruction.
7405 On little endian systems, swap the hi/lo registers after the lqarx
7407 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7408 insure the address is valid for the lqarx instruction.
7409 (store_conditionalti): Insure that the address is a proper indexed
7410 or indirect address for the stqcrx. instruction. On little endian
7411 systems, swap the hi/lo registers before doing the stqcrx.
7413 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7414 insure the address is valid for the stqcrx. instruction.
7416 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7417 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7418 type of quad memory support is available.
7420 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7423 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7424 there is a danger of looping.
7426 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7428 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7429 force flag_ira_loop_pressure if set via command line.
7431 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7433 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7434 (ashr_simd): New builtin handling DI mode.
7435 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7436 (aarch64_sshr_simddi): New match pattern.
7437 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7438 (vshrd_n_s64): Likewise.
7439 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7441 2014-01-23 Nick Clifton <nickc@redhat.com>
7443 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7444 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7445 favour of mcu specific scripts.
7446 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7449 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7450 Alex Velenko <Alex.Velenko@arm.com>
7452 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7453 (vaddv_s16): Likewise.
7454 (vaddv_s32): Likewise.
7455 (vaddv_u8): Likewise.
7456 (vaddv_u16): Likewise.
7457 (vaddv_u32): Likewise.
7458 (vaddvq_s8): Likewise.
7459 (vaddvq_s16): Likewise.
7460 (vaddvq_s32): Likewise.
7461 (vaddvq_s64): Likewise.
7462 (vaddvq_u8): Likewise.
7463 (vaddvq_u16): Likewise.
7464 (vaddvq_u32): Likewise.
7465 (vaddvq_u64): Likewise.
7466 (vaddv_f32): Likewise.
7467 (vaddvq_f32): Likewise.
7468 (vaddvq_f64): Likewise.
7469 (vmaxv_f32): Likewise.
7470 (vmaxv_s8): Likewise.
7471 (vmaxv_s16): Likewise.
7472 (vmaxv_s32): Likewise.
7473 (vmaxv_u8): Likewise.
7474 (vmaxv_u16): Likewise.
7475 (vmaxv_u32): Likewise.
7476 (vmaxvq_f32): Likewise.
7477 (vmaxvq_f64): Likewise.
7478 (vmaxvq_s8): Likewise.
7479 (vmaxvq_s16): Likewise.
7480 (vmaxvq_s32): Likewise.
7481 (vmaxvq_u8): Likewise.
7482 (vmaxvq_u16): Likewise.
7483 (vmaxvq_u32): Likewise.
7484 (vmaxnmv_f32): Likewise.
7485 (vmaxnmvq_f32): Likewise.
7486 (vmaxnmvq_f64): Likewise.
7487 (vminv_f32): Likewise.
7488 (vminv_s8): Likewise.
7489 (vminv_s16): Likewise.
7490 (vminv_s32): Likewise.
7491 (vminv_u8): Likewise.
7492 (vminv_u16): Likewise.
7493 (vminv_u32): Likewise.
7494 (vminvq_f32): Likewise.
7495 (vminvq_f64): Likewise.
7496 (vminvq_s8): Likewise.
7497 (vminvq_s16): Likewise.
7498 (vminvq_s32): Likewise.
7499 (vminvq_u8): Likewise.
7500 (vminvq_u16): Likewise.
7501 (vminvq_u32): Likewise.
7502 (vminnmv_f32): Likewise.
7503 (vminnmvq_f32): Likewise.
7504 (vminnmvq_f64): Likewise.
7506 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7508 * config/aarch64/aarch64-simd.md
7509 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7510 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7511 (*aarch64_mul3_elt<mode>): Likewise.
7512 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7513 (*aarch64_mul3_elt_to_64v2df): Likewise.
7514 (*aarch64_mla_elt<mode>): Likewise.
7515 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7516 (*aarch64_mls_elt<mode>): Likewise.
7517 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7518 (*aarch64_fma4_elt<mode>): Likewise.
7519 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7520 (*aarch64_fma4_elt_to_64v2df): Likewise.
7521 (*aarch64_fnma4_elt<mode>): Likewise.
7522 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7523 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7524 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7525 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7526 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7527 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7528 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7529 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7530 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7532 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7534 * config/aarch64/aarch64-simd.md
7535 (aarch64_be_checked_get_lane<mode>): New define_expand.
7536 * config/aarch64/aarch64-simd-builtins.def
7537 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7538 New builtin definition.
7539 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7540 Use new safe be builtin.
7542 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7544 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7546 (aarch64_be_st1<mode>): Likewise.
7547 (aarch_ld1<VALL:mode>): Define_expand modified.
7548 (aarch_st1<VALL:mode>): Likewise.
7549 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7550 (UNSPEC_ST1): Likewise.
7552 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7554 * config/microblaze/microblaze.md: Add trap insn and attribute
7556 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7558 PR preprocessor/58580
7559 * input.h (location_get_source_line): Take an additional line_size
7561 (void diagnostics_file_cache_fini): Declare new function.
7562 * input.c (struct fcache): New type.
7563 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7564 New static constants.
7565 (diagnostic_file_cache_init, total_lines_num)
7566 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7567 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7568 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7569 (get_next_line, read_next_line, goto_next_line, read_line_num):
7570 New static function definitions.
7571 (diagnostic_file_cache_fini): New function.
7572 (location_get_source_line): Take an additional output line_len
7573 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7575 * diagnostic.c (diagnostic_finish): Call
7576 diagnostic_file_cache_fini.
