| blob | c72c2536fc480a7fa9a43ebe9f1c518d63e73e81 |
1 2014-01-29 Richard Biener <rguenther@suse.de>
3 PR tree-optimization/58742
4 * tree-ssa-forwprop.c (associate_plusminus): Return true
5 if we changed sth, defer EH cleanup to ...
6 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
7 (simplify_mult): New function.
9 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11 PR middle-end/59917
12 PR tree-optimization/59920
13 * tree.c (build_common_builtin_nodes): Remove
14 __builtin_setjmp_dispatcher initialization.
15 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
16 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
17 instead of gsi_after_labels + manually skipping debug stmts.
18 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
19 ignore bbs with IFN_ABNORMAL_DISPATCHER.
20 * tree-inline.c (copy_edges_for_bb): Remove
21 can_make_abnormal_goto argument, instead add abnormal_goto_dest
22 argument. Ignore computed_goto_p stmts. Don't call
23 make_abnormal_goto_edges. If a call might need abnormal edges
24 for non-local gotos, see if it already has an edge to
25 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
26 with true argument, don't do anything then, otherwise add
27 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
28 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
29 caller.
30 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
31 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
32 (lower_stmt): Don't set data->calls_builtin_setjmp.
33 (lower_builtin_setjmp): Adjust comment.
34 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
35 * tree-cfg.c (found_computed_goto): Remove.
36 (factor_computed_gotos): Remove.
37 (make_goto_expr_edges): Return bool, true for computed gotos.
38 Don't call make_abnormal_goto_edges.
39 (build_gimple_cfg): Don't set found_computed_goto, don't call
40 factor_computed_gotos.
41 (computed_goto_p): No longer static.
42 (make_blocks): Don't set found_computed_goto.
43 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
44 (make_edges): If make_goto_expr_edges returns true, push bb
45 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
46 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
47 vector. Record mapping between bbs and OpenMP regions if there
48 are any, adjust make_gimple_omp_edges caller. Call
49 handle_abnormal_edges.
50 (make_abnormal_goto_edges): Remove.
51 * tree-cfg.h (make_abnormal_goto_edges): Remove.
52 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
53 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
54 * builtins.c (expand_builtin): Don't handle
55 BUILT_IN_SETJMP_DISPATCHER.
56 * internal-fn.def (ABNORMAL_DISPATCHER): New.
57 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
58 filling *region also set *region_idx to (*region)->entry->index.
60 PR other/58712
61 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
62 For REGs set ORIGINAL_REGNO.
64 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
66 * doc/md.texi: Mention that a target shouldn't implement
67 vec_widen_(s|u)mul_even/odd pair if it is less efficient
68 than hi/lo pair.
70 2014-01-29 Jakub Jelinek <jakub@redhat.com>
72 PR tree-optimization/59594
73 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
74 a copy of the datarefs vector rather than the vector itself.
76 2014-01-28 Jason Merrill <jason@redhat.com>
78 PR c++/53756
79 * dwarf2out.c (auto_die): New static.
80 (gen_type_die_with_usage): Handle C++1y 'auto'.
81 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
82 on definition.
84 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
86 PR target/59672
87 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
88 (SPEC_X32): Likewise.
89 (SPEC_64): Likewise.
90 * config/i386/i386.c (ix86_option_override_internal): Turn off
91 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
92 for TARGET_16BIT.
93 (x86_file_start): Output .code16gcc for TARGET_16BIT.
94 * config/i386/i386.h (TARGET_16BIT): New macro.
95 (TARGET_16BIT_P): Likewise.
96 * config/i386/i386.opt: Add m16.
97 * doc/invoke.texi: Document -m16.
99 2014-01-28 Jakub Jelinek <jakub@redhat.com>
101 PR preprocessor/59935
102 * input.c (location_get_source_line): Bail out on when line number
103 is zero, and test the return value of
104 lookup_or_add_file_to_cache_tab.
106 2014-01-28 Richard Biener <rguenther@suse.de>
108 PR tree-optimization/58742
109 * tree-ssa-forwprop.c (associate_plusminus): Handle
110 pointer subtraction of the form (T)(P + A) - (T)P.
112 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
114 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
115 at const_int_cost.
117 2014-01-28 Richard Biener <rguenther@suse.de>
119 Revert
120 2014-01-28 Richard Biener <rguenther@suse.de>
122 PR rtl-optimization/45364
123 PR rtl-optimization/59890
124 * var-tracking.c (local_get_addr_clear_given_value): Handle
125 already cleared slot.
126 (val_reset): Handle not allocated local_get_addr_cache.
127 (vt_find_locations): Use post-order on the inverted CFG.
129 2014-01-28 Richard Biener <rguenther@suse.de>
131 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps):
132 Remove.
134 2014-01-28 Richard Biener <rguenther@suse.de>
136 PR rtl-optimization/45364
137 PR rtl-optimization/59890
138 * var-tracking.c (local_get_addr_clear_given_value): Handle
139 already cleared slot.
140 (val_reset): Handle not allocated local_get_addr_cache.
141 (vt_find_locations): Use post-order on the inverted CFG.
143 2014-01-28 Alan Modra <amodra@gmail.com>
145 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
146 * configure.ac <recursive call for build != host>: Define
147 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
148 and LD_FOR_BUILD too.
149 * configure: Regenerate.
151 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
153 * config/i386/i386.c (get_builtin_code_for_version): Separate
154 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
155 Broadwell from Haswell.
157 2014-01-27 Steve Ellcey <sellcey@mips.com>
159 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
160 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
161 * config/mips/mips.c (mips_option_override): Change setting
162 of TARGET_DSP.
163 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
164 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
165 Change from Mask to Var.
167 2014-01-27 Jeff Law <law@redhat.com>
169 * ipa-inline.c (inline_small_functions): Fix typo.
171 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
173 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
174 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
175 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
176 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
177 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
178 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
179 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
180 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
181 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
182 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
183 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
184 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
185 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
186 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
187 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
188 (_mm512_storeu_epi64): Ditto.
189 (_mm512_cmpge_epi32_mask): Ditto.
190 (_mm512_cmpge_epu32_mask): Ditto.
191 (_mm512_cmpge_epi64_mask): Ditto.
192 (_mm512_cmpge_epu64_mask): Ditto.
193 (_mm512_cmple_epi32_mask): Ditto.
194 (_mm512_cmple_epu32_mask): Ditto.
195 (_mm512_cmple_epi64_mask): Ditto.
196 (_mm512_cmple_epu64_mask): Ditto.
197 (_mm512_cmplt_epi32_mask): Ditto.
198 (_mm512_cmplt_epu32_mask): Ditto.
199 (_mm512_cmplt_epi64_mask): Ditto.
200 (_mm512_cmplt_epu64_mask): Ditto.
201 (_mm512_cmpneq_epi32_mask): Ditto.
202 (_mm512_cmpneq_epu32_mask): Ditto.
203 (_mm512_cmpneq_epi64_mask): Ditto.
204 (_mm512_cmpneq_epu64_mask): Ditto.
205 (_mm512_expand_pd): Ditto.
206 (_mm512_expand_ps): Ditto.
207 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
208 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
209 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
210 * config/i386/i386.c (ix86_builtins): Add
211 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
212 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
213 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
214 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
215 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
216 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
217 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
218 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
219 IX86_BUILTIN_PMOVUSQW512_MEM.
220 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
221 __builtin_ia32_pmovsqd512mem_mask,
222 __builtin_ia32_pmovqd512mem_mask,
223 __builtin_ia32_pmovusqw512mem_mask,
224 __builtin_ia32_pmovsqw512mem_mask,
225 __builtin_ia32_pmovqw512mem_mask,
226 __builtin_ia32_pmovusdw512mem_mask,
227 __builtin_ia32_pmovsdw512mem_mask,
228 __builtin_ia32_pmovdw512mem_mask,
229 __builtin_ia32_pmovqb512mem_mask,
230 __builtin_ia32_pmovusqb512mem_mask,
231 __builtin_ia32_pmovsqb512mem_mask,
232 __builtin_ia32_pmovusdb512mem_mask,
233 __builtin_ia32_pmovsdb512mem_mask,
234 __builtin_ia32_pmovdb512mem_mask.
