]> git.ipfire.org Git - thirdparty/gcc.git/blob - gcc/ChangeLog
89930b0042c246fb044bdc0b45d72842fd7da515
[thirdparty/gcc.git] / gcc / ChangeLog
1 2014-04-28 Jeff Law <law@redhat.com>
2
3 PR tree-optimization/60902
4 * tree-ssa-threadedge.c
5 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
6 over real defs when invalidating outputs from statements that do not
7 produce useful outputs for threading.
8
9 2014-04-28 Richard Biener <rguenther@suse.de>
10
11 PR tree-optimization/60979
12 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
13 SCOPs that end in a block with a successor with abnormal
14 predecessors.
15
16 2014-04-28 Richard Biener <rguenther@suse.de>
17
18 * tree-pass.h (execute_pass_list): Adjust prototype.
19 * passes.c (pass_manager::execute_early_local_passes):
20 Adjust.
21 (do_per_function): Change callback signature, push all actual
22 work to the callbals.
23 (do_per_function_toporder): Likewise.
24 (execute_function_dump): Adjust.
25 (execute_function_todo): Likewise.
26 (clear_last_verified): Likewise.
27 (verify_curr_properties): Likewise.
28 (update_properties_after_pass): Likewise.
29 (execute_pass_list_1): Split out from ...
30 (execute_pass_list): ... here. Adjust.
31 (execute_ipa_pass_list): Likewise.
32 * cgraphunit.c (cgraph_add_new_function): Adjust.
33 (analyze_function): Likewise.
34 (expand_function): Likewise.
35 * cgraph.c (release_function_body): Free dominance info
36 here instead of asserting it was magically freed elsewhere.
37
38 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
39
40 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
41 * configure: Regenerate.
42 * config/sparc/sparc.opt (muser-mode): New option.
43 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
44 for LEON3.
45 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
46 * doc/invoke.texi (SPARC options): Document -muser-mode.
47
48 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
49
50 * cselib.c (find_slot_memmode): Delete.
51 (cselib_hasher): Change compare_type to a struct.
52 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
53 constants.
54 (preserve_constants_and_equivs): Adjust for new compare_type.
55 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
56 (wrap_constant): Delete.
57 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
58
59 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
60
61 * doc/install.texi (Building with profile feedback): Remove
62 outdated sentence.
63
64 2014-04-26 Tom de Vries <tom@codesourcery.com>
65
66 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
67 array accesses.
68
69 2014-04-25 Cary Coutant <ccoutant@google.com>
70
71 PR debug/60929
72 * dwarf2out.c (should_move_die_to_comdat): A type definition
73 can contain a subprogram definition, but don't move it to a
74 comdat unit.
75 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
76 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
77 from original DIE.
78 (clone_tree_hash): Rename to...
79 (clone_tree_partial): ...this; change callers. Copy
80 DW_TAG_subprogram DIEs as declarations.
81 (copy_decls_walk): Don't copy children of a declaration into a
82 type unit.
83
84 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
85
86 PR target/60969
87 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
88 alternative 12.
89
90 2014-04-25 Jiong Wang <jiong.wang@arm.com>
91
92 * config/arm/predicates.md (call_insn_operand): Add long_call check.
93 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
94 reg for long_call.
95 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
96 restriction.
97
98 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
99
100 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
101 T16-related fields.
102
103 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
104
105 PR tree-optimization/60930
106 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
107 creating a multiply candidate by folding two constant
108 multiplicands when the result overflows.
109
110 2014-04-25 Jakub Jelinek <jakub@redhat.com>
111
112 PR tree-optimization/60960
113 * tree-vect-generic.c (expand_vector_operation): Only call
114 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
115
116 2014-04-25 Tom de Vries <tom@codesourcery.com>
117
118 * expr.c (clobber_reg_mode): New function.
119 * expr.h (clobber_reg): New function.
120
121 2014-04-25 Tom de Vries <tom@codesourcery.com>
122
123 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
124 clobbers.
125
126 2014-04-25 Radovan Obradovic <robradovic@mips.com>
127 Tom de Vries <tom@codesourcery.com>
128
129 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
130 handle.
131 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
132 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
133 new argument to find_all_hard_reg_sets call.
134
135 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
136
137 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
138 Use HOST_WIDE_INT_C for mask literal.
139 (aarch_rev16_shleft_mask_imm_p): Likewise.
140
141 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
142
143 PR target/60941
144 * config/sparc/sparc.md (ashlsi3_extend): Delete.
145
146 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
147
148 PR preprocessor/56540
149 * config/i386/i386-c.c (ix86_target_macros): Define
150 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
151
152 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
153
154 * configure.ac (tga_func): Remove.
155 (LIB_TLS_SPEC): Remove.
156 * configure: Regenerate.
157 * config.in: Regenerate.
158 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
159
160 2014-04-25 Richard Biener <rguenther@suse.de>
161
162 PR ipa/60912
163 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
164 call stmt use/clobber sets during stmt walk instead of
165 walking the possibly incomplete set of caller edges.
166
167 2014-04-25 Richard Biener <rguenther@suse.de>
168
169 PR ipa/60911
170 * passes.c (apply_ipa_transforms): Inline into only caller ...
171 (execute_one_pass): ... here. Properly bring in function
172 bodies for nodes we want to apply IPA transforms to.
173
174 2014-04-24 Cong Hou <congh@google.com>
175
176 PR tree-optimization/60896
177 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
178 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
179 (vect_mark_pattern_stmts): Set the def type of all statements in
180 PATTERN_DEF_SEQ as vect_internal_def.
181
182 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
183
184 * doc/extend.texi (PowerPC Built-in Functions): Document new
185 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
186 functions.