7577 (adjust_line): Take an additional input parameter for the length
7578 of the line, rather than calculating it with strlen.
7579 (diagnostic_show_locus): Adjust the use of
7580 location_get_source_line and adjust_line with respect to their new
7581 signature. While displaying a line now, do not stop at the first
7582 null byte. Rather, display the zero byte as a space and keep
7583 going until we reach the size of the line.
7584 * Makefile.in: Add vec.o to OBJS-libcommon
7586 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7587 Ilya Tocar <ilya.tocar@intel.com>
7589 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7590 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7591 (__builtin_ia32_kmov16): Ditto.
7592 * config/i386/i386.md (UNSPEC_KMOV): New.
7595 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7597 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7598 (_mm512_storeu_si512): Ditto.
7600 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7603 * rtl.h (get_referenced_operands): Declare.
7604 * recog.c (get_referenced_operands): New function.
7605 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7606 operands have been referenced when recording LO_SUM references.
7608 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7610 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7612 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7614 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7615 Enable for generic and recent AMD targets.
7617 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7619 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7620 ARG_SIZE note when adjustment was eliminated.
7622 2014-01-22 Jeff Law <law@redhat.com>
7624 PR tree-optimization/59597
7625 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7626 in file. Accept new argument REGISTERING and use it to modify
7627 dump output appropriately.
7628 (register_jump_thread): Corresponding changes.
7629 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7630 thread paths involving joiner blocks. Add code to dump cancelled
7631 jump threading paths.
7633 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7635 PR rtl-optimization/59477
7636 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7637 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7639 2014-01-22 Tom Tromey <tromey@redhat.com>
7641 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7643 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7645 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7647 PR rtl-optimization/59896
7648 * lra-constraints.c (process_alt_operands): Check unused note for
7649 matched operands of insn with no output reloads.
7651 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7653 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7654 (mips_move_from_gpr_cost): Likewise.
7656 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7658 PR rtl-optimization/59858
7659 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7660 ira_class_hard_regs_num.
7661 (process_alt_operands): Increase reject for dying matched operand.
7663 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7666 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7667 smaller than size, perform several stores or loads and stores
7668 at dst + count - size to store or copy all of size bytes, rather
7669 than just last modesize bytes.
7671 2014-01-20 DJ Delorie <dj@redhat.com>
7673 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7674 that CLOBBERs are REGs before propogating their values.
7676 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7679 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7680 (cgraph_inline_failed_type): New function.
7681 * cgraph.h (DEFCIFCODE): Add type.
7682 (cgraph_inline_failed_type_t): New enum.
7683 (cgraph_inline_failed_type): New prototype.
7684 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7685 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7686 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7687 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7688 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7689 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7690 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7691 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7692 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7693 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7694 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7695 OPTIMIZATION_MISMATCH.
7696 * tree-inline.c (expand_call_inline): Emit errors during
7697 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7699 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7702 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7703 mode attribute in insn output.
7705 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7707 * output.h (output_constant): Delete.
7708 * varasm.c (output_constant): Make private.
7710 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7712 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7714 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7717 * tree.h (fold_builtin_strcat): New prototype.
7718 * builtins.c (fold_builtin_strcat): No longer static. Add len
7719 argument, if non-NULL, don't call c_strlen. Optimize
7720 directly into __builtin_memcpy instead of __builtin_strcpy.
7721 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7722 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7724 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7726 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7727 for SImode_address_operand operands, having only a REG argument.
7729 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7731 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7732 loader name using mbig-endian.
7733 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7735 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7737 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7741 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7743 * config/aarch64/aarch64-protos.h
7744 (aarch64_cannot_change_mode_class_ptr): Declare.
7745 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7746 aarch64_cannot_change_mode_class_ptr): New.
7747 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7748 backend hook aarch64_cannot_change_mode_class.
7750 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7752 * common/config/aarch64/aarch64-common.c
7753 (aarch64_handle_option): Don't handle any option order logic here.
7754 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7755 selected_cpu, warn on architecture version mismatch.
7756 (aarch64_override_options): Fix parsing order for option strings.
7758 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7759 Iain Sandoe <iain@codesourcery.com>
7762 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7763 warning. Amend comment to reflect current functionality.
7765 2014-01-20 Richard Biener <rguenther@suse.de>
7768 * builtins.c (fold_builtin_strcat): Remove case better handled
7769 by tree-ssa-strlen.c.
7771 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7773 * config/aarch64/aarch64.opt
7774 (mcpu, march, mtune): Make case-insensitive.
7776 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7779 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7780 if operands[1] is a REG or ZERO_EXTEND of a REG.
7782 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7784 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7786 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7788 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7789 long non-pic millicode calls.
7791 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7793 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7795 2014-01-19 Kito Cheng <kito@0xlab.org>
7797 * builtins.c (expand_movstr): Check movstr expand done or fail.
7799 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7800 H.J. Lu <hongjiu.lu@intel.com>
7803 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7804 to DImode for zero-extended addresses.
7806 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7808 PR rtl-optimization/57763
7809 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7810 on the new indirect jump_insn and increment LABEL_NUSES (label).
7812 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7816 * config.gcc (x86_archs): New variable.
7817 (x86_64_archs): Likewise.
7818 (x86_cpus): Likewise.
7819 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7820 --with-arch/--with-cpu= options.
7821 Support --with-arch=/--with-cpu={nehalem,westmere,
7822 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7824 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7826 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7827 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7829 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7831 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7833 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7836 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7837 clear cpp_get_options (parse_in)->warn_unused_macros for
7838 ix86_target_macros_internal with cpp_define.