235 (bdesc_args): Add __builtin_ia32_expanddf512,
236 __builtin_ia32_expandsf512.
237 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
238 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
239 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
240 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
241 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
242 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
243 (avx512f_<code>v8div16qi2_mask_store): This.
244 (avx512f_expand<mode>): New.
246 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
248 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
249 New.
250 (_mm512_mask_prefetch_i64gather_pd): Ditto.
251 (_mm512_prefetch_i32scatter_pd): Ditto.
252 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
253 (_mm512_prefetch_i64scatter_pd): Ditto.
254 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
255 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
256 (_mm512_mask_prefetch_i64gather_ps): Ditto.
257 (_mm512_prefetch_i32scatter_ps): Ditto.
258 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
259 (_mm512_prefetch_i64scatter_ps): Ditto.
260 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
261 * config/i386/i386-builtin-types.def: Define
262 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
263 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
264 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
265 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
266 IX86_BUILTIN_SCATTERPFQPD.
267 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
268 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
269 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
270 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
271 __builtin_ia32_scatterpfqps.
272 (ix86_expand_builtin): Expand new built-ins.
273 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
274 fix memory access data type.
275 (*avx512pf_gatherpf<mode>_mask): Ditto.
276 (*avx512pf_gatherpf<mode>): Ditto.
277 (avx512pf_scatterpf<mode>): Ditto.
278 (*avx512pf_scatterpf<mode>_mask): Ditto.
279 (*avx512pf_scatterpf<mode>): Ditto.
280 (GATHER_SCATTER_SF_MEM_MODE): New.
281 (avx512pf_gatherpf<mode>df): Ditto.
282 (*avx512pf_gatherpf<mode>df_mask): Ditto.
283 (*avx512pf_scatterpf<mode>df): Ditto.
285 2014-01-27 Jakub Jelinek <jakub@redhat.com>
287 PR bootstrap/59934
288 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
289 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
290 reached.
292 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
294 * common/config/arm/arm-common.c
295 (arm_rewrite_mcpu): Handle multiple names.
296 * config/arm/arm.h
297 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
299 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
301 * gimple-builder.h (create_gimple_tmp): Delete.
303 2014-01-27 Christian Bruel <christian.bruel@st.com>
305 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
306 words comparisons.
308 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
310 * config/pa/pa.md (call): Generate indirect long calls to non-local
311 functions when outputing 32-bit code.
312 (call_value): Likewise except for special call to buggy powf function.
314 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
315 portable runtime and PIC indirect calls.
316 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
317 and PIC call sequences. Use ldo instead of blr to set return register
318 in PIC call sequence.
320 2014-01-25 Walter Lee <walt@tilera.com>
322 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
323 avoid clobbering a live register.
325 2014-01-25 Walter Lee <walt@tilera.com>
327 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
328 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
329 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
330 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
332 2014-01-25 Walter Lee <walt@tilera.com>
334 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
335 arguments on even registers.
336 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
337 STACK_BOUNDARY.
338 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
339 (BIGGEST_ALIGNMENT): Ditto.
340 (BIGGEST_FIELD_ALIGNMENT): Ditto.
342 2014-01-25 Walter Lee <walt@tilera.com>
344 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
345 insns before bundling.
346 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
348 2014-01-25 Walter Lee <walt@tilera.com>
350 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
351 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
352 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
354 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
356 * config/mips/constraints.md (kl): Delete.
357 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
358 define expands, using...
359 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
360 instructions for MIPS16.
361 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
362 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
364 2014-01-25 Walter Lee <walt@tilera.com>
366 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
367 (clzdi2): Ditto.
368 (ffsdi2): Ditto.
370 2014-01-25 Walter Lee <walt@tilera.com>
372 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
373 (TARGET_EXPAND_TO_RTL_HOOK): Define.
375 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
377 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
378 Handle XOR.
380 2014-01-25 Jakub Jelinek <jakub@redhat.com>
382 * print-rtl.c (in_call_function_usage): New var.
383 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
384 EXPR_LIST mode as mode and not as reg note name.
386 PR middle-end/59561
387 * cfgloopmanip.c (copy_loop_info): If
388 loop->warned_aggressive_loop_optimizations, make sure
389 the flag is set in target loop too.
391 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
393 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
394 flag_cilkplus.
395 * builtins.def: Likewise.
396 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
397 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
398 * ira.c (ira_setup_eliminable_regset): Likewise.
399 * omp-low.c (gate_expand_omp): Likewise.
400 (execute_lower_omp): Likewise.
401 (diagnose_sb_0): Likewise.
402 (gate_diagnose_omp_blocks): Likewise.
403 (simd_clone_clauses_extract): Likewise.
404 (gate): Likewise.
406 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
408 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
409 correction for little endian...
410 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
411 here.
413 2014-01-24 Jeff Law <law@redhat.com>
415 PR tree-optimization/59919
416 * tree-vrp.c (find_assert_locations_1): Do not register asserts
417 for non-returning calls.
419 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
421 * common/config/aarch64/aarch64-common.c
422 (aarch64_rewrite_mcpu): Handle multiple names.
423 * config/aarch64/aarch64.h
424 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
426 2014-01-24 Dodji Seketeli <dodji@redhat.com>
428 * input.c (add_file_to_cache_tab): Handle the case where fopen
429 returns NULL.
431 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
433 PR target/59929
434 * config/i386/i386.md (pushsf splitter): Get stack adjustment
435 from push operand if code of push isn't PRE_DEC.
437 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
439 PR target/59909
440 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
441 -mquad-memory-atomic. Update -mquad-memory documentation to say
442 it is only used for non-atomic loads/stores.
444 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
445 -mquad-memory or -mquad-memory-atomic switches.
447 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
448 -mquad-memory-atomic to ISA 2.07 support.
450 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
451 to separate support of normal quad word memory operations (ldq, stq)
452 from the atomic quad word memory operations.
454 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
455 support to separate non-atomic quad word operations from atomic
456 quad word operations. Disable non-atomic quad word operations in
457 little endian mode so that we don't have to swap words after the
458 load and before the store.
459 (quad_load_store_p): Add comment about atomic quad word support.
460 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
461 options printed with -mdebug=reg.
463 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
464 -mquad-memory-atomic as the test for whether we have quad word
465 atomic instructions.
466 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
467 or -mp8-vector are used, allow byte/half-word atomic operations.
469 * config/rs6000/sync.md (load_lockedti): Insure that the address
470 is a proper indexed or indirect address for the lqarx instruction.
471 On little endian systems, swap the hi/lo registers after the lqarx
472 instruction.
473 (load_lockedpti): Use indexed_or_indirect_operand predicate to
474 insure the address is valid for the lqarx instruction.
475 (store_conditionalti): Insure that the address is a proper indexed
476 or indirect address for the stqcrx. instruction. On little endian
477 systems, swap the hi/lo registers before doing the stqcrx.
478 instruction.
479 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
480 insure the address is valid for the stqcrx. instruction.
482 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
483 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
484 type of quad memory support is available.
486 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
488 PR regression/59915
489 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
490 there is a danger of looping.
492 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
494 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
495 force flag_ira_loop_pressure if set via command line.
497 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
499 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
500 (ashr_simd): New builtin handling DI mode.
501 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
502 (aarch64_sshr_simddi): New match pattern.
503 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
504 (vshrd_n_s64): Likewise.
505 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
507 2014-01-23 Nick Clifton <nickc@redhat.com>
509 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
510 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
511 favour of mcu specific scripts.
512 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
513 430x multilibs.
515 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
516 Alex Velenko <Alex.Velenko@arm.com>
518 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
519 (vaddv_s16): Likewise.
520 (vaddv_s32): Likewise.
521 (vaddv_u8): Likewise.
522 (vaddv_u16): Likewise.
523 (vaddv_u32): Likewise.
524 (vaddvq_s8): Likewise.