187 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
188
189 * config/rs6000/predicates.md (const_0_to_3_operand): New
190 predicate to match 0..3 integer constants.
191
192 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
193 to support adding miscellaneous builtin functions.
194 (BU_DFP_MISC_2): Likewise.
195 (BU_P7_MISC_1): Likewise.
196 (BU_P7_MISC_2): Likewise.
197 (BU_P8V_MISC_3): Likewise.
198 (BU_MISC_1): Likewise.
199 (BU_MISC_2): Likewise.
200 (DIVWE): Add extended divide builtin functions.
201 (DIVWEO): Likewise.
202 (DIVWEU): Likewise.
203 (DIVWEUO): Likewise.
204 (DIVDE): Likewise.
205 (DIVDEO): Likewise.
206 (DIVDEU): Likewise.
207 (DIVDEUO): Likewise.
208 (DXEX): Add decimal floating-point builtin functions.
209 (DXEXQ): Likewise.
210 (DDEDPD): Likewise.
211 (DDEDPDQ): Likewise.
212 (DENBCD): Likewise.
213 (DENBCDQ): Likewise.
214 (DIEX): Likewise.
215 (DIEXQ): Likewise.
216 (DSCLI): Likewise.
217 (DSCLIQ): Likewise.
218 (DSCRI): Likewise.
219 (DSCRIQ): Likewise.
220 (CDTBCD): Add new BCD builtin functions.
221 (CBCDTD): Likewise.
222 (ADDG6S): Likewise.
223 (BCDADD): Likewise.
224 (BCDADD_LT): Likewise.
225 (BCDADD_EQ): Likewise.
226 (BCDADD_GT): Likewise.
227 (BCDADD_OV): Likewise.
228 (BCDSUB): Likewise.
229 (BCDSUB_LT): Likewise.
230 (BCDSUB_EQ): Likewise.
231 (BCDSUB_GT): Likewise.
232 (BCDSUB_OV): Likewise.
233 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
234 (UNPACK_TD): Likewise.
235 (PACK_TF): Likewise.
236 (UNPACK_TF): Likewise.
237 (UNPACK_TF_0): Likewise.
238 (UNPACK_TF_1): Likewise.
239 (PACK_V1TI): Likewise.
240 (UNPACK_V1TI): Likewise.
241
242 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
243 support for decimal floating point builtin functions.
244 (rs6000_expand_ternop_builtin): Add checks for the new builtin
245 functions that take constant arguments.
246 (rs6000_invalid_builtin): Add decimal floating point builtin
247 support.
248 (rs6000_init_builtins): Setup long double, _Decimal64, and
249 _Decimal128 types for new builtin functions.
250 (builtin_function_type): Set the unsigned flags appropriately for
251 the new builtin functions.
252 (rs6000_opt_masks): Add support for decimal floating point builtin
253 functions.
254
255 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
256 floating point builtin functions.
257 (RS6000_BTM_COMMON): Likewise.
258 (RS6000_BTI_long_double): Likewise.
259 (RS6000_BTI_dfloat64): Likewise.
260 (RS6000_BTI_dfloat128): Likewise.
261 (long_double_type_internal_node): Likewise.
262 (dfloat64_type_internal_node): Likewise.
263 (dfloat128_type_internal_node): Likewise.
264
265 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
266 2.07 bcd arithmetic instructions.
267 (UNSPEC_BCDSUB): Likewise.
268 (UNSPEC_BCD_OVERFLOW): Likewise.
269 (UNSPEC_BCD_ADD_SUB): Likewise.
270 (bcd_add_sub): Likewise.
271 (BCD_TEST): Likewise.
272 (bcd<bcd_add_sub>): Likewise.
273 (bcd<bcd_add_sub>_test): Likewise.
274 (bcd<bcd_add_sub>_test2): Likewise.
275 (bcd<bcd_add_sub>_<code>): Likewise.
276 (peephole2 for combined bcd ops): Likewise.
277
278 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
279 decimal floating point builtin functions.
280 (UNSPEC_DENBCD): Likewise.
281 (UNSPEC_DXEX): Likewise.
282 (UNSPEC_DIEX): Likewise.
283 (UNSPEC_DSCLI): Likewise.
284 (UNSPEC_DSCRI): Likewise.
285 (D64_D128): Likewise.
286 (dfp_suffix): Likewise.
287 (dfp_ddedpd_<mode>): Likewise.
288 (dfp_denbcd_<mode>): Likewise.
289 (dfp_dxex_<mode>): Likewise.
290 (dfp_diex_<mode>): Likewise.
291 (dfp_dscli_<mode>): Likewise.
292 (dfp_dscri_<mode>): Likewise.
293
294 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
295 builtin functions.
296 (UNSPEC_CDTBCD): Likewise.
297 (UNSPEC_CBCDTD): Likewise.
298 (UNSPEC_DIVE): Add support for new extended divide builtin
299 functions.
300 (UNSPEC_DIVEO): Likewise.
301 (UNSPEC_DIVEU): Likewise.
302 (UNSPEC_DIVEUO): Likewise.
303 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
304 pack/unpack 128-bit types.
305 (UNSPEC_PACK_128BIT): Likewise.
306 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
307 (udiv<mode>3): Use idiv_ldiv mode attribute.
308 (div<mode>3): Likewise.
309 (addg6s): Add new BCD builtin functions.
