1 2014-01-29 DJ Delorie <dj@redhat.com>
3 * config/msp430/msp430.opt (-minrt): New.
4 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6 (ENDFILE_SPEC): Likewise.
8 2014-01-29 Jan Hubicka <jh@suse.cz>
10 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
11 (estimate_function_body_sizes): Use it.
13 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
16 * dwarf2out.c (is_cxx_auto): New.
17 (is_base_type): Use it.
18 (gen_type_die_with_usage): Likewise.
20 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
23 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
24 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
25 -maltivec=be with LE targets.
26 (vsx_mergeh_<mode>): Likewise.
27 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New
29 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
30 (altivec_vmrghb): Replace with define_expand and new
31 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE
33 (altivec_vmrghb_direct): New define_insn.
34 (altivec_vmrghh): Replace with define_expand and new
35 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE
37 (altivec_vmrghh_direct): New define_insn.
38 (altivec_vmrghw): Replace with define_expand and new
39 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE
41 (altivec_vmrghw_direct): New define_insn.
42 (*altivec_vmrghsf): Adjust for endianness.
43 (altivec_vmrglb): Replace with define_expand and new
44 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE
46 (altivec_vmrglb_direct): New define_insn.
47 (altivec_vmrglh): Replace with define_expand and new
48 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE
50 (altivec_vmrglh_direct): New define_insn.
51 (altivec_vmrglw): Replace with define_expand and new
52 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE
54 (altivec_vmrglw_direct): New define_insn.
55 (*altivec_vmrglsf): Adjust for endianness.
56 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
57 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
58 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
59 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
60 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
61 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
62 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
63 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
65 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
67 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
68 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
71 2014-01-29 Richard Biener <rguenther@suse.de>
73 PR tree-optimization/58742
74 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
75 associate_pointerplus_align.
76 (associate_pointerplus_diff): New function.
77 (associate_pointerplus): Likewise. Call associate_pointerplus_align
78 and associate_pointerplus_diff.
80 2014-01-29 Richard Biener <rguenther@suse.de>
82 * lto-streamer.h (LTO_major_version): Bump to 3.
83 (LTO_minor_version): Reset to 0.
85 2014-01-29 Renlin Li <Renlin.Li@arm.com>
87 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
88 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
89 (arm_file_start): Generate correct asm header for armv7ve.
90 * config/arm/bpabi.h: Add multilib support for armv7ve.
91 * config/arm/driver-arm.c: Change the architectures of cortex-a7
92 and cortex-a15 to armv7ve.
93 * config/arm/t-aprofile: Add multilib support for armv7ve.
94 * doc/invoke.texi: Document -march=armv7ve.
96 2014-01-29 Richard Biener <rguenther@suse.de>
98 PR tree-optimization/58742
99 * tree-ssa-forwprop.c (associate_plusminus): Return true
100 if we changed sth, defer EH cleanup to ...
101 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
102 (simplify_mult): New function.
104 2014-01-29 Jakub Jelinek <jakub@redhat.com>
107 PR tree-optimization/59920
108 * tree.c (build_common_builtin_nodes): Remove
109 __builtin_setjmp_dispatcher initialization.
110 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
111 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
112 instead of gsi_after_labels + manually skipping debug stmts.
113 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
114 ignore bbs with IFN_ABNORMAL_DISPATCHER.
115 * tree-inline.c (copy_edges_for_bb): Remove
116 can_make_abnormal_goto argument, instead add abnormal_goto_dest
117 argument. Ignore computed_goto_p stmts. Don't call
118 make_abnormal_goto_edges. If a call might need abnormal edges
119 for non-local gotos, see if it already has an edge to
120 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
121 with true argument, don't do anything then, otherwise add
122 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
123 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
125 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
126 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
127 (lower_stmt): Don't set data->calls_builtin_setjmp.
128 (lower_builtin_setjmp): Adjust comment.
129 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
130 * tree-cfg.c (found_computed_goto): Remove.
131 (factor_computed_gotos): Remove.
132 (make_goto_expr_edges): Return bool, true for computed gotos.
133 Don't call make_abnormal_goto_edges.
134 (build_gimple_cfg): Don't set found_computed_goto, don't call
135 factor_computed_gotos.
136 (computed_goto_p): No longer static.
137 (make_blocks): Don't set found_computed_goto.
138 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
139 (make_edges): If make_goto_expr_edges returns true, push bb
140 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
141 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
142 vector. Record mapping between bbs and OpenMP regions if there
143 are any, adjust make_gimple_omp_edges caller. Call
144 handle_abnormal_edges.
145 (make_abnormal_goto_edges): Remove.
146 * tree-cfg.h (make_abnormal_goto_edges): Remove.
147 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
148 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
149 * builtins.c (expand_builtin): Don't handle
150 BUILT_IN_SETJMP_DISPATCHER.
151 * internal-fn.def (ABNORMAL_DISPATCHER): New.
152 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
153 filling *region also set *region_idx to (*region)->entry->index.
156 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
157 For REGs set ORIGINAL_REGNO.
159 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
161 * doc/md.texi: Mention that a target shouldn't implement
162 vec_widen_(s|u)mul_even/odd pair if it is less efficient
165 2014-01-29 Jakub Jelinek <jakub@redhat.com>
167 PR tree-optimization/59594
168 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
169 a copy of the datarefs vector rather than the vector itself.
171 2014-01-28 Jason Merrill <jason@redhat.com>
174 * dwarf2out.c (auto_die): New static.
175 (gen_type_die_with_usage): Handle C++1y 'auto'.
176 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
179 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
182 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
183 (SPEC_X32): Likewise.
185 * config/i386/i386.c (ix86_option_override_internal): Turn off
186 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
188 (x86_file_start): Output .code16gcc for TARGET_16BIT.
189 * config/i386/i386.h (TARGET_16BIT): New macro.
190 (TARGET_16BIT_P): Likewise.
191 * config/i386/i386.opt: Add m16.
192 * doc/invoke.texi: Document -m16.
194 2014-01-28 Jakub Jelinek <jakub@redhat.com>
196 PR preprocessor/59935
197 * input.c (location_get_source_line): Bail out on when line number
198 is zero, and test the return value of
199 lookup_or_add_file_to_cache_tab.
201 2014-01-28 Richard Biener <rguenther@suse.de>
203 PR tree-optimization/58742
204 * tree-ssa-forwprop.c (associate_plusminus): Handle
205 pointer subtraction of the form (T)(P + A) - (T)P.
207 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
209 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
212 2014-01-28 Richard Biener <rguenther@suse.de>
215 2014-01-28 Richard Biener <rguenther@suse.de>
217 PR rtl-optimization/45364
218 PR rtl-optimization/59890
219 * var-tracking.c (local_get_addr_clear_given_value): Handle
220 already cleared slot.
221 (val_reset): Handle not allocated local_get_addr_cache.
222 (vt_find_locations): Use post-order on the inverted CFG.
224 2014-01-28 Richard Biener <rguenther@suse.de>
226 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps):
229 2014-01-28 Richard Biener <rguenther@suse.de>
231 PR rtl-optimization/45364
232 PR rtl-optimization/59890
233 * var-tracking.c (local_get_addr_clear_given_value): Handle
234 already cleared slot.
235 (val_reset): Handle not allocated local_get_addr_cache.
236 (vt_find_locations): Use post-order on the inverted CFG.
238 2014-01-28 Alan Modra <amodra@gmail.com>
240 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
241 * configure.ac <recursive call for build != host>: Define
242 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
243 and LD_FOR_BUILD too.
244 * configure: Regenerate.
246 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
248 * config/i386/i386.c (get_builtin_code_for_version): Separate
249 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
250 Broadwell from Haswell.
252 2014-01-27 Steve Ellcey <sellcey@mips.com>
254 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
255 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
256 * config/mips/mips.c (mips_option_override): Change setting
258 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
259 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
260 Change from Mask to Var.
262 2014-01-27 Jeff Law <law@redhat.com>
264 * ipa-inline.c (inline_small_functions): Fix typo.