525 (vaddvq_s16): Likewise.
526 (vaddvq_s32): Likewise.
527 (vaddvq_s64): Likewise.
528 (vaddvq_u8): Likewise.
529 (vaddvq_u16): Likewise.
530 (vaddvq_u32): Likewise.
531 (vaddvq_u64): Likewise.
532 (vaddv_f32): Likewise.
533 (vaddvq_f32): Likewise.
534 (vaddvq_f64): Likewise.
535 (vmaxv_f32): Likewise.
536 (vmaxv_s8): Likewise.
537 (vmaxv_s16): Likewise.
538 (vmaxv_s32): Likewise.
539 (vmaxv_u8): Likewise.
540 (vmaxv_u16): Likewise.
541 (vmaxv_u32): Likewise.
542 (vmaxvq_f32): Likewise.
543 (vmaxvq_f64): Likewise.
544 (vmaxvq_s8): Likewise.
545 (vmaxvq_s16): Likewise.
546 (vmaxvq_s32): Likewise.
547 (vmaxvq_u8): Likewise.
548 (vmaxvq_u16): Likewise.
549 (vmaxvq_u32): Likewise.
550 (vmaxnmv_f32): Likewise.
551 (vmaxnmvq_f32): Likewise.
552 (vmaxnmvq_f64): Likewise.
553 (vminv_f32): Likewise.
554 (vminv_s8): Likewise.
555 (vminv_s16): Likewise.
556 (vminv_s32): Likewise.
557 (vminv_u8): Likewise.
558 (vminv_u16): Likewise.
559 (vminv_u32): Likewise.
560 (vminvq_f32): Likewise.
561 (vminvq_f64): Likewise.
562 (vminvq_s8): Likewise.
563 (vminvq_s16): Likewise.
564 (vminvq_s32): Likewise.
565 (vminvq_u8): Likewise.
566 (vminvq_u16): Likewise.
567 (vminvq_u32): Likewise.
568 (vminnmv_f32): Likewise.
569 (vminnmvq_f32): Likewise.
570 (vminnmvq_f64): Likewise.
572 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
574 * config/aarch64/aarch64-simd.md
575 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
576 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
577 (*aarch64_mul3_elt<mode>): Likewise.
578 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
579 (*aarch64_mul3_elt_to_64v2df): Likewise.
580 (*aarch64_mla_elt<mode>): Likewise.
581 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
582 (*aarch64_mls_elt<mode>): Likewise.
583 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
584 (*aarch64_fma4_elt<mode>): Likewise.
585 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
586 (*aarch64_fma4_elt_to_64v2df): Likewise.
587 (*aarch64_fnma4_elt<mode>): Likewise.
588 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
589 (*aarch64_fnma4_elt_to_64v2df): Likewise.
590 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
591 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
592 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
593 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
594 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
595 (aarch64_sqdmull_lane<mode>_internal): Likewise.
596 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
598 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
600 * config/aarch64/aarch64-simd.md
601 (aarch64_be_checked_get_lane<mode>): New define_expand.
602 * config/aarch64/aarch64-simd-builtins.def
603 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
604 New builtin definition.
605 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
606 Use new safe be builtin.
608 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
610 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
611 New define_insn.
612 (aarch64_be_st1<mode>): Likewise.
613 (aarch_ld1<VALL:mode>): Define_expand modified.
614 (aarch_st1<VALL:mode>): Likewise.
615 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
616 (UNSPEC_ST1): Likewise.
618 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
620 * config/microblaze/microblaze.md: Add trap insn and attribute
622 2014-01-23 Dodji Seketeli <dodji@redhat.com>
624 PR preprocessor/58580
625 * input.h (location_get_source_line): Take an additional line_size
626 parameter.
627 (void diagnostics_file_cache_fini): Declare new function.
628 * input.c (struct fcache): New type.
629 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
630 New static constants.
631 (diagnostic_file_cache_init, total_lines_num)
632 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
633 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
634 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
635 (get_next_line, read_next_line, goto_next_line, read_line_num):
636 New static function definitions.
637 (diagnostic_file_cache_fini): New function.
638 (location_get_source_line): Take an additional output line_len
639 parameter. Re-write using lookup_or_add_file_to_cache_tab and
640 read_line_num.
641 * diagnostic.c (diagnostic_finish): Call
642 diagnostic_file_cache_fini.
643 (adjust_line): Take an additional input parameter for the length
644 of the line, rather than calculating it with strlen.
645 (diagnostic_show_locus): Adjust the use of
646 location_get_source_line and adjust_line with respect to their new
647 signature. While displaying a line now, do not stop at the first
648 null byte. Rather, display the zero byte as a space and keep
649 going until we reach the size of the line.
650 * Makefile.in: Add vec.o to OBJS-libcommon
652 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
653 Ilya Tocar <ilya.tocar@intel.com>
655 * config/i386/avx512fintrin.h (_mm512_kmov): New.
656 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
657 (__builtin_ia32_kmov16): Ditto.
658 * config/i386/i386.md (UNSPEC_KMOV): New.
659 (kmovw): Ditto.
661 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
663 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
664 (_mm512_storeu_si512): Ditto.
666 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
668 PR target/52125
669 * rtl.h (get_referenced_operands): Declare.
670 * recog.c (get_referenced_operands): New function.
671 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
672 operands have been referenced when recording LO_SUM references.
674 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
676 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
678 2014-01-22 Jan Hubicka <jh@suse.cz>
680 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
681 Enable for generic and recent AMD targets.
683 2014-01-22 Jan Hubicka <jh@suse.cz>
685 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
686 ARG_SIZE note when adjustment was eliminated.
688 2014-01-22 Jeff Law <law@redhat.com>
690 PR tree-optimization/59597
691 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
692 in file. Accept new argument REGISTERING and use it to modify
693 dump output appropriately.
694 (register_jump_thread): Corresponding changes.
695 (mark_threaded_blocks): Reinstate code to cancel unprofitable
696 thread paths involving joiner blocks. Add code to dump cancelled
697 jump threading paths.
699 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
701 PR rtl-optimization/59477
702 * lra-constraints.c (inherit_in_ebb): Process call for living hard
703 regs. Update reloads_num and potential_reload_hard_regs for all insns.
705 2014-01-22 Tom Tromey <tromey@redhat.com>
707 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
708 PARAMS.
709 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
711 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
713 PR rtl-optimization/59896
714 * lra-constraints.c (process_alt_operands): Check unused note for
715 matched operands of insn with no output reloads.
717 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
719 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
720 (mips_move_from_gpr_cost): Likewise.
722 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
724 PR rtl-optimization/59858
725 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
726 ira_class_hard_regs_num.
727 (process_alt_operands): Increase reject for dying matched operand.
729 2014-01-21 Jakub Jelinek <jakub@redhat.com>
731 PR target/59003
732 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
733 smaller than size, perform several stores or loads and stores
734 at dst + count - size to store or copy all of size bytes, rather
735 than just last modesize bytes.
737 2014-01-20 DJ Delorie <dj@redhat.com>
739 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
740 that CLOBBERs are REGs before propogating their values.
742 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
744 PR middle-end/59789
745 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
746 (cgraph_inline_failed_type): New function.
747 * cgraph.h (DEFCIFCODE): Add type.
748 (cgraph_inline_failed_type_t): New enum.
749 (cgraph_inline_failed_type): New prototype.
750 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
751 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
752 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
753 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
754 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
755 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
756 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
757 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
758 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
759 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
760 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
761 OPTIMIZATION_MISMATCH.
762 * tree-inline.c (expand_call_inline): Emit errors during
763 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
765 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
767 PR target/59685
768 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
769 mode attribute in insn output.
771 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
773 * output.h (output_constant): Delete.
774 * varasm.c (output_constant): Make private.
776 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
778 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
780 2014-01-20 Jakub Jelinek <jakub@redhat.com>
782 PR middle-end/59860
783 * tree.h (fold_builtin_strcat): New prototype.
784 * builtins.c (fold_builtin_strcat): No longer static. Add len
785 argument, if non-NULL, don't call c_strlen. Optimize
786 directly into __builtin_memcpy instead of __builtin_strcpy.