310 (cdtbcd): Likewise.
311 (cbcdtd): Likewise.
312 (UNSPEC_DIV_EXTEND): Add support for new extended divide
313 instructions.
314 (div_extend): Likewise.
315 (div<div_extend>_<mode>"): Likewise.
316 (FP128_64): Add support for new builtin functions to pack/unpack
317 128-bit types.
318 (unpack<mode>): Likewise.
319 (unpacktf_0): Likewise.
320 (unpacktf_1): Likewise.
321 (unpack<mode>_dm): Likewise.
322 (unpack<mode>_nodm): Likewise.
323 (pack<mode>): Likewise.
324 (unpackv1ti): Likewise.
325 (packv1ti): Likewise.
326
327 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
328
329 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
330 is disabled.
331
332 2014-04-24 Jakub Jelinek <jakub@redhat.com>
333
334 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
335 * gimplify.c (omp_is_private): Change last argument's type to int.
336 Only diagnose lastprivate if the simd argument is 1, only diagnose
337 linear if the simd argument is 2.
338 (gimplify_omp_for): Adjust omp_is_private callers. When adding
339 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
340 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
341 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
342 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
343 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
344 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
345 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
346 * tree-nested.c (convert_nonlocal_omp_clauses,
347 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
348
349 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
350
351 PR target/60822
352 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
353 operand 1.
354
355 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
356
357 * flag-types.h (enum ivar_visibility): Add.
358
359 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
360
361 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
362 function * argument.
363
364 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
365
366 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
367
368 2014-04-24 Radovan Obradovic <robradovic@mips.com>
369 Tom de Vries <tom@codesourcery.com>
370
371 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
372 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
373 reg-note.
374 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
375 * emit-rtl.c (try_split): Same.
376
377 2014-04-24 Radovan Obradovic <robradovic@mips.com>
378 Tom de Vries <tom@codesourcery.com>
379
380 * common.opt (fuse-caller-save): New option.
381
382 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
383
384 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
385 for big-endian.
386
387 2014-04-24 Richard Biener <rguenther@suse.de>
388
389 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
390 during TER and instead use the sepops interface for expanding
391 non-GIMPLE_SINGLE_RHS.
392
393 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
394
395 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
396 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
397
398 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
399
400 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
401 assembler 64-bit option.
402 * configure: Regenerate.
403
404 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
405
406 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
407 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
408 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
409 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
410 (TARGET_CRYPTO): Take TARGET_SIMD into account.
411
412 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
413
414 * config/aarch64/aarch64-builtins.c
415 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
416 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
417 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
418 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
419 builtins.
420 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
421 (Vrevsuff): New mode attribute.
422
423 2014-04-24 Terry Guo <terry.guo@arm.com>
424
425 * config/arm/arm.h (machine_function): Define variable
426 after_arm_reorg here.
427 * config/arm/arm.c (after_arm_reorg): Remove the definition.
428 (arm_split_constant): Update the way to access variable
429 after_arm_reorg.
430 (arm_reorg): Ditto.
431 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
432
433 2014-04-23 Tom de Vries <tom@codesourcery.com>
434
435 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
436
437 2014-04-23 David Malcolm <dmalcolm@redhat.com>
438
439 * is-a.h: Update comments to reflect the following changes to the
440 "pointerness" of the API, making the template parameter match the
441 return type, allowing use of is-a.h with typedefs of pointers.
442 (is_a_helper::cast): Return a T rather then a pointer to a T, so
443 that the return type matches the parameter to the is_a_helper.
444 (as_a): Likewise.
445 (dyn_cast): Likewise.
446
447 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
448 pointer from the is-a.h API.
449
450 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
451 (is_a_helper <cgraph_node *>::test): ...this, matching change to
452 is-a.h API.
453 (is_a_helper <varpool_node>::test): Likewise, convert to...
454 (is_a_helper <varpool_node *>::test): ...this.
455
456 (varpool_first_variable): Update for removal of implicit pointer
457 from the is-a.h API.
458 (varpool_next_variable): Likewise.
459 (varpool_first_static_initializer): Likewise.
460 (varpool_next_static_initializer): Likewise.
461 (varpool_first_defined_variable): Likewise.
462 (varpool_next_defined_variable): Likewise.
463 (cgraph_first_defined_function): Likewise.
464 (cgraph_next_defined_function): Likewise.
465 (cgraph_first_function): Likewise.
466 (cgraph_next_function): Likewise.
467 (cgraph_first_function_with_gimple_body): Likewise.
468 (cgraph_next_function_with_gimple_body): Likewise.
469 (cgraph_alias_target): Likewise.
470 (varpool_alias_target): Likewise.
471 (cgraph_function_or_thunk_node): Likewise.
472 (varpool_variable_node): Likewise.
473 (symtab_real_symbol_p): Likewise.
474 * cgraphunit.c (referred_to_p): Likewise.
475 (analyze_functions): Likewise.
476 (handle_alias_pairs): Likewise.
477 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
478 * gimple-ssa.h (gimple_vuse_op): Likewise.
479 (gimple_vdef_op): Likewise.
480 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
481 * gimple.c (gimple_build_asm_1): Likewise.
482 (gimple_build_try): Likewise.
483 (gimple_build_resx): Likewise.
484 (gimple_build_eh_dispatch): Likewise.
485 (gimple_build_omp_for): Likewise.
486 (gimple_omp_for_set_clauses): Likewise.
487
488 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
489 (is_a_helper <gimple_statement_asm *>::test): ...this.