266 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
268 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
269 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
270 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
271 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
272 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
273 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
274 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
275 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
276 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
277 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
278 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
279 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
280 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
281 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
282 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
283 (_mm512_storeu_epi64): Ditto.
284 (_mm512_cmpge_epi32_mask): Ditto.
285 (_mm512_cmpge_epu32_mask): Ditto.
286 (_mm512_cmpge_epi64_mask): Ditto.
287 (_mm512_cmpge_epu64_mask): Ditto.
288 (_mm512_cmple_epi32_mask): Ditto.
289 (_mm512_cmple_epu32_mask): Ditto.
290 (_mm512_cmple_epi64_mask): Ditto.
291 (_mm512_cmple_epu64_mask): Ditto.
292 (_mm512_cmplt_epi32_mask): Ditto.
293 (_mm512_cmplt_epu32_mask): Ditto.
294 (_mm512_cmplt_epi64_mask): Ditto.
295 (_mm512_cmplt_epu64_mask): Ditto.
296 (_mm512_cmpneq_epi32_mask): Ditto.
297 (_mm512_cmpneq_epu32_mask): Ditto.
298 (_mm512_cmpneq_epi64_mask): Ditto.
299 (_mm512_cmpneq_epu64_mask): Ditto.
300 (_mm512_expand_pd): Ditto.
301 (_mm512_expand_ps): Ditto.
302 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
303 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
304 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
305 * config/i386/i386.c (ix86_builtins): Add
306 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
307 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
308 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
309 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
310 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
311 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
312 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
313 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
314 IX86_BUILTIN_PMOVUSQW512_MEM.
315 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
316 __builtin_ia32_pmovsqd512mem_mask,
317 __builtin_ia32_pmovqd512mem_mask,
318 __builtin_ia32_pmovusqw512mem_mask,
319 __builtin_ia32_pmovsqw512mem_mask,
320 __builtin_ia32_pmovqw512mem_mask,
321 __builtin_ia32_pmovusdw512mem_mask,
322 __builtin_ia32_pmovsdw512mem_mask,
323 __builtin_ia32_pmovdw512mem_mask,
324 __builtin_ia32_pmovqb512mem_mask,
325 __builtin_ia32_pmovusqb512mem_mask,
326 __builtin_ia32_pmovsqb512mem_mask,
327 __builtin_ia32_pmovusdb512mem_mask,
328 __builtin_ia32_pmovsdb512mem_mask,
329 __builtin_ia32_pmovdb512mem_mask.
330 (bdesc_args): Add __builtin_ia32_expanddf512,
331 __builtin_ia32_expandsf512.
332 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
333 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
334 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
335 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
336 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
337 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
338 (avx512f_<code>v8div16qi2_mask_store): This.
339 (avx512f_expand<mode>): New.
341 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
343 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
345 (_mm512_mask_prefetch_i64gather_pd): Ditto.
346 (_mm512_prefetch_i32scatter_pd): Ditto.
347 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
348 (_mm512_prefetch_i64scatter_pd): Ditto.
349 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
350 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
351 (_mm512_mask_prefetch_i64gather_ps): Ditto.
352 (_mm512_prefetch_i32scatter_ps): Ditto.
353 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
354 (_mm512_prefetch_i64scatter_ps): Ditto.
355 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
356 * config/i386/i386-builtin-types.def: Define
357 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
358 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
359 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
360 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
361 IX86_BUILTIN_SCATTERPFQPD.
362 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
363 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
364 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
365 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
366 __builtin_ia32_scatterpfqps.
367 (ix86_expand_builtin): Expand new built-ins.
368 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
369 fix memory access data type.
370 (*avx512pf_gatherpf<mode>_mask): Ditto.
371 (*avx512pf_gatherpf<mode>): Ditto.
372 (avx512pf_scatterpf<mode>): Ditto.
373 (*avx512pf_scatterpf<mode>_mask): Ditto.
374 (*avx512pf_scatterpf<mode>): Ditto.
375 (GATHER_SCATTER_SF_MEM_MODE): New.
376 (avx512pf_gatherpf<mode>df): Ditto.
377 (*avx512pf_gatherpf<mode>df_mask): Ditto.
378 (*avx512pf_scatterpf<mode>df): Ditto.
380 2014-01-27 Jakub Jelinek <jakub@redhat.com>
383 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
384 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
387 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
389 * common/config/arm/arm-common.c
390 (arm_rewrite_mcpu): Handle multiple names.
392 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
394 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
396 * gimple-builder.h (create_gimple_tmp): Delete.
398 2014-01-27 Christian Bruel <christian.bruel@st.com>
400 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
403 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
405 * config/pa/pa.md (call): Generate indirect long calls to non-local
406 functions when outputing 32-bit code.
407 (call_value): Likewise except for special call to buggy powf function.
409 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
410 portable runtime and PIC indirect calls.
411 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
412 and PIC call sequences. Use ldo instead of blr to set return register
413 in PIC call sequence.
415 2014-01-25 Walter Lee <walt@tilera.com>
417 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
418 avoid clobbering a live register.
420 2014-01-25 Walter Lee <walt@tilera.com>
422 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
423 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
424 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
425 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
427 2014-01-25 Walter Lee <walt@tilera.com>
429 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
430 arguments on even registers.
431 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
433 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
434 (BIGGEST_ALIGNMENT): Ditto.
435 (BIGGEST_FIELD_ALIGNMENT): Ditto.
437 2014-01-25 Walter Lee <walt@tilera.com>
439 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
440 insns before bundling.
441 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
443 2014-01-25 Walter Lee <walt@tilera.com>
445 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
446 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
447 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
449 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
451 * config/mips/constraints.md (kl): Delete.
452 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
453 define expands, using...
454 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
455 instructions for MIPS16.
456 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
457 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
459 2014-01-25 Walter Lee <walt@tilera.com>
461 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
465 2014-01-25 Walter Lee <walt@tilera.com>
467 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
468 (TARGET_EXPAND_TO_RTL_HOOK): Define.
470 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
472 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
475 2014-01-25 Jakub Jelinek <jakub@redhat.com>
477 * print-rtl.c (in_call_function_usage): New var.
478 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
479 EXPR_LIST mode as mode and not as reg note name.
482 * cfgloopmanip.c (copy_loop_info): If
483 loop->warned_aggressive_loop_optimizations, make sure
484 the flag is set in target loop too.
486 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
488 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
490 * builtins.def: Likewise.
491 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
492 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
493 * ira.c (ira_setup_eliminable_regset): Likewise.
494 * omp-low.c (gate_expand_omp): Likewise.
495 (execute_lower_omp): Likewise.
496 (diagnose_sb_0): Likewise.
497 (gate_diagnose_omp_blocks): Likewise.
498 (simd_clone_clauses_extract): Likewise.
501 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
503 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
504 correction for little endian...
505 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
508 2014-01-24 Jeff Law <law@redhat.com>
510 PR tree-optimization/59919
511 * tree-vrp.c (find_assert_locations_1): Do not register asserts
512 for non-returning calls.
514 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
516 * common/config/aarch64/aarch64-common.c
517 (aarch64_rewrite_mcpu): Handle multiple names.
518 * config/aarch64/aarch64.h
519 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
521 2014-01-24 Dodji Seketeli <dodji@redhat.com>
523 * input.c (add_file_to_cache_tab): Handle the case where fopen
526 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
529 * config/i386/i386.md (pushsf splitter): Get stack adjustment
530 from push operand if code of push isn't PRE_DEC.
532 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
535 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
536 -mquad-memory-atomic. Update -mquad-memory documentation to say
537 it is only used for non-atomic loads/stores.
539 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
540 -mquad-memory or -mquad-memory-atomic switches.
542 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
543 -mquad-memory-atomic to ISA 2.07 support.