787 (fold_builtin_2): Adjust fold_builtin_strcat caller.
788 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
790 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
792 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
793 for SImode_address_operand operands, having only a REG argument.
795 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
797 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
798 loader name using mbig-endian.
799 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
801 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
803 * doc/invoke.texi (-march): Clarify documentation for AArch64.
804 (-mtune): Likewise.
805 (-mcpu): Likewise.
807 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
809 * config/aarch64/aarch64-protos.h
810 (aarch64_cannot_change_mode_class_ptr): Declare.
811 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
812 aarch64_cannot_change_mode_class_ptr): New.
813 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
814 backend hook aarch64_cannot_change_mode_class.
816 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
818 * common/config/aarch64/aarch64-common.c
819 (aarch64_handle_option): Don't handle any option order logic here.
820 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
821 selected_cpu, warn on architecture version mismatch.
822 (aarch64_override_options): Fix parsing order for option strings.
824 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
825 Iain Sandoe <iain@codesourcery.com>
827 PR bootstrap/59496
828 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
829 warning. Amend comment to reflect current functionality.
831 2014-01-20 Richard Biener <rguenther@suse.de>
833 PR middle-end/59860
834 * builtins.c (fold_builtin_strcat): Remove case better handled
835 by tree-ssa-strlen.c.
837 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
839 * config/aarch64/aarch64.opt
840 (mcpu, march, mtune): Make case-insensitive.
842 2014-01-20 Jakub Jelinek <jakub@redhat.com>
844 PR target/59880
845 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
846 if operands[1] is a REG or ZERO_EXTEND of a REG.
848 2014-01-19 Jan Hubicka <jh@suse.cz>
850 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
852 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
854 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
855 long non-pic millicode calls.
857 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
859 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
861 2014-01-19 Kito Cheng <kito@0xlab.org>
863 * builtins.c (expand_movstr): Check movstr expand done or fail.
865 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
866 H.J. Lu <hongjiu.lu@intel.com>
868 PR target/59379
869 * config/i386/i386.md (*lea<mode>): Zero-extend return register
870 to DImode for zero-extended addresses.
872 2014-01-19 Jakub Jelinek <jakub@redhat.com>
874 PR rtl-optimization/57763
875 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
876 on the new indirect jump_insn and increment LABEL_NUSES (label).
878 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
880 PR bootstrap/59580
881 PR bootstrap/59583
882 * config.gcc (x86_archs): New variable.
883 (x86_64_archs): Likewise.
884 (x86_cpus): Likewise.
885 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
886 --with-arch/--with-cpu= options.
887 Support --with-arch=/--with-cpu={nehalem,westmere,
888 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
890 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
892 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
893 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
895 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
897 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
899 2014-01-18 Jakub Jelinek <jakub@redhat.com>
901 PR target/58944
902 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
903 clear cpp_get_options (parse_in)->warn_unused_macros for
904 ix86_target_macros_internal with cpp_define.
906 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
908 * jump.c (delete_related_insns): Keep (use (insn))s.
909 * reorg.c (redundant_insn): Check for barriers too.
911 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
913 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
915 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
917 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
918 call to $$dyncall when TARGET_LONG_CALLS is true.
920 2014-01-17 Jeff Law <law@redhat.com>
922 * ree.c (combine_set_extension): Temporarily disable test for
923 changing number of hard registers.
925 2014-01-17 Jan Hubicka <jh@suse.cz>
927 PR middle-end/58125
928 * ipa-inline-analysis.c (inline_free_summary):
929 Do not free summary of aliases.
931 2014-01-17 Jakub Jelinek <jakub@redhat.com>
933 PR middle-end/59706
934 * gimplify.c (gimplify_expr): Use create_tmp_var
935 instead of create_tmp_var_raw. If cond doesn't have
936 integral type, don't add the IFN_ANNOTATE builtin at all.
938 2014-01-17 Martin Jambor <mjambor@suse.cz>
940 PR ipa/59736
941 * ipa-cp.c (prev_edge_clone): New variable.
942 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
943 Also resize prev_edge_clone vector.
944 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
945 (ipcp_edge_removal_hook): New function.
946 (ipcp_driver): Register ipcp_edge_removal_hook.
948 2014-01-17 Andrew Pinski <apinski@cavium.com>
949 Steve Ellcey <sellcey@mips.com>
951 PR target/59462
952 * config/mips/mips.c (mips_print_operand): Check operand mode instead
953 of operator mode.
955 2014-01-17 Jeff Law <law@redhat.com>
957 PR middle-end/57904
958 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
959 so that pass_ccp runs first.
961 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
963 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
964 (ix86_adjust_cost): Use !TARGET_XXX.
965 (do_reorder_for_imul): Likewise.
966 (swap_top_of_ready_list): Likewise.
967 (ix86_sched_reorder): Likewise.
969 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
971 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
972 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
973 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
974 (intel_memset): New. Duplicate slm_memset.
975 (intel_cost): New. Duplicate slm_cost.
976 (m_INTEL): New macro.
977 (processor_target_table): Add "intel".
978 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
979 with PROCESSOR_INTEL for "intel".
980 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
981 PROCESSOR_SILVERMONT.
982 (ix86_issue_rate): Likewise.
983 (ix86_adjust_cost): Likewise.
984 (ia32_multipass_dfa_lookahead): Likewise.
985 (swap_top_of_ready_list): Likewise.
986 (ix86_sched_reorder): Likewise.
987 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
988 instead of TARGET_OPT_AGU.
989 * config/i386/i386.h (TARGET_INTEL): New.
990 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
991 (processor_type): Add PROCESSOR_INTEL.
992 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
993 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
995 2014-01-17 Marek Polacek <polacek@redhat.com>
997 PR c/58346
998 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
999 size is zero.
1001 2014-01-17 Richard Biener <rguenther@suse.de>
1003 PR tree-optimization/46590
1004 * opts.c (default_options_table): Add entries for
1005 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
1006 all enabled at -O1 but not for -Og.
1007 * common.opt (fbranch-count-reg): Remove Init(1).
1008 (fmove-loop-invariants): Likewise.
1009 (ftree-pta): Likewise.
1011 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1013 * config/i386/i386.c (ix86_data_alignment): For compatibility with
1014 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
1015 decls to at least the GCC 4.8 used alignments.
1017 PR fortran/59440
1018 * tree-nested.c (convert_nonlocal_reference_stmt,
1019 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
1020 of GIMPLE_BIND stmts, adjust associated decls.
1022 2014-01-17 Richard Biener <rguenther@suse.de>
1024 PR tree-optimization/46590
1025 * vec.h (vec<>::bseach): New member function implementing
1026 binary search according to C89 bsearch.
1027 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
1028 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
1029 bitmap pointer again. Make accesses_in_loop a flat array.
1030 (mem_ref_obstack): New global.
1031 (outermost_indep_loop): Adjust for mem_ref->stored changes.
1032 (mark_ref_stored): Likewise.
1033 (ref_indep_loop_p_2): Likewise.
1034 (set_ref_stored_in_loop): New helper function.
1035 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
1036 (memref_free): Adjust.
1037 (record_mem_ref_loc): Simplify.
1038 (gather_mem_refs_stmt): Adjust.
1039 (sort_locs_in_loop_postorder_cmp): New function.
1040 (analyze_memory_references): Sort accesses_in_loop after
1041 loop postorder number.
1042 (find_ref_loc_in_loop_cmp): New function.
1043 (for_all_locs_in_loop): Find relevant cluster of locs in
1044 accesses_in_loop and iterate without recursion.
1045 (execute_sm): Avoid uninit warning.
1046 (struct ref_always_accessed): Simplify.
1047 (ref_always_accessed::operator ()): Likewise.
1048 (ref_always_accessed_p): Likewise.
1049 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
1050 loop postorder numbers here.
1051 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
1052 numbers.
1054 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
1056 PR c++/57945
1057 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
1058 on decls for which assemble_alias has been called.