7840 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7842 * jump.c (delete_related_insns): Keep (use (insn))s.
7843 * reorg.c (redundant_insn): Check for barriers too.
7845 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7847 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7849 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7851 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7852 call to $$dyncall when TARGET_LONG_CALLS is true.
7854 2014-01-17 Jeff Law <law@redhat.com>
7856 * ree.c (combine_set_extension): Temporarily disable test for
7857 changing number of hard registers.
7859 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7862 * ipa-inline-analysis.c (inline_free_summary):
7863 Do not free summary of aliases.
7865 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7868 * gimplify.c (gimplify_expr): Use create_tmp_var
7869 instead of create_tmp_var_raw. If cond doesn't have
7870 integral type, don't add the IFN_ANNOTATE builtin at all.
7872 2014-01-17 Martin Jambor <mjambor@suse.cz>
7875 * ipa-cp.c (prev_edge_clone): New variable.
7876 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7877 Also resize prev_edge_clone vector.
7878 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7879 (ipcp_edge_removal_hook): New function.
7880 (ipcp_driver): Register ipcp_edge_removal_hook.
7882 2014-01-17 Andrew Pinski <apinski@cavium.com>
7883 Steve Ellcey <sellcey@mips.com>
7886 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7889 2014-01-17 Jeff Law <law@redhat.com>
7892 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7893 so that pass_ccp runs first.
7895 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7897 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7898 (ix86_adjust_cost): Use !TARGET_XXX.
7899 (do_reorder_for_imul): Likewise.
7900 (swap_top_of_ready_list): Likewise.
7901 (ix86_sched_reorder): Likewise.
7903 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7905 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7906 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7907 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7908 (intel_memset): New. Duplicate slm_memset.
7909 (intel_cost): New. Duplicate slm_cost.
7910 (m_INTEL): New macro.
7911 (processor_target_table): Add "intel".
7912 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7913 with PROCESSOR_INTEL for "intel".
7914 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7915 PROCESSOR_SILVERMONT.
7916 (ix86_issue_rate): Likewise.
7917 (ix86_adjust_cost): Likewise.
7918 (ia32_multipass_dfa_lookahead): Likewise.
7919 (swap_top_of_ready_list): Likewise.
7920 (ix86_sched_reorder): Likewise.
7921 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7922 instead of TARGET_OPT_AGU.
7923 * config/i386/i386.h (TARGET_INTEL): New.
7924 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7925 (processor_type): Add PROCESSOR_INTEL.
7926 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7927 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7929 2014-01-17 Marek Polacek <polacek@redhat.com>
7932 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7935 2014-01-17 Richard Biener <rguenther@suse.de>
7937 PR tree-optimization/46590
7938 * opts.c (default_options_table): Add entries for
7939 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7940 all enabled at -O1 but not for -Og.
7941 * common.opt (fbranch-count-reg): Remove Init(1).
7942 (fmove-loop-invariants): Likewise.
7943 (ftree-pta): Likewise.
7945 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7947 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7948 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7949 decls to at least the GCC 4.8 used alignments.
7952 * tree-nested.c (convert_nonlocal_reference_stmt,
7953 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7954 of GIMPLE_BIND stmts, adjust associated decls.
7956 2014-01-17 Richard Biener <rguenther@suse.de>
7958 PR tree-optimization/46590
7959 * vec.h (vec<>::bseach): New member function implementing
7960 binary search according to C89 bsearch.
7961 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7962 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7963 bitmap pointer again. Make accesses_in_loop a flat array.
7964 (mem_ref_obstack): New global.
7965 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7966 (mark_ref_stored): Likewise.
7967 (ref_indep_loop_p_2): Likewise.
7968 (set_ref_stored_in_loop): New helper function.
7969 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7970 (memref_free): Adjust.
7971 (record_mem_ref_loc): Simplify.
7972 (gather_mem_refs_stmt): Adjust.
7973 (sort_locs_in_loop_postorder_cmp): New function.
7974 (analyze_memory_references): Sort accesses_in_loop after
7975 loop postorder number.
7976 (find_ref_loc_in_loop_cmp): New function.
7977 (for_all_locs_in_loop): Find relevant cluster of locs in
7978 accesses_in_loop and iterate without recursion.
7979 (execute_sm): Avoid uninit warning.
7980 (struct ref_always_accessed): Simplify.
7981 (ref_always_accessed::operator ()): Likewise.
7982 (ref_always_accessed_p): Likewise.
7983 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7984 loop postorder numbers here.
7985 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7988 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7991 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7992 on decls for which assemble_alias has been called.
7994 2014-01-17 Nick Clifton <nickc@redhat.com>
7996 * config/msp430/msp430.opt: (mcpu): New option.
7997 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7998 (msp430_option_override): Parse target_cpu. If the MCU name
7999 matches a generic string, clear target_mcu.
8000 (msp430_attr): Allow numeric interrupt values up to 63.
8001 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
8002 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
8004 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
8006 * config/msp430/msp430.md (popm): Use %J rather than %I.
8007 (addsi3): Use msp430_nonimmediate_operand for operand 2.
8008 (addhi_cy_i): Use immediate_operand for operand 2.
8009 * doc/invoke.texi: Document -mcpu option.
8011 2014-01-17 Richard Biener <rguenther@suse.de>
8013 PR rtl-optimization/38518
8014 * df.h (df_analyze_loop): Declare.
8015 * df-core.c: Include cfgloop.h.