545 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
546 to separate support of normal quad word memory operations (ldq, stq)
547 from the atomic quad word memory operations.
549 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
550 support to separate non-atomic quad word operations from atomic
551 quad word operations. Disable non-atomic quad word operations in
552 little endian mode so that we don't have to swap words after the
553 load and before the store.
554 (quad_load_store_p): Add comment about atomic quad word support.
555 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
556 options printed with -mdebug=reg.
558 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
559 -mquad-memory-atomic as the test for whether we have quad word
561 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
562 or -mp8-vector are used, allow byte/half-word atomic operations.
564 * config/rs6000/sync.md (load_lockedti): Insure that the address
565 is a proper indexed or indirect address for the lqarx instruction.
566 On little endian systems, swap the hi/lo registers after the lqarx
568 (load_lockedpti): Use indexed_or_indirect_operand predicate to
569 insure the address is valid for the lqarx instruction.
570 (store_conditionalti): Insure that the address is a proper indexed
571 or indirect address for the stqcrx. instruction. On little endian
572 systems, swap the hi/lo registers before doing the stqcrx.
574 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
575 insure the address is valid for the stqcrx. instruction.
577 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
578 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
579 type of quad memory support is available.
581 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
584 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
585 there is a danger of looping.
587 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
589 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
590 force flag_ira_loop_pressure if set via command line.
592 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
594 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
595 (ashr_simd): New builtin handling DI mode.
596 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
597 (aarch64_sshr_simddi): New match pattern.
598 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
599 (vshrd_n_s64): Likewise.
600 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
602 2014-01-23 Nick Clifton <nickc@redhat.com>
604 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
605 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
606 favour of mcu specific scripts.
607 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
610 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
611 Alex Velenko <Alex.Velenko@arm.com>
613 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
614 (vaddv_s16): Likewise.
615 (vaddv_s32): Likewise.
616 (vaddv_u8): Likewise.
617 (vaddv_u16): Likewise.
618 (vaddv_u32): Likewise.
619 (vaddvq_s8): Likewise.
620 (vaddvq_s16): Likewise.
621 (vaddvq_s32): Likewise.
622 (vaddvq_s64): Likewise.
623 (vaddvq_u8): Likewise.
624 (vaddvq_u16): Likewise.
625 (vaddvq_u32): Likewise.
626 (vaddvq_u64): Likewise.
627 (vaddv_f32): Likewise.
628 (vaddvq_f32): Likewise.
629 (vaddvq_f64): Likewise.
630 (vmaxv_f32): Likewise.
631 (vmaxv_s8): Likewise.
632 (vmaxv_s16): Likewise.
633 (vmaxv_s32): Likewise.
634 (vmaxv_u8): Likewise.
635 (vmaxv_u16): Likewise.
636 (vmaxv_u32): Likewise.
637 (vmaxvq_f32): Likewise.
638 (vmaxvq_f64): Likewise.
639 (vmaxvq_s8): Likewise.
640 (vmaxvq_s16): Likewise.
641 (vmaxvq_s32): Likewise.
642 (vmaxvq_u8): Likewise.
643 (vmaxvq_u16): Likewise.
644 (vmaxvq_u32): Likewise.
645 (vmaxnmv_f32): Likewise.
646 (vmaxnmvq_f32): Likewise.
647 (vmaxnmvq_f64): Likewise.
648 (vminv_f32): Likewise.
649 (vminv_s8): Likewise.
650 (vminv_s16): Likewise.
651 (vminv_s32): Likewise.
652 (vminv_u8): Likewise.
653 (vminv_u16): Likewise.
654 (vminv_u32): Likewise.
655 (vminvq_f32): Likewise.
656 (vminvq_f64): Likewise.
657 (vminvq_s8): Likewise.
658 (vminvq_s16): Likewise.
659 (vminvq_s32): Likewise.
660 (vminvq_u8): Likewise.
661 (vminvq_u16): Likewise.
662 (vminvq_u32): Likewise.
663 (vminnmv_f32): Likewise.
664 (vminnmvq_f32): Likewise.
665 (vminnmvq_f64): Likewise.
667 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
669 * config/aarch64/aarch64-simd.md
670 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
671 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
672 (*aarch64_mul3_elt<mode>): Likewise.
673 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
674 (*aarch64_mul3_elt_to_64v2df): Likewise.
675 (*aarch64_mla_elt<mode>): Likewise.
676 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
677 (*aarch64_mls_elt<mode>): Likewise.
678 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
679 (*aarch64_fma4_elt<mode>): Likewise.
680 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
681 (*aarch64_fma4_elt_to_64v2df): Likewise.
682 (*aarch64_fnma4_elt<mode>): Likewise.
683 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
684 (*aarch64_fnma4_elt_to_64v2df): Likewise.
685 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
686 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
687 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
688 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
689 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
690 (aarch64_sqdmull_lane<mode>_internal): Likewise.
691 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
693 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
695 * config/aarch64/aarch64-simd.md
696 (aarch64_be_checked_get_lane<mode>): New define_expand.
697 * config/aarch64/aarch64-simd-builtins.def
698 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
699 New builtin definition.
700 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
701 Use new safe be builtin.
703 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
705 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
707 (aarch64_be_st1<mode>): Likewise.
708 (aarch_ld1<VALL:mode>): Define_expand modified.
709 (aarch_st1<VALL:mode>): Likewise.
710 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
711 (UNSPEC_ST1): Likewise.
713 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
715 * config/microblaze/microblaze.md: Add trap insn and attribute
717 2014-01-23 Dodji Seketeli <dodji@redhat.com>
719 PR preprocessor/58580
720 * input.h (location_get_source_line): Take an additional line_size
722 (void diagnostics_file_cache_fini): Declare new function.
723 * input.c (struct fcache): New type.
724 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
725 New static constants.
726 (diagnostic_file_cache_init, total_lines_num)
727 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
728 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
729 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
730 (get_next_line, read_next_line, goto_next_line, read_line_num):
731 New static function definitions.
732 (diagnostic_file_cache_fini): New function.
733 (location_get_source_line): Take an additional output line_len
734 parameter. Re-write using lookup_or_add_file_to_cache_tab and
736 * diagnostic.c (diagnostic_finish): Call
737 diagnostic_file_cache_fini.
738 (adjust_line): Take an additional input parameter for the length
739 of the line, rather than calculating it with strlen.
740 (diagnostic_show_locus): Adjust the use of
741 location_get_source_line and adjust_line with respect to their new
742 signature. While displaying a line now, do not stop at the first
743 null byte. Rather, display the zero byte as a space and keep
744 going until we reach the size of the line.
745 * Makefile.in: Add vec.o to OBJS-libcommon
747 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
748 Ilya Tocar <ilya.tocar@intel.com>
750 * config/i386/avx512fintrin.h (_mm512_kmov): New.
751 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
752 (__builtin_ia32_kmov16): Ditto.
753 * config/i386/i386.md (UNSPEC_KMOV): New.
756 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
758 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
759 (_mm512_storeu_si512): Ditto.
761 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
764 * rtl.h (get_referenced_operands): Declare.
765 * recog.c (get_referenced_operands): New function.
766 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
767 operands have been referenced when recording LO_SUM references.
769 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
771 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
773 2014-01-22 Jan Hubicka <jh@suse.cz>
775 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
776 Enable for generic and recent AMD targets.
778 2014-01-22 Jan Hubicka <jh@suse.cz>
780 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
781 ARG_SIZE note when adjustment was eliminated.
783 2014-01-22 Jeff Law <law@redhat.com>
785 PR tree-optimization/59597
786 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
787 in file. Accept new argument REGISTERING and use it to modify
788 dump output appropriately.
789 (register_jump_thread): Corresponding changes.
790 (mark_threaded_blocks): Reinstate code to cancel unprofitable
791 thread paths involving joiner blocks. Add code to dump cancelled
792 jump threading paths.