1060 2014-01-17 Nick Clifton <nickc@redhat.com>
1062 * config/msp430/msp430.opt: (mcpu): New option.
1063 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
1064 (msp430_option_override): Parse target_cpu. If the MCU name
1065 matches a generic string, clear target_mcu.
1066 (msp430_attr): Allow numeric interrupt values up to 63.
1067 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
1068 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
1069 option.
1070 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
1071 Add mcpu matches.
1072 * config/msp430/msp430.md (popm): Use %J rather than %I.
1073 (addsi3): Use msp430_nonimmediate_operand for operand 2.
1074 (addhi_cy_i): Use immediate_operand for operand 2.
1075 * doc/invoke.texi: Document -mcpu option.
1077 2014-01-17 Richard Biener <rguenther@suse.de>
1079 PR rtl-optimization/38518
1080 * df.h (df_analyze_loop): Declare.
1081 * df-core.c: Include cfgloop.h.
1082 (df_analyze_1): Split out main part of df_analyze.
1083 (df_analyze): Adjust.
1084 (loop_inverted_post_order_compute): New function.
1085 (loop_post_order_compute): Likewise.
1086 (df_analyze_loop): New function avoiding whole-function
1087 postorder computes.
1088 * loop-invariant.c (find_defs): Use df_analyze_loop.
1089 (find_invariants): Adjust.
1090 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
1092 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
1094 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
1095 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
1097 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
1099 * ipa-ref.c (ipa_remove_stmt_references): Fix references
1100 traversal when removing references.
1102 2014-01-16 Jan Hubicka <jh@suse.cz>
1104 PR ipa/59775
1105 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
1107 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
1109 PR middle-end/56791
1110 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
1111 pushing a reload for an autoinc when we had previously reloaded an
1112 inner part of the address.
1114 2014-01-16 Jakub Jelinek <jakub@redhat.com>
1116 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
1117 field.
1118 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
1119 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
1120 when not giving up or versioning for alias only because of
1121 loop->safelen.
1122 (vect_analyze_data_ref_dependences): Set to true.
1123 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
1124 is a GIMPLE_PHI.
1125 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
1126 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
1127 to the condition.
1129 PR middle-end/58344
1130 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
1132 PR target/59839
1133 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
1134 operand 0 predicate for gathers, use a new pseudo as subtarget.
1136 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1138 PR middle-end/59609
1139 * lra-constraints.c (process_alt_operands): Add printing debug info.
1140 Check absence of input/output reloads for matched operands too.
1142 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1144 PR rtl-optimization/59835
1145 * ira.c (ira_init_register_move_cost): Increase cost for
1146 impossible modes.
1148 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
1150 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
1152 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
1154 PR target/59780
1155 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
1156 non-register objects. Use gen_(high/low)part more consistently.
1157 Fix assertions.
1159 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1161 PR target/59844
1162 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
1163 endian support, remove tests for WORDS_BIG_ENDIAN.
1164 (p8_mfvsrd_3_<mode>): Likewise.
1165 (reload_gpr_from_vsx<mode>): Likewise.
1166 (reload_gpr_from_vsxsf): Likewise.
1167 (p8_mfvsrd_4_disf): Likewise.
1169 2014-01-16 Richard Biener <rguenther@suse.de>
1171 PR rtl-optimization/46590
1172 * lcm.c (compute_antinout_edge): Use postorder iteration.
1173 (compute_laterin): Use inverted postorder iteration.
1175 2014-01-16 Nick Clifton <nickc@redhat.com>
1177 PR middle-end/28865
1178 * varasm.c (output_constant): Return the number of bytes actually
1179 emitted.
1180 (output_constructor_array_range): Update the field size with the
1181 number of bytes emitted by output_constant.
1182 (output_constructor_regular_field): Likewise. Also do not
1183 complain if the total number of bytes emitted is now greater
1184 than the expected fieldpos.
1185 * output.h (output_constant): Update prototype and descriptive comment.
1187 2014-01-16 Marek Polacek <polacek@redhat.com>
1189 PR middle-end/59827
1190 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
1191 it is error_mark_node.
1193 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
1195 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
1196 VALID_AVX256_REG_OR_OI_MODE.
1198 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
1200 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
1201 current procedure should be profiled.
1203 2014-01-15 Andrew Pinski <apinski@cavium.com>
1205 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
1206 of moving from/to the STACK_REG register class.
1208 2014-01-15 Richard Henderson <rth@redhat.com>
1210 PR debug/54694
1211 * reginfo.c (global_regs_decl): Globalize.
1212 * rtl.h (global_regs_decl): Declare.
1213 * ira.c (do_reload): Diagnose frame_pointer_needed and it
1214 reserved via global_regs.
1216 2014-01-15 Teresa Johnson <tejohnson@google.com>
1218 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
1220 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1222 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
1223 and vmulosh rather than call gen_vec_widen_smult_*.
1224 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
1225 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
1226 (vec_widen_smult_even_v16qi): Likewise.
1227 (vec_widen_umult_even_v8hi): Likewise.
1228 (vec_widen_smult_even_v8hi): Likewise.
1229 (vec_widen_umult_odd_v16qi): Likewise.
1230 (vec_widen_smult_odd_v16qi): Likewise.
1231 (vec_widen_umult_odd_v8hi): Likewise.
1232 (vec_widen_smult_odd_v8hi): Likewise.
1233 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
1234 vmuloub rather than call gen_vec_widen_umult_*.
1235 (vec_widen_umult_lo_v16qi): Likewise.
1236 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
1237 vmulosb rather than call gen_vec_widen_smult_*.
1238 (vec_widen_smult_lo_v16qi): Likewise.
1239 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
1240 rather than call gen_vec_widen_umult_*.
1241 (vec_widen_umult_lo_v8hi): Likewise.
1242 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
1243 rather than call gen_vec_widen_smult_*.
1244 (vec_widen_smult_lo_v8hi): Likewise.
1246 2014-01-15 Jeff Law <law@redhat.com>
1248 PR tree-optimization/59747
1249 * ree.c (find_and_remove_re): Properly handle case where a second
1250 eliminated extension requires widening a copy created for elimination
1251 of a prior extension.
1252 (combine_set_extension): Ensure that the number of hard regs needed
1253 for a destination register does not change when we widen it.
1255 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
1257 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
1258 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
1259 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
1260 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
1261 (avr-*-rtems*): Likewise.
1262 (bfin*-rtems*): Likewise.
1263 (moxie-*-rtems*): Likewise.
1264 (h8300-*-rtems*): Likewise.
1265 (i[34567]86-*-rtems*): Likewise.
1266 (lm32-*-rtems*): Likewise.
1267 (m32r-*-rtems*): Likewise.
1268 (m68k-*-rtems*): Likewise.
1269 (microblaze*-*-rtems*): Likewise.
1270 (mips*-*-rtems*): Likewise.
1271 (powerpc-*-rtems*): Likewise.
1272 (sh-*-rtems*): Likewise.
1273 (sparc-*-rtems*): Likewise.
1274 (sparc64-*-rtems*): Likewise.
1275 (v850-*-rtems*): Likewise.
1276 (m32c-*-rtems*): Likewise.
1278 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
1280 PR rtl-optimization/59511
1281 * ira.c (ira_init_register_move_cost): Use memory costs for some
1282 cases of register move cost calculations.
1283 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
1284 instead of BB frequency.
1285 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
1286 * lra-assigns.c (find_hard_regno_for): Ditto.
1288 2014-01-15 Richard Biener <rguenther@suse.de>
1290 PR tree-optimization/59822
1291 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
1292 (vectorizable_load): Use it to hoist defs of uses of invariant
1293 loads out of the loop.
1295 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1296 Kugan Vivekanandarajah <kuganv@linaro.org>
1298 PR target/59695
1299 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
1300 truncation.
1302 2014-01-15 Richard Biener <rguenther@suse.de>
1304 PR rtl-optimization/59802
1305 * lcm.c (compute_available): Use inverted postorder to seed
1306 the initial worklist.