8016 (df_analyze_1): Split out main part of df_analyze.
8017 (df_analyze): Adjust.
8018 (loop_inverted_post_order_compute): New function.
8019 (loop_post_order_compute): Likewise.
8020 (df_analyze_loop): New function avoiding whole-function
8022 * loop-invariant.c (find_defs): Use df_analyze_loop.
8023 (find_invariants): Adjust.
8024 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
8026 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
8028 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
8029 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
8031 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
8033 * ipa-ref.c (ipa_remove_stmt_references): Fix references
8034 traversal when removing references.
8036 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
8039 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
8041 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
8044 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
8045 pushing a reload for an autoinc when we had previously reloaded an
8046 inner part of the address.
8048 2014-01-16 Jakub Jelinek <jakub@redhat.com>
8050 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
8052 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
8053 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
8054 when not giving up or versioning for alias only because of
8056 (vect_analyze_data_ref_dependences): Set to true.
8057 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
8059 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
8060 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
8064 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
8067 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
8068 operand 0 predicate for gathers, use a new pseudo as subtarget.
8070 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8073 * lra-constraints.c (process_alt_operands): Add printing debug info.
8074 Check absence of input/output reloads for matched operands too.
8076 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
8078 PR rtl-optimization/59835
8079 * ira.c (ira_init_register_move_cost): Increase cost for
8082 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
8084 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
8086 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
8089 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
8090 non-register objects. Use gen_(high/low)part more consistently.
8093 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8096 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
8097 endian support, remove tests for WORDS_BIG_ENDIAN.
8098 (p8_mfvsrd_3_<mode>): Likewise.
8099 (reload_gpr_from_vsx<mode>): Likewise.
8100 (reload_gpr_from_vsxsf): Likewise.
8101 (p8_mfvsrd_4_disf): Likewise.
8103 2014-01-16 Richard Biener <rguenther@suse.de>
8105 PR rtl-optimization/46590
8106 * lcm.c (compute_antinout_edge): Use postorder iteration.
8107 (compute_laterin): Use inverted postorder iteration.
8109 2014-01-16 Nick Clifton <nickc@redhat.com>
8112 * varasm.c (output_constant): Return the number of bytes actually
8114 (output_constructor_array_range): Update the field size with the
8115 number of bytes emitted by output_constant.
8116 (output_constructor_regular_field): Likewise. Also do not
8117 complain if the total number of bytes emitted is now greater
8118 than the expected fieldpos.
8119 * output.h (output_constant): Update prototype and descriptive comment.
8121 2014-01-16 Marek Polacek <polacek@redhat.com>
8124 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
8125 it is error_mark_node.
8127 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8129 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8130 VALID_AVX256_REG_OR_OI_MODE.
8132 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8134 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8135 current procedure should be profiled.
8137 2014-01-15 Andrew Pinski <apinski@cavium.com>
8139 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8140 of moving from/to the STACK_REG register class.
8142 2014-01-15 Richard Henderson <rth@redhat.com>
8145 * reginfo.c (global_regs_decl): Globalize.
8146 * rtl.h (global_regs_decl): Declare.
8147 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8148 reserved via global_regs.
8150 2014-01-15 Teresa Johnson <tejohnson@google.com>
8152 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8154 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8156 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8157 and vmulosh rather than call gen_vec_widen_smult_*.
8158 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8159 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8160 (vec_widen_smult_even_v16qi): Likewise.
8161 (vec_widen_umult_even_v8hi): Likewise.
8162 (vec_widen_smult_even_v8hi): Likewise.
8163 (vec_widen_umult_odd_v16qi): Likewise.
8164 (vec_widen_smult_odd_v16qi): Likewise.
8165 (vec_widen_umult_odd_v8hi): Likewise.
8166 (vec_widen_smult_odd_v8hi): Likewise.
8167 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8168 vmuloub rather than call gen_vec_widen_umult_*.
8169 (vec_widen_umult_lo_v16qi): Likewise.
8170 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
8171 vmulosb rather than call gen_vec_widen_smult_*.
8172 (vec_widen_smult_lo_v16qi): Likewise.
8173 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
8174 rather than call gen_vec_widen_umult_*.
8175 (vec_widen_umult_lo_v8hi): Likewise.
8176 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
8177 rather than call gen_vec_widen_smult_*.
8178 (vec_widen_smult_lo_v8hi): Likewise.
8180 2014-01-15 Jeff Law <law@redhat.com>
8182 PR tree-optimization/59747
8183 * ree.c (find_and_remove_re): Properly handle case where a second
8184 eliminated extension requires widening a copy created for elimination
8185 of a prior extension.
8186 (combine_set_extension): Ensure that the number of hard regs needed
8187 for a destination register does not change when we widen it.
8189 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
8191 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
8192 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
8193 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
8194 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
8195 (avr-*-rtems*): Likewise.
8196 (bfin*-rtems*): Likewise.
8197 (moxie-*-rtems*): Likewise.
8198 (h8300-*-rtems*): Likewise.
8199 (i[34567]86-*-rtems*): Likewise.
8200 (lm32-*-rtems*): Likewise.
8201 (m32r-*-rtems*): Likewise.
8202 (m68k-*-rtems*): Likewise.
8203 (microblaze*-*-rtems*): Likewise.
8204 (mips*-*-rtems*): Likewise.
8205 (powerpc-*-rtems*): Likewise.
8206 (sh-*-rtems*): Likewise.
8207 (sparc-*-rtems*): Likewise.
8208 (sparc64-*-rtems*): Likewise.