490 (is_a_helper <gimple_statement_bind>::test): Convert to...
491 (is_a_helper <gimple_statement_bind *>::test): ...this.
492 (is_a_helper <gimple_statement_call>::test): Convert to...
493 (is_a_helper <gimple_statement_call *>::test): ...this.
494 (is_a_helper <gimple_statement_catch>::test): Convert to...
495 (is_a_helper <gimple_statement_catch *>::test): ...this.
496 (is_a_helper <gimple_statement_resx>::test): Convert to...
497 (is_a_helper <gimple_statement_resx *>::test): ...this.
498 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
499 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
500 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
501 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
502 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
503 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
504 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
505 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
506 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
507 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
508 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
509 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
510 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
511 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
512 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
513 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
514 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
515 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
516 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
517 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
518 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
519 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
520 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
521 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
522 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
523 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
524 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
525 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
526 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
527 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
528 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
529 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
530 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
531 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
532 (is_a_helper <gimple_statement_phi>::test): Convert to...
533 (is_a_helper <gimple_statement_phi *>::test): ...this.
534 (is_a_helper <gimple_statement_transaction>::test): Convert to...
535 (is_a_helper <gimple_statement_transaction *>::test): ...this.
536 (is_a_helper <gimple_statement_try>::test): Convert to...
537 (is_a_helper <gimple_statement_try *>::test): ...this.
538 (is_a_helper <gimple_statement_wce>::test): Convert to...
539 (is_a_helper <gimple_statement_wce *>::test): ...this.
540 (is_a_helper <const gimple_statement_asm>::test): Convert to...
541 (is_a_helper <const gimple_statement_asm *>::test): ...this.
542 (is_a_helper <const gimple_statement_bind>::test): Convert to...
543 (is_a_helper <const gimple_statement_bind *>::test): ...this.
544 (is_a_helper <const gimple_statement_call>::test): Convert to...
545 (is_a_helper <const gimple_statement_call *>::test): ...this.
546 (is_a_helper <const gimple_statement_catch>::test): Convert to...
547 (is_a_helper <const gimple_statement_catch *>::test): ...this.
548 (is_a_helper <const gimple_statement_resx>::test): Convert to...
549 (is_a_helper <const gimple_statement_resx *>::test): ...this.
550 (is_a_helper <const gimple_statement_eh_dispatch>::test):
551 Convert to...
552 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
553 ...this.
554 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
555 to...
556 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
557 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
558 Convert to...
559 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
560 ...this.
561 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
562 Convert to...
563 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
564 ...this.
565 (is_a_helper <const gimple_statement_omp_return>::test): Convert
566 to...
567 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
568 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
569 to...
570 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
571 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
572 to...
573 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
574 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
575 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
576 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
577 to...
578 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
579 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
580 to...
581 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
582 (is_a_helper <const gimple_statement_omp_target>::test): Convert
583 to...
584 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
585 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
586 to...
587 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
588 (is_a_helper <const gimple_statement_omp_single>::test): Convert
589 to...
590 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
591 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
592 to...
593 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
594 (is_a_helper <const gimple_statement_omp_task>::test): Convert
595 to...
596 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
597 (is_a_helper <const gimple_statement_phi>::test): Convert to...
598 (is_a_helper <const gimple_statement_phi *>::test): ...this.
599 (is_a_helper <const gimple_statement_transaction>::test): Convert
600 to...
601 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
602 (is_a_helper <const gimple_statement_with_ops>::test): Convert
603 to...
604 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
605 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
606 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
607 (is_a_helper <const gimple_statement_with_memory_ops>::test):
608 Convert to...
609 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
610 ...this.
611 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
612 to...
613 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
614
615 (gimple_use_ops): Update for removal of implicit pointer from the
616 is-a.h API.
617 (gimple_set_use_ops): Likewise.
618 (gimple_vuse): Likewise.
619 (gimple_vdef): Likewise.
620 (gimple_vuse_ptr): Likewise.
621 (gimple_vdef_ptr): Likewise.
622 (gimple_set_vuse): Likewise.
623 (gimple_set_vdef): Likewise.
624 (gimple_omp_return_set_lhs): Likewise.
625 (gimple_omp_return_lhs): Likewise.
626 (gimple_omp_return_lhs_ptr): Likewise.
627 (gimple_call_fntype): Likewise.
628 (gimple_call_set_fntype): Likewise.
629 (gimple_call_set_internal_fn): Likewise.
630 (gimple_call_use_set): Likewise.
631 (gimple_call_clobber_set): Likewise.
632 (gimple_bind_vars): Likewise.
633 (gimple_bind_set_vars): Likewise.
634 (gimple_bind_body_ptr): Likewise.
635 (gimple_bind_set_body): Likewise.
636 (gimple_bind_add_stmt): Likewise.
637 (gimple_bind_block): Likewise.
638 (gimple_bind_set_block): Likewise.
639 (gimple_asm_ninputs): Likewise.
640 (gimple_asm_noutputs): Likewise.
641 (gimple_asm_nclobbers): Likewise.
642 (gimple_asm_nlabels): Likewise.
643 (gimple_asm_input_op): Likewise.
644 (gimple_asm_input_op_ptr): Likewise.
645 (gimple_asm_output_op): Likewise.
646 (gimple_asm_output_op_ptr): Likewise.
647 (gimple_asm_set_output_op): Likewise.
648 (gimple_asm_clobber_op): Likewise.
649 (gimple_asm_set_clobber_op): Likewise.
650 (gimple_asm_label_op): Likewise.
651 (gimple_asm_set_label_op): Likewise.