1308 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1310 PR target/59803
1311 * config/s390/s390.c (s390_preferred_reload_class): Don't return
1312 ADDR_REGS for invalid symrefs in non-PIC code.
1314 2014-01-15 Jakub Jelinek <jakub@redhat.com>
1316 PR other/58712
1317 * builtins.c (determine_block_size): Initialize *probable_max_size
1318 even if len_rtx is CONST_INT.
1320 2014-01-14 Andrew Pinski <apinski@cavium.com>
1322 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
1323 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
1324 (cortexa53_tunings): Likewise.
1325 (aarch64_sched_issue_rate): New function.
1326 (TARGET_SCHED_ISSUE_RATE): Define.
1328 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1330 * ira-costs.c (find_costs_and_classes): Add missed
1331 ira_init_register_move_cost_if_necessary.
1333 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1335 PR target/59787
1336 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
1338 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
1340 PR target/59794
1341 * config/i386/i386.c (type_natural_mode): Add a bool parameter
1342 to indicate if type is used for function return value. Warn ABI
1343 change if the vector mode isn't available for function return value.
1344 (ix86_function_arg_advance): Pass false to type_natural_mode.
1345 (ix86_function_arg): Likewise.
1346 (ix86_gimplify_va_arg): Likewise.
1347 (function_arg_32): Don't warn ABI change.
1348 (ix86_function_value): Pass true to type_natural_mode.
1349 (ix86_return_in_memory): Likewise.
1350 (ix86_struct_value_rtx): Removed.
1351 (TARGET_STRUCT_VALUE_RTX): Likewise.
1353 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1355 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
1356 converting a conditional jump into a conditional return.
1358 2014-01-14 Richard Biener <rguenther@suse.de>
1360 PR tree-optimization/58921
1361 PR tree-optimization/59006
1362 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
1363 hoisting invariant stmts.
1364 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
1365 invariant loads on the preheader edge if possible.
1367 2014-01-14 Joey Ye <joey.ye@arm.com>
1369 * doc/plugin.texi (Building GCC plugins): Update to C++.
1371 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
1373 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
1374 (_mm_rcp28_round_ss): Ditto.
1375 (_mm_rsqrt28_round_sd): Ditto.
1376 (_mm_rsqrt28_round_ss): Ditto.
1377 (_mm_rcp28_sd): Ditto.
1378 (_mm_rcp28_ss): Ditto.
1379 (_mm_rsqrt28_sd): Ditto.
1380 (_mm_rsqrt28_ss): Ditto.
1381 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
1382 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
1383 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
1384 (IX86_BUILTIN_RCP28SD): Ditto.
1385 (IX86_BUILTIN_RCP28SS): Ditto.
1386 (IX86_BUILTIN_RSQRT28SD): Ditto.
1387 (IX86_BUILTIN_RSQRT28SS): Ditto.
1388 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
1389 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
1390 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
1391 (ix86_expand_special_args_builtin): Expand new FTYPE.
1392 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
1393 (srcp14<mode>): Make insn unary.
1394 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
1395 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
1396 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
1397 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
1398 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
1399 Fix rounding: make it SAE only.
1400 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
1401 Ditto.
1402 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
1403 Ditto.
1404 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
1405 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
1406 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
1407 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
1408 (round_saeonly_mask_scalar_operand4): Ditto.
1409 (round_saeonly_mask_scalar_op3): Ditto.
1410 (round_saeonly_mask_scalar_op4): Ditto.
1412 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1414 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
1415 Implement -maltivec=be for vec_insert and vec_extract.
1417 2014-01-10 DJ Delorie <dj@redhat.com>
1419 * config/msp430/msp430.md (call_internal): Don't allow memory
1420 references with SP as the base register.
1421 (call_value_internal): Likewise.
1422 * config/msp430/constraints.md (Yc): New. For memory references
1423 that don't use SP as a base register.
1425 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
1426 "an integer without a # prefix"
1427 * config/msp430/msp430.md (epilogue_helper): Use it.
1429 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1431 PR target/59617
1432 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
1433 AVX512F gather builtins.
1434 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
1435 on gather decls with INTEGER_TYPE masktype.
1436 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
1437 directly into the builtin rather than hoisting it before loop.
1439 PR tree-optimization/59387
1440 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
1441 (scev_const_prop): If folded_casts and type has undefined overflow,
1442 use force_gimple_operand instead of force_gimple_operand_gsi and
1443 for each added stmt if it is assign with
1444 arith_code_with_undefined_signed_overflow, call
1445 rewrite_to_defined_overflow.
1446 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
1447 gimple-fold.h instead.
1448 (arith_code_with_undefined_signed_overflow,
1449 rewrite_to_defined_overflow): Moved to ...
1450 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
1451 rewrite_to_defined_overflow): ... here. No longer static.
1452 Include gimplify-me.h.
1453 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
1454 rewrite_to_defined_overflow): New prototypes.
1456 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1458 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
1460 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
1462 * builtins.c (get_object_alignment_2): Minor tweak.
1463 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
1465 2014-01-13 Christian Bruel <christian.bruel@st.com>
1467 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
1468 optimized non constant lengths.
1470 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1472 PR libgomp/59194
1473 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
1474 load as __atomic_load_N if possible.
1476 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1478 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
1479 target parameter.
1480 (rs6000_expand_builtin): Adjust call.
1482 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1484 PR target/58115
1485 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
1486 * config/rs6000/rs6000.c: Include target-globals.h.
1487 (rs6000_set_current_function): Instead of doing target_reinit
1488 unconditionally, use save_target_globals_default_opts and
1489 restore_target_globals.
1491 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
1492 FPSCR.
1493 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
1494 (rs6000_expand_builtin): Handle mffs and mtfsf.
1495 (rs6000_init_builtins): Define mffs and mtfsf.
1496 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
1497 (rs6000_mffs): New pattern.
1498 (rs6000_mtfsf): New pattern.
1500 2014-01-11 Bin Cheng <bin.cheng@arm.com>
1502 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
1503 Start narrowing with START. Apply candidate-use pair
1504 and check overall cost in narrowing.
1505 (iv_ca_prune): Pass new argument.
1507 2014-01-10 Jeff Law <law@redhat.com>
1509 PR middle-end/59743
1510 * ree.c (combine_reaching_defs): Ensure the defining statement
1511 occurs before the extension when optimizing extensions with
1512 different source and destination hard registers.
1514 2014-01-10 Jan Hubicka <jh@suse.cz>
1516 PR ipa/58585
1517 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
1518 vtables into the type inheritance graph.
1520 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1522 PR rtl-optimization/59754
1523 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
1524 modes in the REGNO != REGNO case.
1526 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1528 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
1530 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1532 PR tree-optimization/59745
1533 * tree-predcom.c (tree_predictive_commoning_loop): Call
1534 free_affine_expand_cache if giving up because components is NULL.
1536 * target-globals.c (save_target_globals): Allocate < 4KB structs using
1537 GC in payload of target_globals struct instead of allocating them on
1538 the heap and the larger structs separately using GC.
1539 * target-globals.h (struct target_globals): Make regs, hard_regs,
1540 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
1541 of GTY((skip)) and change type to void *.
1542 (reset_target_globals): Cast loads from those fields to corresponding
1543 types.
1545 2014-01-10 Steve Ellcey <sellcey@mips.com>
1547 PR plugins/59335
1548 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
1549 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
1550 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
1552 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
1554 PR target/59744
1555 * aarch64-modes.def (CC_Zmode): New flags mode.
1556 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
1557 represents an equality.
1558 (aarch64_get_condition_code): Handle CC_Zmode.
1559 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
1561 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1563 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
1564 extraction in good case.
1566 2014-01-10 Richard Biener <rguenther@suse.de>
1568 PR tree-optimization/59374
1569 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
1570 checking after SLP discovery. Mark stmts not participating
1571 in any SLP instance properly.
1573 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1575 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
1576 when handling a SET rtx.
1578 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1580 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
1581 (cortex-a57): Likewise.
1582 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
1584 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1586 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
1587 non-iwmmxt builtins.