8209 (v850-*-rtems*): Likewise.
8210 (m32c-*-rtems*): Likewise.
8212 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
8214 PR rtl-optimization/59511
8215 * ira.c (ira_init_register_move_cost): Use memory costs for some
8216 cases of register move cost calculations.
8217 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
8218 instead of BB frequency.
8219 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
8220 * lra-assigns.c (find_hard_regno_for): Ditto.
8222 2014-01-15 Richard Biener <rguenther@suse.de>
8224 PR tree-optimization/59822
8225 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
8226 (vectorizable_load): Use it to hoist defs of uses of invariant
8227 loads out of the loop.
8229 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8230 Kugan Vivekanandarajah <kuganv@linaro.org>
8233 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
8236 2014-01-15 Richard Biener <rguenther@suse.de>
8238 PR rtl-optimization/59802
8239 * lcm.c (compute_available): Use inverted postorder to seed
8240 the initial worklist.
8242 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8245 * config/s390/s390.c (s390_preferred_reload_class): Don't return
8246 ADDR_REGS for invalid symrefs in non-PIC code.
8248 2014-01-15 Jakub Jelinek <jakub@redhat.com>
8251 * builtins.c (determine_block_size): Initialize *probable_max_size
8252 even if len_rtx is CONST_INT.
8254 2014-01-14 Andrew Pinski <apinski@cavium.com>
8256 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
8257 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
8258 (cortexa53_tunings): Likewise.
8259 (aarch64_sched_issue_rate): New function.
8260 (TARGET_SCHED_ISSUE_RATE): Define.
8262 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8264 * ira-costs.c (find_costs_and_classes): Add missed
8265 ira_init_register_move_cost_if_necessary.
8267 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8270 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
8272 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
8275 * config/i386/i386.c (type_natural_mode): Add a bool parameter
8276 to indicate if type is used for function return value. Warn ABI
8277 change if the vector mode isn't available for function return value.
8278 (ix86_function_arg_advance): Pass false to type_natural_mode.
8279 (ix86_function_arg): Likewise.
8280 (ix86_gimplify_va_arg): Likewise.
8281 (function_arg_32): Don't warn ABI change.
8282 (ix86_function_value): Pass true to type_natural_mode.
8283 (ix86_return_in_memory): Likewise.
8284 (ix86_struct_value_rtx): Removed.
8285 (TARGET_STRUCT_VALUE_RTX): Likewise.
8287 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8289 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8290 converting a conditional jump into a conditional return.
8292 2014-01-14 Richard Biener <rguenther@suse.de>
8294 PR tree-optimization/58921
8295 PR tree-optimization/59006
8296 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8297 hoisting invariant stmts.
8298 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8299 invariant loads on the preheader edge if possible.
8301 2014-01-14 Joey Ye <joey.ye@arm.com>
8303 * doc/plugin.texi (Building GCC plugins): Update to C++.
8305 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8307 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8308 (_mm_rcp28_round_ss): Ditto.
8309 (_mm_rsqrt28_round_sd): Ditto.
8310 (_mm_rsqrt28_round_ss): Ditto.
8311 (_mm_rcp28_sd): Ditto.
8312 (_mm_rcp28_ss): Ditto.
8313 (_mm_rsqrt28_sd): Ditto.
8314 (_mm_rsqrt28_ss): Ditto.
8315 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8316 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8317 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8318 (IX86_BUILTIN_RCP28SD): Ditto.
8319 (IX86_BUILTIN_RCP28SS): Ditto.
8320 (IX86_BUILTIN_RSQRT28SD): Ditto.
8321 (IX86_BUILTIN_RSQRT28SS): Ditto.
8322 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8323 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8324 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8325 (ix86_expand_special_args_builtin): Expand new FTYPE.
8326 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8327 (srcp14<mode>): Make insn unary.
8328 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8329 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8330 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8331 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8332 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8333 Fix rounding: make it SAE only.
8334 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8336 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8338 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8339 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8340 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8341 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8342 (round_saeonly_mask_scalar_operand4): Ditto.
8343 (round_saeonly_mask_scalar_op3): Ditto.
8344 (round_saeonly_mask_scalar_op4): Ditto.
8346 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8348 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8349 Implement -maltivec=be for vec_insert and vec_extract.
8351 2014-01-10 DJ Delorie <dj@redhat.com>
8353 * config/msp430/msp430.md (call_internal): Don't allow memory
8354 references with SP as the base register.
8355 (call_value_internal): Likewise.
8356 * config/msp430/constraints.md (Yc): New. For memory references
8357 that don't use SP as a base register.
8359 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8360 "an integer without a # prefix"
8361 * config/msp430/msp430.md (epilogue_helper): Use it.
8363 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8366 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8367 AVX512F gather builtins.
8368 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8369 on gather decls with INTEGER_TYPE masktype.
8370 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8371 directly into the builtin rather than hoisting it before loop.
8373 PR tree-optimization/59387
8374 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8375 (scev_const_prop): If folded_casts and type has undefined overflow,
8376 use force_gimple_operand instead of force_gimple_operand_gsi and
8377 for each added stmt if it is assign with
8378 arith_code_with_undefined_signed_overflow, call
8379 rewrite_to_defined_overflow.
8380 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8381 gimple-fold.h instead.
8382 (arith_code_with_undefined_signed_overflow,
8383 rewrite_to_defined_overflow): Moved to ...
8384 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8385 rewrite_to_defined_overflow): ... here. No longer static.
8386 Include gimplify-me.h.
8387 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8388 rewrite_to_defined_overflow): New prototypes.