794 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
796 PR rtl-optimization/59477
797 * lra-constraints.c (inherit_in_ebb): Process call for living hard
798 regs. Update reloads_num and potential_reload_hard_regs for all insns.
800 2014-01-22 Tom Tromey <tromey@redhat.com>
802 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
804 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
806 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
808 PR rtl-optimization/59896
809 * lra-constraints.c (process_alt_operands): Check unused note for
810 matched operands of insn with no output reloads.
812 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
814 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
815 (mips_move_from_gpr_cost): Likewise.
817 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
819 PR rtl-optimization/59858
820 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
821 ira_class_hard_regs_num.
822 (process_alt_operands): Increase reject for dying matched operand.
824 2014-01-21 Jakub Jelinek <jakub@redhat.com>
827 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
828 smaller than size, perform several stores or loads and stores
829 at dst + count - size to store or copy all of size bytes, rather
830 than just last modesize bytes.
832 2014-01-20 DJ Delorie <dj@redhat.com>
834 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
835 that CLOBBERs are REGs before propogating their values.
837 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
840 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
841 (cgraph_inline_failed_type): New function.
842 * cgraph.h (DEFCIFCODE): Add type.
843 (cgraph_inline_failed_type_t): New enum.
844 (cgraph_inline_failed_type): New prototype.
845 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
846 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
847 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
848 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
849 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
850 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
851 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
852 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
853 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
854 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
855 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
856 OPTIMIZATION_MISMATCH.
857 * tree-inline.c (expand_call_inline): Emit errors during
858 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
860 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
863 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
864 mode attribute in insn output.
866 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
868 * output.h (output_constant): Delete.
869 * varasm.c (output_constant): Make private.
871 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
873 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
875 2014-01-20 Jakub Jelinek <jakub@redhat.com>
878 * tree.h (fold_builtin_strcat): New prototype.
879 * builtins.c (fold_builtin_strcat): No longer static. Add len
880 argument, if non-NULL, don't call c_strlen. Optimize
881 directly into __builtin_memcpy instead of __builtin_strcpy.
882 (fold_builtin_2): Adjust fold_builtin_strcat caller.
883 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
885 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
887 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
888 for SImode_address_operand operands, having only a REG argument.
890 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
892 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
893 loader name using mbig-endian.
894 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
896 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
898 * doc/invoke.texi (-march): Clarify documentation for AArch64.
902 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
904 * config/aarch64/aarch64-protos.h
905 (aarch64_cannot_change_mode_class_ptr): Declare.
906 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
907 aarch64_cannot_change_mode_class_ptr): New.
908 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
909 backend hook aarch64_cannot_change_mode_class.
911 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
913 * common/config/aarch64/aarch64-common.c
914 (aarch64_handle_option): Don't handle any option order logic here.
915 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
916 selected_cpu, warn on architecture version mismatch.
917 (aarch64_override_options): Fix parsing order for option strings.
919 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
920 Iain Sandoe <iain@codesourcery.com>
923 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
924 warning. Amend comment to reflect current functionality.
926 2014-01-20 Richard Biener <rguenther@suse.de>
929 * builtins.c (fold_builtin_strcat): Remove case better handled
930 by tree-ssa-strlen.c.
932 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
934 * config/aarch64/aarch64.opt
935 (mcpu, march, mtune): Make case-insensitive.
937 2014-01-20 Jakub Jelinek <jakub@redhat.com>
940 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
941 if operands[1] is a REG or ZERO_EXTEND of a REG.
943 2014-01-19 Jan Hubicka <jh@suse.cz>
945 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
947 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
949 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
950 long non-pic millicode calls.
952 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
954 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
956 2014-01-19 Kito Cheng <kito@0xlab.org>
958 * builtins.c (expand_movstr): Check movstr expand done or fail.
960 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
961 H.J. Lu <hongjiu.lu@intel.com>
964 * config/i386/i386.md (*lea<mode>): Zero-extend return register
965 to DImode for zero-extended addresses.
967 2014-01-19 Jakub Jelinek <jakub@redhat.com>
969 PR rtl-optimization/57763
970 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
971 on the new indirect jump_insn and increment LABEL_NUSES (label).
973 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
977 * config.gcc (x86_archs): New variable.
978 (x86_64_archs): Likewise.
979 (x86_cpus): Likewise.
980 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
981 --with-arch/--with-cpu= options.
982 Support --with-arch=/--with-cpu={nehalem,westmere,
983 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
985 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
987 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
988 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
990 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
992 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
994 2014-01-18 Jakub Jelinek <jakub@redhat.com>
997 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
998 clear cpp_get_options (parse_in)->warn_unused_macros for
999 ix86_target_macros_internal with cpp_define.
1001 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
1003 * jump.c (delete_related_insns): Keep (use (insn))s.
1004 * reorg.c (redundant_insn): Check for barriers too.
1006 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1008 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
1010 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
1012 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
1013 call to $$dyncall when TARGET_LONG_CALLS is true.
1015 2014-01-17 Jeff Law <law@redhat.com>
1017 * ree.c (combine_set_extension): Temporarily disable test for
1018 changing number of hard registers.
1020 2014-01-17 Jan Hubicka <jh@suse.cz>
1023 * ipa-inline-analysis.c (inline_free_summary):
1024 Do not free summary of aliases.
1026 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1029 * gimplify.c (gimplify_expr): Use create_tmp_var
1030 instead of create_tmp_var_raw. If cond doesn't have
1031 integral type, don't add the IFN_ANNOTATE builtin at all.
1033 2014-01-17 Martin Jambor <mjambor@suse.cz>
1036 * ipa-cp.c (prev_edge_clone): New variable.
1037 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
1038 Also resize prev_edge_clone vector.
1039 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
1040 (ipcp_edge_removal_hook): New function.
1041 (ipcp_driver): Register ipcp_edge_removal_hook.
1043 2014-01-17 Andrew Pinski <apinski@cavium.com>
1044 Steve Ellcey <sellcey@mips.com>
1047 * config/mips/mips.c (mips_print_operand): Check operand mode instead
1050 2014-01-17 Jeff Law <law@redhat.com>
1053 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
1054 so that pass_ccp runs first.
1056 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1058 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
1059 (ix86_adjust_cost): Use !TARGET_XXX.
1060 (do_reorder_for_imul): Likewise.
1061 (swap_top_of_ready_list): Likewise.
1062 (ix86_sched_reorder): Likewise.
1064 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1066 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1067 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
1068 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
1069 (intel_memset): New. Duplicate slm_memset.
1070 (intel_cost): New. Duplicate slm_cost.
1071 (m_INTEL): New macro.
1072 (processor_target_table): Add "intel".
1073 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
1074 with PROCESSOR_INTEL for "intel".
1075 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
1076 PROCESSOR_SILVERMONT.
1077 (ix86_issue_rate): Likewise.
1078 (ix86_adjust_cost): Likewise.
1079 (ia32_multipass_dfa_lookahead): Likewise.
1080 (swap_top_of_ready_list): Likewise.
1081 (ix86_sched_reorder): Likewise.
1082 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
1083 instead of TARGET_OPT_AGU.
1084 * config/i386/i386.h (TARGET_INTEL): New.
1085 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
1086 (processor_type): Add PROCESSOR_INTEL.
1087 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
1088 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
1090 2014-01-17 Marek Polacek <polacek@redhat.com>
1093 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
1096 2014-01-17 Richard Biener <rguenther@suse.de>
1098 PR tree-optimization/46590
1099 * opts.c (default_options_table): Add entries for
1100 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
1101 all enabled at -O1 but not for -Og.
1102 * common.opt (fbranch-count-reg): Remove Init(1).
1103 (fmove-loop-invariants): Likewise.
1104 (ftree-pta): Likewise.
1106 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1108 * config/i386/i386.c (ix86_data_alignment): For compatibility with
1109 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
1110 decls to at least the GCC 4.8 used alignments.