1589 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
1591 PR ipa/58252
1592 PR ipa/59226
1593 * ipa-devirt.c record_target_from_binfo): Take as argument
1594 stack of binfos and lookup matching one for virtual inheritance.
1595 (possible_polymorphic_call_targets_1): Update.
1597 2014-01-10 Huacai Chen <chenhc@lemote.com>
1599 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
1600 kernel strings for Loongson-2E/2F/3A.
1602 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1604 PR middle-end/59670
1605 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
1606 is_gimple_call before calling gimple_call_internal_p.
1608 2014-01-09 Steve Ellcey <sellcey@mips.com>
1610 * Makefile.in (TREE_FLOW_H): Remove.
1611 (TREE_SSA_H): Add file names from tree-flow.h.
1612 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
1613 * tree.h: Remove tree-flow.h reference.
1614 * hash-table.h: Remove tree-flow.h reference.
1615 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
1616 reference with tree-ssa-loop.h.
1618 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1620 * doc/invoke.texi: Add -maltivec={be,le} options, and document
1621 default element-order behavior for -maltivec.
1622 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
1623 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
1624 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
1625 when targeting big endian, at least for now.
1626 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
1628 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1630 PR middle-end/47735
1631 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
1632 var satisfies use_register_for_decl, just take into account type
1633 alignment, rather than decl alignment.
1635 PR tree-optimization/59622
1636 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
1637 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
1638 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
1639 Don't devirtualize for inplace at all. For targets.length () == 1,
1640 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
1642 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
1644 * config/i386/i386.md (cpu): Remove the unused btver1.
1646 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
1648 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
1650 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1652 PR target/58115
1653 * tree-core.h (struct target_globals): New forward declaration.
1654 (struct tree_target_option): Add globals field.
1655 * tree.h (TREE_TARGET_GLOBALS): Define.
1656 (prepare_target_option_nodes_for_pch): New prototype.
1657 * target-globals.h (struct target_globals): Define even if
1658 !SWITCHABLE_TARGET.
1659 * tree.c (prepare_target_option_node_for_pch,
1660 prepare_target_option_nodes_for_pch): New functions.
1661 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
1662 * config/i386/i386.c: Include target-globals.h.
1663 (ix86_set_current_function): Instead of doing target_reinit
1664 unconditionally, use save_target_globals_default_opts and
1665 restore_target_globals.
1667 2014-01-09 Richard Biener <rguenther@suse.de>
1669 PR tree-optimization/59715
1670 * tree-cfg.h (split_critical_edges): Declare.
1671 * tree-cfg.c (split_critical_edges): Export.
1672 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
1674 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
1676 * cfgexpand.c (expand_stack_vars): Optionally disable
1677 asan stack protection.
1678 (expand_used_vars): Likewise.
1679 (partition_stack_vars): Likewise.
1680 * asan.c (asan_emit_stack_protection): Optionally disable
1681 after return stack usage.
1682 (instrument_derefs): Optionally disable memory access instrumentation.
1683 (instrument_builtin_call): Likewise.
1684 (instrument_strlen_call): Likewise.
1685 (asan_protect_global): Optionally disable global variables protection.
1686 * doc/invoke.texi: Added doc for new options.
1687 * params.def: Added new options.
1688 * params.h: Likewise.
1690 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1692 PR rtl-optimization/59724
1693 * ifcvt.c (cond_exec_process_if_block): Don't call
1694 flow_find_head_matching_sequence with 0 longest_match.
1695 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
1696 non-active insns if !stop_after.
1697 (try_head_merge_bb): Revert 2014-01-07 changes.
1699 2014-01-08 Jeff Law <law@redhat.com>
1701 * ree.c (get_sub_rtx): New function, extracted from...
1702 (merge_def_and_ext): Here.
1703 (combine_reaching_defs): Use get_sub_rtx.
1705 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
1707 * cgraph.h (varpool_variable_node): Do not choke on null node.
1709 2014-01-08 Catherine Moore <clm@codesourcery.com>
1711 * config/mips/mips.md (simple_return): Attempt to use JRC
1712 for microMIPS.
1713 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
1715 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
1717 PR rtl-optimization/59137
1718 * reorg.c (steal_delay_list_from_target): Call update_block for
1719 elided insns.
1720 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
1722 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1724 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
1725 two duplicate entries.
1727 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
1729 Revert:
1730 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
1732 * config/mips/mips.c (mips_truncated_op_cost): New function.
1733 (mips_rtx_costs): Adjust test for BADDU.
1734 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
1736 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
1738 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
1739 (*baddu_si): ...this new pattern.
1741 2014-01-08 Jakub Jelinek <jakub@redhat.com>
1743 PR ipa/59722
1744 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
1746 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
1748 PR middle-end/57748
1749 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
1750 inner_reference_p.
1751 (expand_expr, expand_normal): Adjust.
1752 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
1753 inner_reference_p. Use inner_reference_p to expand inner references.
1754 (store_expr): Adjust.
1755 * cfgexpand.c (expand_call_stmt): Adjust.
1757 2014-01-08 Rong Xu <xur@google.com>
1759 * gcov-io.c (gcov_var): Move from gcov-io.h.
1760 (gcov_position): Ditto.
1761 (gcov_is_error): Ditto.
1762 (gcov_rewrite): Ditto.
1763 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
1764 only part to libgcc/libgcov.h.
1766 2014-01-08 Marek Polacek <polacek@redhat.com>
1768 PR middle-end/59669
1769 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
1771 2014-01-08 Marek Polacek <polacek@redhat.com>
1773 PR sanitizer/59667
1774 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
1776 2014-01-08 Jakub Jelinek <jakub@redhat.com>
1778 PR rtl-optimization/59649
1779 * stor-layout.c (get_mode_bounds): For BImode return
1780 0 and STORE_FLAG_VALUE.
1782 2014-01-08 Richard Biener <rguenther@suse.de>
1784 PR middle-end/59630
1785 * gimple.h (is_gimple_builtin_call): Remove.
1786 (gimple_builtin_call_types_compatible_p): New.
1787 (gimple_call_builtin_p): New overload.
1788 * gimple.c (is_gimple_builtin_call): Remove.
1789 (validate_call): Rename to ...
1790 (gimple_builtin_call_types_compatible_p): ... this and export. Also
1791 check return types.
1792 (validate_type): New static function.
1793 (gimple_call_builtin_p): New overload and adjust.
1794 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
1795 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
1796 (gimple_fold_stmt_to_constant_1): Likewise.
1797 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
1799 2014-01-08 Richard Biener <rguenther@suse.de>
1801 PR middle-end/59471
1802 * gimplify.c (gimplify_expr): Gimplify register-register type
1803 VIEW_CONVERT_EXPRs to separate stmts.
1805 2014-01-07 Jeff Law <law@redhat.com>
1807 PR middle-end/53623
1808 * ree.c (combine_set_extension): Handle case where source
1809 and destination registers in an extension insn are different.
1810 (combine_reaching_defs): Allow source and destination registers
1811 in extension to be different under limited circumstances.
1812 (add_removable_extension): Remove restriction that the
1813 source and destination registers in the extension are the same.
1814 (find_and_remove_re): Emit a copy from the extension's
1815 destination to its source after the defining insn if
1816 the source and destination registers are different.
1818 PR middle-end/59285
1819 * ifcvt.c (merge_if_block): If we are merging a block with more than
1820 one successor with a block with no successors, remove any BARRIER
1821 after the second block.
1823 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
1825 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
1827 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
1829 PR target/59652
1830 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
1831 for 14-bit register offsets when INT14_OK_STRICT is false.
1833 2014-01-07 Roland Stigge <stigge@antcom.de>
1834 Michael Meissner <meissner@linux.vnet.ibm.com>
1836 PR 57386/target
1837 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
1838 Only check TFmode for SPE constants. Don't check TImode or TDmode.
1840 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
1842 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
1843 -mcpu.
1845 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
1847 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
1848 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
1849 rtx is const0_rtx or not.
1851 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
1853 PR target/58115
1854 * target-globals.c (save_target_globals): Remove this_fn_optab
1855 handling.