8390 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8392 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8394 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8396 * builtins.c (get_object_alignment_2): Minor tweak.
8397 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8399 2014-01-13 Christian Bruel <christian.bruel@st.com>
8401 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8402 optimized non constant lengths.
8404 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8407 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8408 load as __atomic_load_N if possible.
8410 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8412 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8414 (rs6000_expand_builtin): Adjust call.
8416 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8419 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8420 * config/rs6000/rs6000.c: Include target-globals.h.
8421 (rs6000_set_current_function): Instead of doing target_reinit
8422 unconditionally, use save_target_globals_default_opts and
8423 restore_target_globals.
8425 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8427 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8428 (rs6000_expand_builtin): Handle mffs and mtfsf.
8429 (rs6000_init_builtins): Define mffs and mtfsf.
8430 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8431 (rs6000_mffs): New pattern.
8432 (rs6000_mtfsf): New pattern.
8434 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8436 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8437 Start narrowing with START. Apply candidate-use pair
8438 and check overall cost in narrowing.
8439 (iv_ca_prune): Pass new argument.
8441 2014-01-10 Jeff Law <law@redhat.com>
8444 * ree.c (combine_reaching_defs): Ensure the defining statement
8445 occurs before the extension when optimizing extensions with
8446 different source and destination hard registers.
8448 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8451 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8452 vtables into the type inheritance graph.
8454 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8456 PR rtl-optimization/59754
8457 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8458 modes in the REGNO != REGNO case.
8460 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8462 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8464 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8466 PR tree-optimization/59745
8467 * tree-predcom.c (tree_predictive_commoning_loop): Call
8468 free_affine_expand_cache if giving up because components is NULL.
8470 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8471 GC in payload of target_globals struct instead of allocating them on
8472 the heap and the larger structs separately using GC.
8473 * target-globals.h (struct target_globals): Make regs, hard_regs,
8474 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8475 of GTY((skip)) and change type to void *.
8476 (reset_target_globals): Cast loads from those fields to corresponding
8479 2014-01-10 Steve Ellcey <sellcey@mips.com>
8482 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8483 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8484 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8486 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8489 * aarch64-modes.def (CC_Zmode): New flags mode.
8490 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8491 represents an equality.
8492 (aarch64_get_condition_code): Handle CC_Zmode.
8493 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8495 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8497 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8498 extraction in good case.
8500 2014-01-10 Richard Biener <rguenther@suse.de>
8502 PR tree-optimization/59374
8503 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8504 checking after SLP discovery. Mark stmts not participating
8505 in any SLP instance properly.
8507 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8509 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8510 when handling a SET rtx.
8512 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8514 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8515 (cortex-a57): Likewise.
8516 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8518 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8520 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8521 non-iwmmxt builtins.
8523 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8527 * ipa-devirt.c record_target_from_binfo): Take as argument
8528 stack of binfos and lookup matching one for virtual inheritance.
8529 (possible_polymorphic_call_targets_1): Update.
8531 2014-01-10 Huacai Chen <chenhc@lemote.com>
8533 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8534 kernel strings for Loongson-2E/2F/3A.
8536 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8539 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8540 is_gimple_call before calling gimple_call_internal_p.
8542 2014-01-09 Steve Ellcey <sellcey@mips.com>
8544 * Makefile.in (TREE_FLOW_H): Remove.
8545 (TREE_SSA_H): Add file names from tree-flow.h.
8546 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8547 * tree.h: Remove tree-flow.h reference.
8548 * hash-table.h: Remove tree-flow.h reference.
8549 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8550 reference with tree-ssa-loop.h.
8552 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8554 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8555 default element-order behavior for -maltivec.
8556 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8557 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8558 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8559 when targeting big endian, at least for now.
8560 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8562 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8565 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8566 var satisfies use_register_for_decl, just take into account type
8567 alignment, rather than decl alignment.
8569 PR tree-optimization/59622
8570 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8571 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8572 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8573 Don't devirtualize for inplace at all. For targets.length () == 1,
8574 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8576 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8578 * config/i386/i386.md (cpu): Remove the unused btver1.
8580 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8582 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8584 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8587 * tree-core.h (struct target_globals): New forward declaration.
8588 (struct tree_target_option): Add globals field.
8589 * tree.h (TREE_TARGET_GLOBALS): Define.
8590 (prepare_target_option_nodes_for_pch): New prototype.
8591 * target-globals.h (struct target_globals): Define even if
8593 * tree.c (prepare_target_option_node_for_pch,
8594 prepare_target_option_nodes_for_pch): New functions.
8595 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8596 * config/i386/i386.c: Include target-globals.h.
8597 (ix86_set_current_function): Instead of doing target_reinit
8598 unconditionally, use save_target_globals_default_opts and
8599 restore_target_globals.
8601 2014-01-09 Richard Biener <rguenther@suse.de>
8603 PR tree-optimization/59715
8604 * tree-cfg.h (split_critical_edges): Declare.
8605 * tree-cfg.c (split_critical_edges): Export.
8606 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8608 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8610 * cfgexpand.c (expand_stack_vars): Optionally disable
8611 asan stack protection.
8612 (expand_used_vars): Likewise.
8613 (partition_stack_vars): Likewise.
8614 * asan.c (asan_emit_stack_protection): Optionally disable
8615 after return stack usage.
8616 (instrument_derefs): Optionally disable memory access instrumentation.
8617 (instrument_builtin_call): Likewise.
8618 (instrument_strlen_call): Likewise.