1113 * tree-nested.c (convert_nonlocal_reference_stmt,
1114 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
1115 of GIMPLE_BIND stmts, adjust associated decls.
1117 2014-01-17 Richard Biener <rguenther@suse.de>
1119 PR tree-optimization/46590
1120 * vec.h (vec<>::bseach): New member function implementing
1121 binary search according to C89 bsearch.
1122 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
1123 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
1124 bitmap pointer again. Make accesses_in_loop a flat array.
1125 (mem_ref_obstack): New global.
1126 (outermost_indep_loop): Adjust for mem_ref->stored changes.
1127 (mark_ref_stored): Likewise.
1128 (ref_indep_loop_p_2): Likewise.
1129 (set_ref_stored_in_loop): New helper function.
1130 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
1131 (memref_free): Adjust.
1132 (record_mem_ref_loc): Simplify.
1133 (gather_mem_refs_stmt): Adjust.
1134 (sort_locs_in_loop_postorder_cmp): New function.
1135 (analyze_memory_references): Sort accesses_in_loop after
1136 loop postorder number.
1137 (find_ref_loc_in_loop_cmp): New function.
1138 (for_all_locs_in_loop): Find relevant cluster of locs in
1139 accesses_in_loop and iterate without recursion.
1140 (execute_sm): Avoid uninit warning.
1141 (struct ref_always_accessed): Simplify.
1142 (ref_always_accessed::operator ()): Likewise.
1143 (ref_always_accessed_p): Likewise.
1144 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
1145 loop postorder numbers here.
1146 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
1149 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
1152 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
1153 on decls for which assemble_alias has been called.
1155 2014-01-17 Nick Clifton <nickc@redhat.com>
1157 * config/msp430/msp430.opt: (mcpu): New option.
1158 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
1159 (msp430_option_override): Parse target_cpu. If the MCU name
1160 matches a generic string, clear target_mcu.
1161 (msp430_attr): Allow numeric interrupt values up to 63.
1162 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
1163 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
1165 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
1167 * config/msp430/msp430.md (popm): Use %J rather than %I.
1168 (addsi3): Use msp430_nonimmediate_operand for operand 2.
1169 (addhi_cy_i): Use immediate_operand for operand 2.
1170 * doc/invoke.texi: Document -mcpu option.
1172 2014-01-17 Richard Biener <rguenther@suse.de>
1174 PR rtl-optimization/38518
1175 * df.h (df_analyze_loop): Declare.
1176 * df-core.c: Include cfgloop.h.
1177 (df_analyze_1): Split out main part of df_analyze.
1178 (df_analyze): Adjust.
1179 (loop_inverted_post_order_compute): New function.
1180 (loop_post_order_compute): Likewise.
1181 (df_analyze_loop): New function avoiding whole-function
1183 * loop-invariant.c (find_defs): Use df_analyze_loop.
1184 (find_invariants): Adjust.
1185 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
1187 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
1189 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
1190 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
1192 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
1194 * ipa-ref.c (ipa_remove_stmt_references): Fix references
1195 traversal when removing references.
1197 2014-01-16 Jan Hubicka <jh@suse.cz>
1200 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
1202 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
1205 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
1206 pushing a reload for an autoinc when we had previously reloaded an
1207 inner part of the address.
1209 2014-01-16 Jakub Jelinek <jakub@redhat.com>
1211 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
1213 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
1214 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
1215 when not giving up or versioning for alias only because of
1217 (vect_analyze_data_ref_dependences): Set to true.
1218 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
1220 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
1221 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
1225 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
1228 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
1229 operand 0 predicate for gathers, use a new pseudo as subtarget.
1231 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1234 * lra-constraints.c (process_alt_operands): Add printing debug info.
1235 Check absence of input/output reloads for matched operands too.
1237 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1239 PR rtl-optimization/59835
1240 * ira.c (ira_init_register_move_cost): Increase cost for
1243 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
1245 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
1247 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
1250 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
1251 non-register objects. Use gen_(high/low)part more consistently.
1254 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1257 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
1258 endian support, remove tests for WORDS_BIG_ENDIAN.
1259 (p8_mfvsrd_3_<mode>): Likewise.
1260 (reload_gpr_from_vsx<mode>): Likewise.
1261 (reload_gpr_from_vsxsf): Likewise.
1262 (p8_mfvsrd_4_disf): Likewise.
1264 2014-01-16 Richard Biener <rguenther@suse.de>
1266 PR rtl-optimization/46590
1267 * lcm.c (compute_antinout_edge): Use postorder iteration.
1268 (compute_laterin): Use inverted postorder iteration.
1270 2014-01-16 Nick Clifton <nickc@redhat.com>
1273 * varasm.c (output_constant): Return the number of bytes actually
1275 (output_constructor_array_range): Update the field size with the
1276 number of bytes emitted by output_constant.
1277 (output_constructor_regular_field): Likewise. Also do not
1278 complain if the total number of bytes emitted is now greater
1279 than the expected fieldpos.
1280 * output.h (output_constant): Update prototype and descriptive comment.
1282 2014-01-16 Marek Polacek <polacek@redhat.com>
1285 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
1286 it is error_mark_node.
1288 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
1290 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
1291 VALID_AVX256_REG_OR_OI_MODE.
1293 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
1295 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
1296 current procedure should be profiled.
1298 2014-01-15 Andrew Pinski <apinski@cavium.com>
1300 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
1301 of moving from/to the STACK_REG register class.
1303 2014-01-15 Richard Henderson <rth@redhat.com>
1306 * reginfo.c (global_regs_decl): Globalize.
1307 * rtl.h (global_regs_decl): Declare.
1308 * ira.c (do_reload): Diagnose frame_pointer_needed and it
1309 reserved via global_regs.
1311 2014-01-15 Teresa Johnson <tejohnson@google.com>
1313 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
1315 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1317 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
1318 and vmulosh rather than call gen_vec_widen_smult_*.
1319 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
1320 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
1321 (vec_widen_smult_even_v16qi): Likewise.
1322 (vec_widen_umult_even_v8hi): Likewise.
1323 (vec_widen_smult_even_v8hi): Likewise.
1324 (vec_widen_umult_odd_v16qi): Likewise.
1325 (vec_widen_smult_odd_v16qi): Likewise.
1326 (vec_widen_umult_odd_v8hi): Likewise.
1327 (vec_widen_smult_odd_v8hi): Likewise.
1328 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
1329 vmuloub rather than call gen_vec_widen_umult_*.
1330 (vec_widen_umult_lo_v16qi): Likewise.
1331 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
1332 vmulosb rather than call gen_vec_widen_smult_*.
1333 (vec_widen_smult_lo_v16qi): Likewise.
1334 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
1335 rather than call gen_vec_widen_umult_*.
1336 (vec_widen_umult_lo_v8hi): Likewise.
1337 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
1338 rather than call gen_vec_widen_smult_*.
1339 (vec_widen_smult_lo_v8hi): Likewise.
1341 2014-01-15 Jeff Law <law@redhat.com>
1343 PR tree-optimization/59747
1344 * ree.c (find_and_remove_re): Properly handle case where a second
1345 eliminated extension requires widening a copy created for elimination
1346 of a prior extension.
1347 (combine_set_extension): Ensure that the number of hard regs needed
1348 for a destination register does not change when we widen it.
1350 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
1352 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
1353 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
1354 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
1355 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
1356 (avr-*-rtems*): Likewise.
1357 (bfin*-rtems*): Likewise.
1358 (moxie-*-rtems*): Likewise.
1359 (h8300-*-rtems*): Likewise.
1360 (i[34567]86-*-rtems*): Likewise.
1361 (lm32-*-rtems*): Likewise.
1362 (m32r-*-rtems*): Likewise.
1363 (m68k-*-rtems*): Likewise.
1364 (microblaze*-*-rtems*): Likewise.
1365 (mips*-*-rtems*): Likewise.