652 (gimple_asm_string): Likewise.
653 (gimple_catch_types): Likewise.
654 (gimple_catch_types_ptr): Likewise.
655 (gimple_catch_handler_ptr): Likewise.
656 (gimple_catch_set_types): Likewise.
657 (gimple_catch_set_handler): Likewise.
658 (gimple_eh_filter_types): Likewise.
659 (gimple_eh_filter_types_ptr): Likewise.
660 (gimple_eh_filter_failure_ptr): Likewise.
661 (gimple_eh_filter_set_types): Likewise.
662 (gimple_eh_filter_set_failure): Likewise.
663 (gimple_eh_must_not_throw_fndecl): Likewise.
664 (gimple_eh_must_not_throw_set_fndecl): Likewise.
665 (gimple_eh_else_n_body_ptr): Likewise.
666 (gimple_eh_else_e_body_ptr): Likewise.
667 (gimple_eh_else_set_n_body): Likewise.
668 (gimple_eh_else_set_e_body): Likewise.
669 (gimple_try_eval_ptr): Likewise.
670 (gimple_try_cleanup_ptr): Likewise.
671 (gimple_try_set_eval): Likewise.
672 (gimple_try_set_cleanup): Likewise.
673 (gimple_wce_cleanup_ptr): Likewise.
674 (gimple_wce_set_cleanup): Likewise.
675 (gimple_phi_capacity): Likewise.
676 (gimple_phi_num_args): Likewise.
677 (gimple_phi_result): Likewise.
678 (gimple_phi_result_ptr): Likewise.
679 (gimple_phi_set_result): Likewise.
680 (gimple_phi_arg): Likewise.
681 (gimple_phi_set_arg): Likewise.
682 (gimple_resx_region): Likewise.
683 (gimple_resx_set_region): Likewise.
684 (gimple_eh_dispatch_region): Likewise.
685 (gimple_eh_dispatch_set_region): Likewise.
686 (gimple_omp_critical_name): Likewise.
687 (gimple_omp_critical_name_ptr): Likewise.
688 (gimple_omp_critical_set_name): Likewise.
689 (gimple_omp_for_clauses): Likewise.
690 (gimple_omp_for_clauses_ptr): Likewise.
691 (gimple_omp_for_set_clauses): Likewise.
692 (gimple_omp_for_collapse): Likewise.
693 (gimple_omp_for_index): Likewise.
694 (gimple_omp_for_index_ptr): Likewise.
695 (gimple_omp_for_set_index): Likewise.
696 (gimple_omp_for_initial): Likewise.
697 (gimple_omp_for_initial_ptr): Likewise.
698 (gimple_omp_for_set_initial): Likewise.
699 (gimple_omp_for_final): Likewise.
700 (gimple_omp_for_final_ptr): Likewise.
701 (gimple_omp_for_set_final): Likewise.
702 (gimple_omp_for_incr): Likewise.
703 (gimple_omp_for_incr_ptr): Likewise.
704 (gimple_omp_for_set_incr): Likewise.
705 (gimple_omp_for_pre_body_ptr): Likewise.
706 (gimple_omp_for_set_pre_body): Likewise.
707 (gimple_omp_parallel_clauses): Likewise.
708 (gimple_omp_parallel_clauses_ptr): Likewise.
709 (gimple_omp_parallel_set_clauses): Likewise.
710 (gimple_omp_parallel_child_fn): Likewise.
711 (gimple_omp_parallel_child_fn_ptr): Likewise.
712 (gimple_omp_parallel_set_child_fn): Likewise.
713 (gimple_omp_parallel_data_arg): Likewise.
714 (gimple_omp_parallel_data_arg_ptr): Likewise.
715 (gimple_omp_parallel_set_data_arg): Likewise.
716 (gimple_omp_task_clauses): Likewise.
717 (gimple_omp_task_clauses_ptr): Likewise.
718 (gimple_omp_task_set_clauses): Likewise.
719 (gimple_omp_task_child_fn): Likewise.
720 (gimple_omp_task_child_fn_ptr): Likewise.
721 (gimple_omp_task_set_child_fn): Likewise.
722 (gimple_omp_task_data_arg): Likewise.
723 (gimple_omp_task_data_arg_ptr): Likewise.
724 (gimple_omp_task_set_data_arg): Likewise.
725 (gimple_omp_taskreg_clauses): Likewise.
726 (gimple_omp_taskreg_clauses_ptr): Likewise.
727 (gimple_omp_taskreg_set_clauses): Likewise.
728 (gimple_omp_taskreg_child_fn): Likewise.
729 (gimple_omp_taskreg_child_fn_ptr): Likewise.
730 (gimple_omp_taskreg_set_child_fn): Likewise.
731 (gimple_omp_taskreg_data_arg): Likewise.
732 (gimple_omp_taskreg_data_arg_ptr): Likewise.
733 (gimple_omp_taskreg_set_data_arg): Likewise.
734 (gimple_omp_task_copy_fn): Likewise.
735 (gimple_omp_task_copy_fn_ptr): Likewise.
736 (gimple_omp_task_set_copy_fn): Likewise.
737 (gimple_omp_task_arg_size): Likewise.
738 (gimple_omp_task_arg_size_ptr): Likewise.
739 (gimple_omp_task_set_arg_size): Likewise.
740 (gimple_omp_task_arg_align): Likewise.
741 (gimple_omp_task_arg_align_ptr): Likewise.
742 (gimple_omp_task_set_arg_align): Likewise.