1856 * toplev.c: Include optabs.h.
1857 (target_reinit): Temporarily restore the global options if another
1858 set of options are in force.
1860 2014-01-07 Jakub Jelinek <jakub@redhat.com>
1862 PR rtl-optimization/58668
1863 * cfgcleanup.c (flow_find_cross_jump): Don't count
1864 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
1865 to determine what is counted.
1866 (flow_find_head_matching_sequence): Use active_insn_p to determine
1867 what is counted.
1868 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
1869 counting change.
1870 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
1871 determine what is counted.
1873 PR tree-optimization/59643
1874 * tree-predcom.c (split_data_refs_to_components): If one dr is
1875 read and one write, determine_offset fails and the write isn't
1876 in the bad component, just put the read into the bad component.
1878 2014-01-07 Mike Stump <mikestump@comcast.net>
1879 Jakub Jelinek <jakub@redhat.com>
1881 PR pch/59436
1882 * tree-core.h (struct tree_optimization_option): Change optabs
1883 type from unsigned char * to void *.
1884 * optabs.c (init_tree_optimization_optabs): Adjust
1885 TREE_OPTIMIZATION_OPTABS initialization.
1887 2014-01-06 Jakub Jelinek <jakub@redhat.com>
1889 PR target/59644
1890 * config/i386/i386.h (struct machine_function): Add
1891 no_drap_save_restore field.
1892 * config/i386/i386.c (ix86_save_reg): Use
1893 !cfun->machine->no_drap_save_restore instead of
1894 crtl->stack_realign_needed.
1895 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
1896 this function clears frame_pointer_needed. Set
1897 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
1898 and DRAP reg is needed.
1900 2014-01-06 Marek Polacek <polacek@redhat.com>
1902 PR c/57773
1903 * doc/implement-c.texi: Mention that other integer types are
1904 permitted as bit-field types in strictly conforming mode.
1906 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
1908 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
1909 is newly allocated.
1911 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
1913 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
1915 2014-01-06 Martin Jambor <mjambor@suse.cz>
1917 PR ipa/59008
1918 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
1919 to int.
1920 * ipa-prop.c (ipa_print_node_params): Fix indentation.
1922 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
1924 PR debug/59350
1925 PR debug/59510
1926 * var-tracking.c (add_stores): Preserve the value of the source even if
1927 we don't record the store.
1929 2014-01-06 Terry Guo <terry.guo@arm.com>
1931 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
1933 2014-01-05 Iain Sandoe <iain@codesourcery.com>
1935 PR bootstrap/59541
1936 * config/darwin.c (darwin_function_section): Adjust return values to
1937 correspond to optimisation changes made in r206070.
1939 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
1941 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
1942 from prefetch_block tune setting.
1943 (nocona_cost): Correct size of prefetch block to 64.
1945 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
1947 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
1948 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
1949 used to save the static chain register in the computation of the offset
1950 from which the FP registers need to be restored.
1952 2014-01-04 Jakub Jelinek <jakub@redhat.com>
1954 PR tree-optimization/59519
1955 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
1956 ICE if get_current_def (current_new_name) is already non-NULL, as long
1957 as it is a phi result of some other phi in *new_exit_bb that has
1958 the same argument.
1960 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
1961 or vmovdqu* for misaligned_operand.
1962 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
1963 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
1964 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
1965 aligned_mem for AVX512F masked aligned load and store builtins and for
1966 non-temporal moves.
1968 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
1970 PR tree-optimization/59651
1971 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
1972 Address range for negative step should be added by TYPE_SIZE_UNIT.
1974 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
1976 * config/m68k/m68k.c (handle_move_double): Handle pushes with
1977 overlapping registers also for registers other than the stack pointer.
1979 2014-01-03 Marek Polacek <polacek@redhat.com>
1981 PR other/59661
1982 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
1983 __builtin_FILE.
1985 2014-01-03 Jakub Jelinek <jakub@redhat.com>
1987 PR target/59625
1988 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
1989 asm goto as jump.
1991 * config/i386/i386.md (MODE_SIZE): New mode attribute.
1992 (push splitter): Use <P:MODE_SIZE> instead of
1993 GET_MODE_SIZE (<P:MODE>mode).
1994 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
1995 (mov -1, reg peephole2): Likewise.
1996 * config/i386/sse.md (*mov<mode>_internal,
1997 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
1998 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
1999 *<code><mode>3, *andnot<mode>3<mask_name>,
2000 <mask_codefor><code><mode>3<mask_name>): Likewise.
2001 * config/i386/subst.md (mask_mode512bit_condition,
2002 sd_mask_mode512bit_condition): Likewise.
2004 2014-01-02 Xinliang David Li <davidxl@google.com>
2006 PR tree-optimization/59303
2007 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
2008 (dump_predicates): Better output format.
2009 (pred_equal_p): New function.
2010 (is_neq_relop_p): Ditto.
2011 (is_neq_zero_form_p): Ditto.
2012 (pred_expr_equal_p): Ditto.
2013 (pred_neg_p): Ditto.
2014 (simplify_pred): Ditto.
2015 (simplify_preds_2): Ditto.
2016 (simplify_preds_3): Ditto.
2017 (simplify_preds_4): Ditto.
2018 (simplify_preds): Ditto.
2019 (push_pred): Ditto.
2020 (push_to_worklist): Ditto.
2021 (get_pred_info_from_cmp): Ditto.
2022 (is_degenerated_phi): Ditto.
2023 (normalize_one_pred_1): Ditto.
2024 (normalize_one_pred): Ditto.
2025 (normalize_one_pred_chain): Ditto.
2026 (normalize_preds): Ditto.
2027 (normalize_cond_1): Remove function.
2028 (normalize_cond): Ditto.
2029 (is_gcond_subset_of): Ditto.
2030 (is_subset_of_any): Ditto.
2031 (is_or_set_subset_of): Ditto.
2032 (is_and_set_subset_of): Ditto.
2033 (is_norm_cond_subset_of): Ditto.
2034 (pred_chain_length_cmp): Ditto.
2035 (convert_control_dep_chain_into_preds): Type change.
2036 (find_predicates): Ditto.
2037 (find_def_preds): Ditto.
2038 (destroy_predicates_vecs): Ditto.
2039 (find_matching_predicates_in_rest_chains): Ditto.
2040 (use_pred_not_overlap_with_undef_path_pred): Ditto.
2041 (is_pred_expr_subset): Ditto.
2042 (is_pred_chain_subset_of): Ditto.
2043 (is_included_in): Ditto.
2044 (is_superset_of): Ditto.
2046 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2048 Update copyright years.
2050 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2052 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
2053 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
2054 config/arc/arc.md, config/arc/arc.opt,
2055 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
2056 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
2057 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
2058 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
2059 config/linux-protos.h, config/linux.c, config/winnt-c.c,
2060 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
2061 vtable-verify.c, vtable-verify.h: Use the standard form for the
2062 copyright notice.
2064 2014-01-02 Tobias Burnus <burnus@net-b.de>
2066 * gcc.c (process_command): Update copyright notice dates.
2067 * gcov-dump.c: Ditto.
2068 * gcov.c: Ditto.
2069 * doc/cpp.texi: Bump @copying's copyright year.
2070 * doc/cppinternals.texi: Ditto.
2071 * doc/gcc.texi: Ditto.
2072 * doc/gccint.texi: Ditto.
2073 * doc/gcov.texi: Ditto.
2074 * doc/install.texi: Ditto.
2075 * doc/invoke.texi: Ditto.
2077 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2079 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
2081 2014-01-01 Jakub Jelinek <jakub@redhat.com>
2083 * config/i386/sse.md (*mov<mode>_internal): Guard
2084 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
2086 PR rtl-optimization/59647
2087 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
2088 new_rtx into UNSIGNED_FLOAT rtxes.
2089 \f
2090 Copyright (C) 2014 Free Software Foundation, Inc.
2092 Copying and distribution of this file, with or without modification,
2093 are permitted in any medium without royalty provided the copyright
2094 notice and this notice are preserved.