1366 (powerpc-*-rtems*): Likewise.
1367 (sh-*-rtems*): Likewise.
1368 (sparc-*-rtems*): Likewise.
1369 (sparc64-*-rtems*): Likewise.
1370 (v850-*-rtems*): Likewise.
1371 (m32c-*-rtems*): Likewise.
1373 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
1375 PR rtl-optimization/59511
1376 * ira.c (ira_init_register_move_cost): Use memory costs for some
1377 cases of register move cost calculations.
1378 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
1379 instead of BB frequency.
1380 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
1381 * lra-assigns.c (find_hard_regno_for): Ditto.
1383 2014-01-15 Richard Biener <rguenther@suse.de>
1385 PR tree-optimization/59822
1386 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
1387 (vectorizable_load): Use it to hoist defs of uses of invariant
1388 loads out of the loop.
1390 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1391 Kugan Vivekanandarajah <kuganv@linaro.org>
1394 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
1397 2014-01-15 Richard Biener <rguenther@suse.de>
1399 PR rtl-optimization/59802
1400 * lcm.c (compute_available): Use inverted postorder to seed
1401 the initial worklist.
1403 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1406 * config/s390/s390.c (s390_preferred_reload_class): Don't return
1407 ADDR_REGS for invalid symrefs in non-PIC code.
1409 2014-01-15 Jakub Jelinek <jakub@redhat.com>
1412 * builtins.c (determine_block_size): Initialize *probable_max_size
1413 even if len_rtx is CONST_INT.
1415 2014-01-14 Andrew Pinski <apinski@cavium.com>
1417 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
1418 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
1419 (cortexa53_tunings): Likewise.
1420 (aarch64_sched_issue_rate): New function.
1421 (TARGET_SCHED_ISSUE_RATE): Define.
1423 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1425 * ira-costs.c (find_costs_and_classes): Add missed
1426 ira_init_register_move_cost_if_necessary.
1428 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1431 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
1433 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
1436 * config/i386/i386.c (type_natural_mode): Add a bool parameter
1437 to indicate if type is used for function return value. Warn ABI
1438 change if the vector mode isn't available for function return value.
1439 (ix86_function_arg_advance): Pass false to type_natural_mode.
1440 (ix86_function_arg): Likewise.
1441 (ix86_gimplify_va_arg): Likewise.
1442 (function_arg_32): Don't warn ABI change.
1443 (ix86_function_value): Pass true to type_natural_mode.
1444 (ix86_return_in_memory): Likewise.
1445 (ix86_struct_value_rtx): Removed.
1446 (TARGET_STRUCT_VALUE_RTX): Likewise.
1448 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1450 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
1451 converting a conditional jump into a conditional return.
1453 2014-01-14 Richard Biener <rguenther@suse.de>
1455 PR tree-optimization/58921
1456 PR tree-optimization/59006
1457 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
1458 hoisting invariant stmts.
1459 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
1460 invariant loads on the preheader edge if possible.
1462 2014-01-14 Joey Ye <joey.ye@arm.com>
1464 * doc/plugin.texi (Building GCC plugins): Update to C++.
1466 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
1468 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
1469 (_mm_rcp28_round_ss): Ditto.
1470 (_mm_rsqrt28_round_sd): Ditto.
1471 (_mm_rsqrt28_round_ss): Ditto.
1472 (_mm_rcp28_sd): Ditto.
1473 (_mm_rcp28_ss): Ditto.
1474 (_mm_rsqrt28_sd): Ditto.
1475 (_mm_rsqrt28_ss): Ditto.
1476 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
1477 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
1478 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
1479 (IX86_BUILTIN_RCP28SD): Ditto.
1480 (IX86_BUILTIN_RCP28SS): Ditto.
1481 (IX86_BUILTIN_RSQRT28SD): Ditto.
1482 (IX86_BUILTIN_RSQRT28SS): Ditto.
1483 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
1484 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
1485 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
1486 (ix86_expand_special_args_builtin): Expand new FTYPE.
1487 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
1488 (srcp14<mode>): Make insn unary.
1489 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
1490 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
1491 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
1492 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
1493 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
1494 Fix rounding: make it SAE only.
1495 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
1497 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
1499 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
1500 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
1501 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
1502 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
1503 (round_saeonly_mask_scalar_operand4): Ditto.
1504 (round_saeonly_mask_scalar_op3): Ditto.
1505 (round_saeonly_mask_scalar_op4): Ditto.
1507 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1509 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
1510 Implement -maltivec=be for vec_insert and vec_extract.
1512 2014-01-10 DJ Delorie <dj@redhat.com>
1514 * config/msp430/msp430.md (call_internal): Don't allow memory
1515 references with SP as the base register.
1516 (call_value_internal): Likewise.
1517 * config/msp430/constraints.md (Yc): New. For memory references
1518 that don't use SP as a base register.
1520 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
1521 "an integer without a # prefix"
1522 * config/msp430/msp430.md (epilogue_helper): Use it.
1524 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1527 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
1528 AVX512F gather builtins.
1529 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
1530 on gather decls with INTEGER_TYPE masktype.
1531 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
1532 directly into the builtin rather than hoisting it before loop.
1534 PR tree-optimization/59387
1535 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
1536 (scev_const_prop): If folded_casts and type has undefined overflow,
1537 use force_gimple_operand instead of force_gimple_operand_gsi and
1538 for each added stmt if it is assign with
1539 arith_code_with_undefined_signed_overflow, call
1540 rewrite_to_defined_overflow.
1541 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
1542 gimple-fold.h instead.
1543 (arith_code_with_undefined_signed_overflow,
1544 rewrite_to_defined_overflow): Moved to ...
1545 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
1546 rewrite_to_defined_overflow): ... here. No longer static.
1547 Include gimplify-me.h.
1548 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
1549 rewrite_to_defined_overflow): New prototypes.
1551 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1553 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
1555 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
1557 * builtins.c (get_object_alignment_2): Minor tweak.
1558 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
1560 2014-01-13 Christian Bruel <christian.bruel@st.com>
1562 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
1563 optimized non constant lengths.
1565 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1568 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
1569 load as __atomic_load_N if possible.
1571 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1573 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
1575 (rs6000_expand_builtin): Adjust call.
1577 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1580 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
1581 * config/rs6000/rs6000.c: Include target-globals.h.
1582 (rs6000_set_current_function): Instead of doing target_reinit
1583 unconditionally, use save_target_globals_default_opts and
1584 restore_target_globals.
1586 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
1588 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
1589 (rs6000_expand_builtin): Handle mffs and mtfsf.
1590 (rs6000_init_builtins): Define mffs and mtfsf.
1591 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
1592 (rs6000_mffs): New pattern.
1593 (rs6000_mtfsf): New pattern.
1595 2014-01-11 Bin Cheng <bin.cheng@arm.com>
1597 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
1598 Start narrowing with START. Apply candidate-use pair
1599 and check overall cost in narrowing.
1600 (iv_ca_prune): Pass new argument.
1602 2014-01-10 Jeff Law <law@redhat.com>
1605 * ree.c (combine_reaching_defs): Ensure the defining statement
1606 occurs before the extension when optimizing extensions with
1607 different source and destination hard registers.
1609 2014-01-10 Jan Hubicka <jh@suse.cz>
1612 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
1613 vtables into the type inheritance graph.
1615 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1617 PR rtl-optimization/59754
1618 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
1619 modes in the REGNO != REGNO case.
1621 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1623 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
1625 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1627 PR tree-optimization/59745
1628 * tree-predcom.c (tree_predictive_commoning_loop): Call
1629 free_affine_expand_cache if giving up because components is NULL.
1631 * target-globals.c (save_target_globals): Allocate < 4KB structs using
1632 GC in payload of target_globals struct instead of allocating them on
1633 the heap and the larger structs separately using GC.
1634 * target-globals.h (struct target_globals): Make regs, hard_regs,
1635 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
1636 of GTY((skip)) and change type to void *.