743 (gimple_omp_single_clauses): Likewise.
744 (gimple_omp_single_clauses_ptr): Likewise.
745 (gimple_omp_single_set_clauses): Likewise.
746 (gimple_omp_target_clauses): Likewise.
747 (gimple_omp_target_clauses_ptr): Likewise.
748 (gimple_omp_target_set_clauses): Likewise.
749 (gimple_omp_target_child_fn): Likewise.
750 (gimple_omp_target_child_fn_ptr): Likewise.
751 (gimple_omp_target_set_child_fn): Likewise.
752 (gimple_omp_target_data_arg): Likewise.
753 (gimple_omp_target_data_arg_ptr): Likewise.
754 (gimple_omp_target_set_data_arg): Likewise.
755 (gimple_omp_teams_clauses): Likewise.
756 (gimple_omp_teams_clauses_ptr): Likewise.
757 (gimple_omp_teams_set_clauses): Likewise.
758 (gimple_omp_sections_clauses): Likewise.
759 (gimple_omp_sections_clauses_ptr): Likewise.
760 (gimple_omp_sections_set_clauses): Likewise.
761 (gimple_omp_sections_control): Likewise.
762 (gimple_omp_sections_control_ptr): Likewise.
763 (gimple_omp_sections_set_control): Likewise.
764 (gimple_omp_for_set_cond): Likewise.
765 (gimple_omp_for_cond): Likewise.
766 (gimple_omp_atomic_store_set_val): Likewise.
767 (gimple_omp_atomic_store_val): Likewise.
768 (gimple_omp_atomic_store_val_ptr): Likewise.
769 (gimple_omp_atomic_load_set_lhs): Likewise.
770 (gimple_omp_atomic_load_lhs): Likewise.
771 (gimple_omp_atomic_load_lhs_ptr): Likewise.
772 (gimple_omp_atomic_load_set_rhs): Likewise.
773 (gimple_omp_atomic_load_rhs): Likewise.
774 (gimple_omp_atomic_load_rhs_ptr): Likewise.
775 (gimple_omp_continue_control_def): Likewise.
776 (gimple_omp_continue_control_def_ptr): Likewise.
777 (gimple_omp_continue_set_control_def): Likewise.
778 (gimple_omp_continue_control_use): Likewise.
779 (gimple_omp_continue_control_use_ptr): Likewise.
780 (gimple_omp_continue_set_control_use): Likewise.
781 (gimple_transaction_body_ptr): Likewise.
782 (gimple_transaction_label): Likewise.
783 (gimple_transaction_label_ptr): Likewise.
784 (gimple_transaction_set_body): Likewise.
785 (gimple_transaction_set_label): Likewise.
786
787 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
788 * ipa-inline-analysis.c (inline_write_summary): Likewise.
789 * ipa-ref.c (ipa_record_reference): Likewise.
790 * ipa-reference.c (analyze_function): Likewise.
791 (ipa_reference_write_optimization_summary): Likewise.
792 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
793 (address_taken_from_non_vtable_p): Likewise.
794 (comdat_can_be_unshared_p_1): Likewise.
795 * lto-cgraph.c (lto_output_ref): Likewise.
796 (add_references): Likewise.
797 (compute_ltrans_boundary): Likewise.
798 (output_symtab): Likewise.
799 (input_ref): Likewise.
800 (input_cgraph_1): Likewise.
801 (output_cgraph_opt_summary): Likewise.
802 * lto-streamer-out.c (lto_output): Likewise.
803 (output_symbol_p): Likewise.
804 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
805 (lsei_start_function_in_partition): Likewise.
806 (lsei_next_variable_in_partition): Likewise.
807 (lsei_start_variable_in_partition): Likewise.
808 * symtab.c (insert_to_assembler_name_hash): Likewise.
809 (unlink_from_assembler_name_hash): Likewise.
810 (symtab_unregister_node): Likewise.
811 (symtab_remove_node): Likewise.
812 (dump_symtab_node): Likewise.
813 (verify_symtab_base): Likewise.
814 (verify_symtab_node): Likewise.
815 (symtab_make_decl_local): Likewise.
816 (symtab_alias_ultimate_target): Likewise.
817 (symtab_resolve_alias): Likewise.
818 (symtab_get_symbol_partitioning_class): Likewise.
819 * tree-phinodes.c (allocate_phi_node): Likewise.
820 (reserve_phi_args_for_new_edge): Likewise.
821 (remove_phi_args): Likewise.
822 * varpool.c (varpool_node_for_asm): Likewise.
823 (varpool_remove_unreferenced_decls): Likewise.
824
825 2014-04-23 Jeff Law <law@redhat.com>
826
827 PR tree-optimization/60902
828 * tree-ssa-threadedge.c
829 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
830 invalidate outputs from statements that do not produce useful
831 outputs for threading.
832
833 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
834
835 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
836 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
837 machine descriptions for Stack Smashing Protector.
838
839 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
840
841 * aarch64.md (<optab>_rol<mode>3): New pattern.
842 (<optab>_rolsi3_uxtw): Likewise.
843 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
844
845 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
846
847 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
848 (arm_cortex_a12_tune): Likewise.
849
850 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
851
852 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
853
854 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
855
856 * config/arm/arm.md (arm_rev16si2): New pattern.
857 (arm_rev16si2_alt): Likewise.
858 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
859
860 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
861
862 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
863 (rev16<mode>2_alt): Likewise.
864 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
865 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
866 (aarch_rev16_shleft_mask_imm_p): Likewise.
867 (aarch_rev16_p_1): Likewise.