8619 (asan_protect_global): Optionally disable global variables protection.
8620 * doc/invoke.texi: Added doc for new options.
8621 * params.def: Added new options.
8622 * params.h: Likewise.
8624 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8626 PR rtl-optimization/59724
8627 * ifcvt.c (cond_exec_process_if_block): Don't call
8628 flow_find_head_matching_sequence with 0 longest_match.
8629 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8630 non-active insns if !stop_after.
8631 (try_head_merge_bb): Revert 2014-01-07 changes.
8633 2014-01-08 Jeff Law <law@redhat.com>
8635 * ree.c (get_sub_rtx): New function, extracted from...
8636 (merge_def_and_ext): Here.
8637 (combine_reaching_defs): Use get_sub_rtx.
8639 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8641 * cgraph.h (varpool_variable_node): Do not choke on null node.
8643 2014-01-08 Catherine Moore <clm@codesourcery.com>
8645 * config/mips/mips.md (simple_return): Attempt to use JRC
8647 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8649 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8651 PR rtl-optimization/59137
8652 * reorg.c (steal_delay_list_from_target): Call update_block for
8654 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8656 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8658 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8659 two duplicate entries.
8661 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8664 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8666 * config/mips/mips.c (mips_truncated_op_cost): New function.
8667 (mips_rtx_costs): Adjust test for BADDU.
8668 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8670 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8672 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8673 (*baddu_si): ...this new pattern.
8675 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8678 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8680 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8683 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8685 (expand_expr, expand_normal): Adjust.
8686 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8687 inner_reference_p. Use inner_reference_p to expand inner references.
8688 (store_expr): Adjust.
8689 * cfgexpand.c (expand_call_stmt): Adjust.
8691 2014-01-08 Rong Xu <xur@google.com>
8693 * gcov-io.c (gcov_var): Move from gcov-io.h.
8694 (gcov_position): Ditto.
8695 (gcov_is_error): Ditto.
8696 (gcov_rewrite): Ditto.
8697 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8698 only part to libgcc/libgcov.h.
8700 2014-01-08 Marek Polacek <polacek@redhat.com>
8703 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8705 2014-01-08 Marek Polacek <polacek@redhat.com>
8708 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8710 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8712 PR rtl-optimization/59649
8713 * stor-layout.c (get_mode_bounds): For BImode return
8714 0 and STORE_FLAG_VALUE.
8716 2014-01-08 Richard Biener <rguenther@suse.de>
8719 * gimple.h (is_gimple_builtin_call): Remove.
8720 (gimple_builtin_call_types_compatible_p): New.
8721 (gimple_call_builtin_p): New overload.
8722 * gimple.c (is_gimple_builtin_call): Remove.
8723 (validate_call): Rename to ...
8724 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8726 (validate_type): New static function.
8727 (gimple_call_builtin_p): New overload and adjust.
8728 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8729 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8730 (gimple_fold_stmt_to_constant_1): Likewise.
8731 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8733 2014-01-08 Richard Biener <rguenther@suse.de>
8736 * gimplify.c (gimplify_expr): Gimplify register-register type
8737 VIEW_CONVERT_EXPRs to separate stmts.
8739 2014-01-07 Jeff Law <law@redhat.com>
8742 * ree.c (combine_set_extension): Handle case where source
8743 and destination registers in an extension insn are different.
8744 (combine_reaching_defs): Allow source and destination registers
8745 in extension to be different under limited circumstances.
8746 (add_removable_extension): Remove restriction that the
8747 source and destination registers in the extension are the same.
8748 (find_and_remove_re): Emit a copy from the extension's
8749 destination to its source after the defining insn if
8750 the source and destination registers are different.
8753 * ifcvt.c (merge_if_block): If we are merging a block with more than
8754 one successor with a block with no successors, remove any BARRIER
8755 after the second block.
8757 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8759 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8761 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8764 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8765 for 14-bit register offsets when INT14_OK_STRICT is false.
8767 2014-01-07 Roland Stigge <stigge@antcom.de>
8768 Michael Meissner <meissner@linux.vnet.ibm.com>
8771 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8772 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8774 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8776 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8779 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8781 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8782 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8783 rtx is const0_rtx or not.
8785 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8788 * target-globals.c (save_target_globals): Remove this_fn_optab
8790 * toplev.c: Include optabs.h.
8791 (target_reinit): Temporarily restore the global options if another
8792 set of options are in force.
8794 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8796 PR rtl-optimization/58668
8797 * cfgcleanup.c (flow_find_cross_jump): Don't count
8798 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8799 to determine what is counted.
8800 (flow_find_head_matching_sequence): Use active_insn_p to determine
8802 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8804 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8805 determine what is counted.
8807 PR tree-optimization/59643
8808 * tree-predcom.c (split_data_refs_to_components): If one dr is
8809 read and one write, determine_offset fails and the write isn't
8810 in the bad component, just put the read into the bad component.
8812 2014-01-07 Mike Stump <mikestump@comcast.net>
8813 Jakub Jelinek <jakub@redhat.com>
8816 * tree-core.h (struct tree_optimization_option): Change optabs
8817 type from unsigned char * to void *.
8818 * optabs.c (init_tree_optimization_optabs): Adjust
8819 TREE_OPTIMIZATION_OPTABS initialization.
8821 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8824 * config/i386/i386.h (struct machine_function): Add
8825 no_drap_save_restore field.
8826 * config/i386/i386.c (ix86_save_reg): Use
8827 !cfun->machine->no_drap_save_restore instead of
8828 crtl->stack_realign_needed.