1637 (reset_target_globals): Cast loads from those fields to corresponding
1640 2014-01-10 Steve Ellcey <sellcey@mips.com>
1643 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
1644 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
1645 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
1647 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
1650 * aarch64-modes.def (CC_Zmode): New flags mode.
1651 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
1652 represents an equality.
1653 (aarch64_get_condition_code): Handle CC_Zmode.
1654 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
1656 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1658 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
1659 extraction in good case.
1661 2014-01-10 Richard Biener <rguenther@suse.de>
1663 PR tree-optimization/59374
1664 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
1665 checking after SLP discovery. Mark stmts not participating
1666 in any SLP instance properly.
1668 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1670 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
1671 when handling a SET rtx.
1673 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1675 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
1676 (cortex-a57): Likewise.
1677 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
1679 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1681 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
1682 non-iwmmxt builtins.
1684 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
1688 * ipa-devirt.c record_target_from_binfo): Take as argument
1689 stack of binfos and lookup matching one for virtual inheritance.
1690 (possible_polymorphic_call_targets_1): Update.
1692 2014-01-10 Huacai Chen <chenhc@lemote.com>
1694 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
1695 kernel strings for Loongson-2E/2F/3A.
1697 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1700 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
1701 is_gimple_call before calling gimple_call_internal_p.
1703 2014-01-09 Steve Ellcey <sellcey@mips.com>
1705 * Makefile.in (TREE_FLOW_H): Remove.
1706 (TREE_SSA_H): Add file names from tree-flow.h.
1707 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
1708 * tree.h: Remove tree-flow.h reference.
1709 * hash-table.h: Remove tree-flow.h reference.
1710 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
1711 reference with tree-ssa-loop.h.
1713 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1715 * doc/invoke.texi: Add -maltivec={be,le} options, and document
1716 default element-order behavior for -maltivec.
1717 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
1718 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
1719 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
1720 when targeting big endian, at least for now.
1721 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
1723 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1726 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
1727 var satisfies use_register_for_decl, just take into account type
1728 alignment, rather than decl alignment.
1730 PR tree-optimization/59622
1731 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
1732 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
1733 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
1734 Don't devirtualize for inplace at all. For targets.length () == 1,
1735 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
1737 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
1739 * config/i386/i386.md (cpu): Remove the unused btver1.
1741 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
1743 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
1745 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1748 * tree-core.h (struct target_globals): New forward declaration.
1749 (struct tree_target_option): Add globals field.
1750 * tree.h (TREE_TARGET_GLOBALS): Define.
1751 (prepare_target_option_nodes_for_pch): New prototype.
1752 * target-globals.h (struct target_globals): Define even if
1754 * tree.c (prepare_target_option_node_for_pch,
1755 prepare_target_option_nodes_for_pch): New functions.
1756 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
1757 * config/i386/i386.c: Include target-globals.h.
1758 (ix86_set_current_function): Instead of doing target_reinit
1759 unconditionally, use save_target_globals_default_opts and
1760 restore_target_globals.
1762 2014-01-09 Richard Biener <rguenther@suse.de>
1764 PR tree-optimization/59715
1765 * tree-cfg.h (split_critical_edges): Declare.
1766 * tree-cfg.c (split_critical_edges): Export.
1767 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
1769 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
1771 * cfgexpand.c (expand_stack_vars): Optionally disable
1772 asan stack protection.
1773 (expand_used_vars): Likewise.
1774 (partition_stack_vars): Likewise.
1775 * asan.c (asan_emit_stack_protection): Optionally disable
1776 after return stack usage.
1777 (instrument_derefs): Optionally disable memory access instrumentation.
1778 (instrument_builtin_call): Likewise.
1779 (instrument_strlen_call): Likewise.
1780 (asan_protect_global): Optionally disable global variables protection.
1781 * doc/invoke.texi: Added doc for new options.
1782 * params.def: Added new options.
1783 * params.h: Likewise.
1785 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1787 PR rtl-optimization/59724
1788 * ifcvt.c (cond_exec_process_if_block): Don't call
1789 flow_find_head_matching_sequence with 0 longest_match.
1790 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
1791 non-active insns if !stop_after.
1792 (try_head_merge_bb): Revert 2014-01-07 changes.
1794 2014-01-08 Jeff Law <law@redhat.com>
1796 * ree.c (get_sub_rtx): New function, extracted from...
1797 (merge_def_and_ext): Here.
1798 (combine_reaching_defs): Use get_sub_rtx.
1800 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
1802 * cgraph.h (varpool_variable_node): Do not choke on null node.
1804 2014-01-08 Catherine Moore <clm@codesourcery.com>
1806 * config/mips/mips.md (simple_return): Attempt to use JRC
1808 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
1810 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
1812 PR rtl-optimization/59137
1813 * reorg.c (steal_delay_list_from_target): Call update_block for
1815 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
1817 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1819 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
1820 two duplicate entries.
1822 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
1825 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
1827 * config/mips/mips.c (mips_truncated_op_cost): New function.
1828 (mips_rtx_costs): Adjust test for BADDU.
1829 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
1831 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
1833 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
1834 (*baddu_si): ...this new pattern.
1836 2014-01-08 Jakub Jelinek <jakub@redhat.com>
1839 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
1841 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
1844 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
1846 (expand_expr, expand_normal): Adjust.
1847 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
1848 inner_reference_p. Use inner_reference_p to expand inner references.
1849 (store_expr): Adjust.
1850 * cfgexpand.c (expand_call_stmt): Adjust.
1852 2014-01-08 Rong Xu <xur@google.com>
1854 * gcov-io.c (gcov_var): Move from gcov-io.h.
1855 (gcov_position): Ditto.
1856 (gcov_is_error): Ditto.
1857 (gcov_rewrite): Ditto.
1858 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
1859 only part to libgcc/libgcov.h.
1861 2014-01-08 Marek Polacek <polacek@redhat.com>
1864 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
1866 2014-01-08 Marek Polacek <polacek@redhat.com>
1869 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
1871 2014-01-08 Jakub Jelinek <jakub@redhat.com>
1873 PR rtl-optimization/59649
1874 * stor-layout.c (get_mode_bounds): For BImode return
1875 0 and STORE_FLAG_VALUE.
1877 2014-01-08 Richard Biener <rguenther@suse.de>
1880 * gimple.h (is_gimple_builtin_call): Remove.
1881 (gimple_builtin_call_types_compatible_p): New.
1882 (gimple_call_builtin_p): New overload.
1883 * gimple.c (is_gimple_builtin_call): Remove.
1884 (validate_call): Rename to ...
1885 (gimple_builtin_call_types_compatible_p): ... this and export. Also
1887 (validate_type): New static function.
1888 (gimple_call_builtin_p): New overload and adjust.
1889 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
1890 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
1891 (gimple_fold_stmt_to_constant_1): Likewise.
1892 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
1894 2014-01-08 Richard Biener <rguenther@suse.de>
1897 * gimplify.c (gimplify_expr): Gimplify register-register type
1898 VIEW_CONVERT_EXPRs to separate stmts.
1900 2014-01-07 Jeff Law <law@redhat.com>
1903 * ree.c (combine_set_extension): Handle case where source
1904 and destination registers in an extension insn are different.
1905 (combine_reaching_defs): Allow source and destination registers
1906 in extension to be different under limited circumstances.
1907 (add_removable_extension): Remove restriction that the
1908 source and destination registers in the extension are the same.
1909 (find_and_remove_re): Emit a copy from the extension's
1910 destination to its source after the defining insn if
1911 the source and destination registers are different.
1914 * ifcvt.c (merge_if_block): If we are merging a block with more than
1915 one successor with a block with no successors, remove any BARRIER
1916 after the second block.
1918 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
1920 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
1922 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
1925 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
1926 for 14-bit register offsets when INT14_OK_STRICT is false.