868 (aarch_rev16_p): Likewise.
869 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
870 (aarch_rev16_shright_mask_imm_p): Likewise.
871 (aarch_rev16_shleft_mask_imm_p): Likewise.
872
873 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
874
875 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
876 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
877 rev cost.
878 (cortex_a53_extra_costs): Likewise.
879 (cortex_a57_extra_costs): Likewise.
880 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
881 (cortexa7_extra_costs): Likewise.
882 (cortexa8_extra_costs): Likewise.
883 (cortexa12_extra_costs): Likewise.
884 (cortexa15_extra_costs): Likewise.
885 (v7m_extra_costs): Likewise.
886 (arm_new_rtx_costs): Handle BSWAP.
887
888 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
889
890 * config/arm/arm.c (cortexa8_extra_costs): New table.
891 (arm_cortex_a8_tune): New tuning struct.
892 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
893
894 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
895
896 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
897
898 2014-04-23 Richard Biener <rguenther@suse.de>
899
900 * Makefile.in (OBJS): Remove loop-unswitch.o.
901 * loop-unswitch.c: Delete.
902 * tree-pass.h (make_pass_rtl_unswitch): Remove.
903 * passes.def (pass_rtl_unswitch): Likewise.
904 * loop-init.c (gate_rtl_unswitch): Likewise.
905 (rtl_unswitch): Likewise.
906 (pass_data_rtl_unswitch): Likewise.
907 (pass_rtl_unswitch): Likewise.
908 (make_pass_rtl_unswitch): Likewise.
909 * rtl.h (reversed_condition): Likewise.
910 (compare_and_jump_seq): Likewise.
911 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
912 and make static.
913 * loop-unroll.c (compare_and_jump_seq): Likewise.
914
915 2014-04-23 Richard Biener <rguenther@suse.de>
916
917 PR tree-optimization/60903
918 * tree-ssa-loop-im.c (analyze_memory_references): Remove
919 commented code block.
920 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
921 loop flags to newly created BBs and edges.
922
923 2014-04-23 Nick Clifton <nickc@redhat.com>
924
925 * config/msp430/msp430.c (msp430_handle_option): Move function
926 to msp430-common.c
927 (msp430_option_override): Simplify mcu and mcpu option handling.
928 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
929 support for -mhwmult command line option.
930 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
931 -mhwmult command line option.
932 (msp430_hwmult_enabled): Delete.
933 (msp43o_output_labelref): Add support for -mhwmult command line
934 option.
935 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
936 (umulsidi3): Likewise.
937 * config/msp430/msp430.opt (mmcu): Add Report attribute.
938 (mcpu, mlarge, msmall): Likewise.
939 (mhwmult): New option.
940 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
941 prototype.
942 (msp430_is_f5_mcu): Remove prototype.
943 (msp430_use_f5_series_hwmult): Add prototype.
944 * config/msp430/msp430-opts.h: New file.
945 * common/config/msp430: New directory.
946 * common/config/msp430/msp430-common.c: New file.
947 * config.gcc (msp430): Remove target_has_targetm_common.
948 * doc/invoke.texi: Document -mhwmult command line option.
949
950 2014-04-23 Nick Clifton <nickc@redhat.com>
951
952 * config/i386/cygwin.h (ENDFILE_SPEC): Include
953 default-manifest.o if it can be found in the search path.
954 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
955
956 2014-04-23 Terry Guo <terry.guo@arm.com>
957
958 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
959
960 2014-04-23 Richard Biener <rguenther@suse.de>
961
962 PR middle-end/60895
963 * tree-inline.c (declare_return_variable): Use mark_addressable.
964
965 2014-04-23 Richard Biener <rguenther@suse.de>
966
967 PR middle-end/60891
968 * loop-init.c (loop_optimizer_init): Make sure to apply
969 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
970
971 2014-04-22 Jakub Jelinek <jakub@redhat.com>
972
973 PR sanitizer/60275
974 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
975 New options.
976 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
977 if flag_sanitize_undefined_trap_on_error.
978 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
979 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
980 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
981 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
982 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
983 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
984 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
985 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
986 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
987 * ubsan.c (ubsan_instrument_unreachable): Return
988 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
989 (ubsan_expand_null_ifn): Emit __builtin_trap ()
990 if flag_sanitize_undefined_trap_on_error and
991 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
992 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
993 instrument_bool_enum_load): Emit __builtin_trap () if
994 flag_sanitize_undefined_trap_on_error and
995 __builtin_handle_*_abort () if !flag_sanitize_recover.
996 * doc/invoke.texi (-fsanitize-recover,
997 -fsanitize-undefined-trap-on-error): Document.
998
999 2014-04-22 Christian Bruel <christian.bruel@st.com>
1000
1001 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1002 Force immediates to SImode.
1003
1004 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1005
1006 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1007 (lroundsfsi2): New.
1008 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1009 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1010 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1011 (nios2_fpu_insn): Add entry for round.
1012 (N2FPU_NO_ERRNO_P): Define.
1013 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1014 flag_errno_math.
1015 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1016
1017 2014-04-22 Richard Henderson <rth@redhat.com>
1018
1019 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1020 (add<GPI>3_compare0): Remove leading * from name.
1021 (add<GPI>3_carryin): Likewise.
1022 (sub<GPI>3_compare0): Likewise.
1023 (sub<GPI>3_carryin): Likewise.
1024 (<su_optab>mulditi3): New expander.
1025 (multi3): New expander.
1026 (madd<GPI>): Remove leading * from name.