8829 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8830 this function clears frame_pointer_needed. Set
8831 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8832 and DRAP reg is needed.
8834 2014-01-06 Marek Polacek <polacek@redhat.com>
8837 * doc/implement-c.texi: Mention that other integer types are
8838 permitted as bit-field types in strictly conforming mode.
8840 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8842 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8845 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8847 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8849 2014-01-06 Martin Jambor <mjambor@suse.cz>
8852 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8854 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8856 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8860 * var-tracking.c (add_stores): Preserve the value of the source even if
8861 we don't record the store.
8863 2014-01-06 Terry Guo <terry.guo@arm.com>
8865 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8867 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8870 * config/darwin.c (darwin_function_section): Adjust return values to
8871 correspond to optimisation changes made in r206070.
8873 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8875 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8876 from prefetch_block tune setting.
8877 (nocona_cost): Correct size of prefetch block to 64.
8879 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8881 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8882 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8883 used to save the static chain register in the computation of the offset
8884 from which the FP registers need to be restored.
8886 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8888 PR tree-optimization/59519
8889 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8890 ICE if get_current_def (current_new_name) is already non-NULL, as long
8891 as it is a phi result of some other phi in *new_exit_bb that has
8894 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8895 or vmovdqu* for misaligned_operand.
8896 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8897 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8898 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8899 aligned_mem for AVX512F masked aligned load and store builtins and for
8902 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8904 PR tree-optimization/59651
8905 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8906 Address range for negative step should be added by TYPE_SIZE_UNIT.
8908 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8910 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8911 overlapping registers also for registers other than the stack pointer.
8913 2014-01-03 Marek Polacek <polacek@redhat.com>
8916 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8919 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8922 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8925 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8926 (push splitter): Use <P:MODE_SIZE> instead of
8927 GET_MODE_SIZE (<P:MODE>mode).
8928 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8929 (mov -1, reg peephole2): Likewise.
8930 * config/i386/sse.md (*mov<mode>_internal,
8931 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8932 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8933 *<code><mode>3, *andnot<mode>3<mask_name>,
8934 <mask_codefor><code><mode>3<mask_name>): Likewise.
8935 * config/i386/subst.md (mask_mode512bit_condition,
8936 sd_mask_mode512bit_condition): Likewise.
8938 2014-01-02 Xinliang David Li <davidxl@google.com>
8940 PR tree-optimization/59303
8941 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8942 (dump_predicates): Better output format.
8943 (pred_equal_p): New function.
8944 (is_neq_relop_p): Ditto.
8945 (is_neq_zero_form_p): Ditto.
8946 (pred_expr_equal_p): Ditto.
8947 (pred_neg_p): Ditto.
8948 (simplify_pred): Ditto.
8949 (simplify_preds_2): Ditto.
8950 (simplify_preds_3): Ditto.
8951 (simplify_preds_4): Ditto.
8952 (simplify_preds): Ditto.
8954 (push_to_worklist): Ditto.
8955 (get_pred_info_from_cmp): Ditto.
8956 (is_degenerated_phi): Ditto.
8957 (normalize_one_pred_1): Ditto.
8958 (normalize_one_pred): Ditto.
8959 (normalize_one_pred_chain): Ditto.
8960 (normalize_preds): Ditto.
8961 (normalize_cond_1): Remove function.
8962 (normalize_cond): Ditto.
8963 (is_gcond_subset_of): Ditto.
8964 (is_subset_of_any): Ditto.
8965 (is_or_set_subset_of): Ditto.
8966 (is_and_set_subset_of): Ditto.
8967 (is_norm_cond_subset_of): Ditto.
8968 (pred_chain_length_cmp): Ditto.
8969 (convert_control_dep_chain_into_preds): Type change.
8970 (find_predicates): Ditto.
8971 (find_def_preds): Ditto.
8972 (destroy_predicates_vecs): Ditto.
8973 (find_matching_predicates_in_rest_chains): Ditto.
8974 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8975 (is_pred_expr_subset): Ditto.
8976 (is_pred_chain_subset_of): Ditto.
8977 (is_included_in): Ditto.
8978 (is_superset_of): Ditto.
8980 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8982 Update copyright years.
8984 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8986 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8987 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8988 config/arc/arc.md, config/arc/arc.opt,
8989 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8990 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8991 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8992 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8993 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8994 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8995 vtable-verify.c, vtable-verify.h: Use the standard form for the
8998 2014-01-02 Tobias Burnus <burnus@net-b.de>
9000 * gcc.c (process_command): Update copyright notice dates.
9001 * gcov-dump.c: Ditto.
9003 * doc/cpp.texi: Bump @copying's copyright year.
9004 * doc/cppinternals.texi: Ditto.
9005 * doc/gcc.texi: Ditto.
9006 * doc/gccint.texi: Ditto.
9007 * doc/gcov.texi: Ditto.
9008 * doc/install.texi: Ditto.
9009 * doc/invoke.texi: Ditto.
9011 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9013 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
9015 2014-01-01 Jakub Jelinek <jakub@redhat.com>
9017 * config/i386/sse.md (*mov<mode>_internal): Guard
9018 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
9020 PR rtl-optimization/59647
9021 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
9022 new_rtx into UNSIGNED_FLOAT rtxes.
9024 Copyright (C) 2014 Free Software Foundation, Inc.
9026 Copying and distribution of this file, with or without modification,
9027 are permitted in any medium without royalty provided the copyright
9028 notice and this notice are preserved.