1928 2014-01-07 Roland Stigge <stigge@antcom.de>
1929 Michael Meissner <meissner@linux.vnet.ibm.com>
1932 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
1933 Only check TFmode for SPE constants. Don't check TImode or TDmode.
1935 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
1937 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
1940 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
1942 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
1943 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
1944 rtx is const0_rtx or not.
1946 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
1949 * target-globals.c (save_target_globals): Remove this_fn_optab
1951 * toplev.c: Include optabs.h.
1952 (target_reinit): Temporarily restore the global options if another
1953 set of options are in force.
1955 2014-01-07 Jakub Jelinek <jakub@redhat.com>
1957 PR rtl-optimization/58668
1958 * cfgcleanup.c (flow_find_cross_jump): Don't count
1959 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
1960 to determine what is counted.
1961 (flow_find_head_matching_sequence): Use active_insn_p to determine
1963 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
1965 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
1966 determine what is counted.
1968 PR tree-optimization/59643
1969 * tree-predcom.c (split_data_refs_to_components): If one dr is
1970 read and one write, determine_offset fails and the write isn't
1971 in the bad component, just put the read into the bad component.
1973 2014-01-07 Mike Stump <mikestump@comcast.net>
1974 Jakub Jelinek <jakub@redhat.com>
1977 * tree-core.h (struct tree_optimization_option): Change optabs
1978 type from unsigned char * to void *.
1979 * optabs.c (init_tree_optimization_optabs): Adjust
1980 TREE_OPTIMIZATION_OPTABS initialization.
1982 2014-01-06 Jakub Jelinek <jakub@redhat.com>
1985 * config/i386/i386.h (struct machine_function): Add
1986 no_drap_save_restore field.
1987 * config/i386/i386.c (ix86_save_reg): Use
1988 !cfun->machine->no_drap_save_restore instead of
1989 crtl->stack_realign_needed.
1990 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
1991 this function clears frame_pointer_needed. Set
1992 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
1993 and DRAP reg is needed.
1995 2014-01-06 Marek Polacek <polacek@redhat.com>
1998 * doc/implement-c.texi: Mention that other integer types are
1999 permitted as bit-field types in strictly conforming mode.
2001 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
2003 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
2006 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
2008 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
2010 2014-01-06 Martin Jambor <mjambor@suse.cz>
2013 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
2015 * ipa-prop.c (ipa_print_node_params): Fix indentation.
2017 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
2021 * var-tracking.c (add_stores): Preserve the value of the source even if
2022 we don't record the store.
2024 2014-01-06 Terry Guo <terry.guo@arm.com>
2026 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
2028 2014-01-05 Iain Sandoe <iain@codesourcery.com>
2031 * config/darwin.c (darwin_function_section): Adjust return values to
2032 correspond to optimisation changes made in r206070.
2034 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
2036 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
2037 from prefetch_block tune setting.
2038 (nocona_cost): Correct size of prefetch block to 64.
2040 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
2042 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
2043 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
2044 used to save the static chain register in the computation of the offset
2045 from which the FP registers need to be restored.
2047 2014-01-04 Jakub Jelinek <jakub@redhat.com>
2049 PR tree-optimization/59519
2050 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
2051 ICE if get_current_def (current_new_name) is already non-NULL, as long
2052 as it is a phi result of some other phi in *new_exit_bb that has
2055 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
2056 or vmovdqu* for misaligned_operand.
2057 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
2058 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
2059 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
2060 aligned_mem for AVX512F masked aligned load and store builtins and for
2063 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
2065 PR tree-optimization/59651
2066 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
2067 Address range for negative step should be added by TYPE_SIZE_UNIT.
2069 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
2071 * config/m68k/m68k.c (handle_move_double): Handle pushes with
2072 overlapping registers also for registers other than the stack pointer.
2074 2014-01-03 Marek Polacek <polacek@redhat.com>
2077 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
2080 2014-01-03 Jakub Jelinek <jakub@redhat.com>
2083 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
2086 * config/i386/i386.md (MODE_SIZE): New mode attribute.
2087 (push splitter): Use <P:MODE_SIZE> instead of
2088 GET_MODE_SIZE (<P:MODE>mode).
2089 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
2090 (mov -1, reg peephole2): Likewise.
2091 * config/i386/sse.md (*mov<mode>_internal,
2092 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
2093 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
2094 *<code><mode>3, *andnot<mode>3<mask_name>,
2095 <mask_codefor><code><mode>3<mask_name>): Likewise.
2096 * config/i386/subst.md (mask_mode512bit_condition,
2097 sd_mask_mode512bit_condition): Likewise.
2099 2014-01-02 Xinliang David Li <davidxl@google.com>
2101 PR tree-optimization/59303
2102 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
2103 (dump_predicates): Better output format.
2104 (pred_equal_p): New function.
2105 (is_neq_relop_p): Ditto.
2106 (is_neq_zero_form_p): Ditto.
2107 (pred_expr_equal_p): Ditto.
2108 (pred_neg_p): Ditto.
2109 (simplify_pred): Ditto.
2110 (simplify_preds_2): Ditto.
2111 (simplify_preds_3): Ditto.
2112 (simplify_preds_4): Ditto.
2113 (simplify_preds): Ditto.
2115 (push_to_worklist): Ditto.
2116 (get_pred_info_from_cmp): Ditto.
2117 (is_degenerated_phi): Ditto.
2118 (normalize_one_pred_1): Ditto.
2119 (normalize_one_pred): Ditto.
2120 (normalize_one_pred_chain): Ditto.
2121 (normalize_preds): Ditto.
2122 (normalize_cond_1): Remove function.
2123 (normalize_cond): Ditto.
2124 (is_gcond_subset_of): Ditto.
2125 (is_subset_of_any): Ditto.
2126 (is_or_set_subset_of): Ditto.
2127 (is_and_set_subset_of): Ditto.
2128 (is_norm_cond_subset_of): Ditto.
2129 (pred_chain_length_cmp): Ditto.
2130 (convert_control_dep_chain_into_preds): Type change.
2131 (find_predicates): Ditto.
2132 (find_def_preds): Ditto.
2133 (destroy_predicates_vecs): Ditto.
2134 (find_matching_predicates_in_rest_chains): Ditto.
2135 (use_pred_not_overlap_with_undef_path_pred): Ditto.
2136 (is_pred_expr_subset): Ditto.
2137 (is_pred_chain_subset_of): Ditto.
2138 (is_included_in): Ditto.
2139 (is_superset_of): Ditto.
2141 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2143 Update copyright years.
2145 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2147 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
2148 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
2149 config/arc/arc.md, config/arc/arc.opt,
2150 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
2151 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
2152 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
2153 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
2154 config/linux-protos.h, config/linux.c, config/winnt-c.c,
2155 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
2156 vtable-verify.c, vtable-verify.h: Use the standard form for the
2159 2014-01-02 Tobias Burnus <burnus@net-b.de>
2161 * gcc.c (process_command): Update copyright notice dates.
2162 * gcov-dump.c: Ditto.
2164 * doc/cpp.texi: Bump @copying's copyright year.
2165 * doc/cppinternals.texi: Ditto.
2166 * doc/gcc.texi: Ditto.
2167 * doc/gccint.texi: Ditto.
2168 * doc/gcov.texi: Ditto.
2169 * doc/install.texi: Ditto.
2170 * doc/invoke.texi: Ditto.
2172 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2174 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
2176 2014-01-01 Jakub Jelinek <jakub@redhat.com>
2178 * config/i386/sse.md (*mov<mode>_internal): Guard
2179 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
2181 PR rtl-optimization/59647
2182 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
2183 new_rtx into UNSIGNED_FLOAT rtxes.
2185 Copyright (C) 2014 Free Software Foundation, Inc.
2187 Copying and distribution of this file, with or without modification,
2188 are permitted in any medium without royalty provided the copyright
2189 notice and this notice are preserved.