1027
1028 2014-04-22 Martin Jambor <mjambor@suse.cz>
1029
1030 * cgraphclones.c (cgraph_function_versioning): Copy
1031 ipa_transforms_to_apply instead of asserting it is empty.
1032
1033 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1034
1035 PR target/60868
1036 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1037 on count_exp to get mode.
1038
1039 2014-04-22 Andrew Pinski <apinski@cavium.com>
1040
1041 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1042 Handle TLS for ILP32.
1043 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1044 (tlsie_small_<mode>): this and handle PTR.
1045 (tlsie_small_sidi): New pattern.
1046 (tlsle_small): Change to an expand to handle ILP32.
1047 (tlsle_small_<mode>): New pattern.
1048 (tlsdesc_small): Rename to ...
1049 (tlsdesc_small_<mode>): this and handle PTR.
1050
1051 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1052
1053 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1054
1055 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1056
1057 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1058 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1059 (aarch64_types_signed_poly_qualifiers): Likewise.
1060 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1061 (aarch64_types_poly_signed_qualifiers): Likewise.
1062 (TYPES_REINTERP_SS): Type macro added.
1063 (TYPES_REINTERP_SU): Likewise.
1064 (TYPES_REINTERP_SP): Likewise.
1065 (TYPES_REINTERP_US): Likewise.
1066 (TYPES_REINTERP_PS): Likewise.
1067 (aarch64_fold_builtin): New expression folding added.
1068 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1069 Declarations removed.
1070 (REINTERP_SS): Declarations added.
1071 (REINTERP_US): Likewise.
1072 (REINTERP_PS): Likewise.
1073 (REINTERP_SU): Likewise.
1074 (REINTERP_SP): Likewise.
1075 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1076 (vreinterpretq_p8_f64): Likewise.
1077 (vreinterpret_p16_f64): Likewise.
1078 (vreinterpretq_p16_f64): Likewise.
1079 (vreinterpret_f32_f64): Likewise.
1080 (vreinterpretq_f32_f64): Likewise.
1081 (vreinterpret_f64_f32): Likewise.
1082 (vreinterpret_f64_p8): Likewise.
1083 (vreinterpret_f64_p16): Likewise.
1084 (vreinterpret_f64_s8): Likewise.
1085 (vreinterpret_f64_s16): Likewise.
1086 (vreinterpret_f64_s32): Likewise.
1087 (vreinterpret_f64_s64): Likewise.
1088 (vreinterpret_f64_u8): Likewise.
1089 (vreinterpret_f64_u16): Likewise.
1090 (vreinterpret_f64_u32): Likewise.
1091 (vreinterpret_f64_u64): Likewise.
1092 (vreinterpretq_f64_f32): Likewise.
1093 (vreinterpretq_f64_p8): Likewise.
1094 (vreinterpretq_f64_p16): Likewise.
1095 (vreinterpretq_f64_s8): Likewise.
1096 (vreinterpretq_f64_s16): Likewise.
1097 (vreinterpretq_f64_s32): Likewise.
1098 (vreinterpretq_f64_s64): Likewise.
1099 (vreinterpretq_f64_u8): Likewise.
1100 (vreinterpretq_f64_u16): Likewise.
1101 (vreinterpretq_f64_u32): Likewise.
1102 (vreinterpretq_f64_u64): Likewise.
1103 (vreinterpret_s64_f64): Likewise.
1104 (vreinterpretq_s64_f64): Likewise.
1105 (vreinterpret_u64_f64): Likewise.
1106 (vreinterpretq_u64_f64): Likewise.
1107 (vreinterpret_s8_f64): Likewise.
1108 (vreinterpretq_s8_f64): Likewise.
1109 (vreinterpret_s16_f64): Likewise.
1110 (vreinterpretq_s16_f64): Likewise.
1111 (vreinterpret_s32_f64): Likewise.
1112 (vreinterpretq_s32_f64): Likewise.
1113 (vreinterpret_u8_f64): Likewise.
1114 (vreinterpretq_u8_f64): Likewise.
1115 (vreinterpret_u16_f64): Likewise.
1116 (vreinterpretq_u16_f64): Likewise.
1117 (vreinterpret_u32_f64): Likewise.
1118 (vreinterpretq_u32_f64): Likewise.
1119
1120 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1121
1122 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1123 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1124 (vreinterpret_p8_s8): Likewise.
1125 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1126 (vreinterpret_p8_s16): Likewise.
1127 (vreinterpret_p8_s32): Likewise.
1128 (vreinterpret_p8_s64): Likewise.
1129 (vreinterpret_p8_f32): Likewise.
1130 (vreinterpret_p8_u8): Likewise.
1131 (vreinterpret_p8_u16): Likewise.
1132 (vreinterpret_p8_u32): Likewise.
1133 (vreinterpret_p8_u64): Likewise.
1134 (vreinterpret_p8_p16): Likewise.
1135 (vreinterpretq_p8_s8): Likewise.
1136 (vreinterpretq_p8_s16): Likewise.
1137 (vreinterpretq_p8_s32): Likewise.
1138 (vreinterpretq_p8_s64): Likewise.
1139 (vreinterpretq_p8_f32): Likewise.
1140 (vreinterpretq_p8_u8): Likewise.
1141 (vreinterpretq_p8_u16): Likewise.
1142 (vreinterpretq_p8_u32): Likewise.
1143 (vreinterpretq_p8_u64): Likewise.
1144 (vreinterpretq_p8_p16): Likewise.
1145 (vreinterpret_p16_s8): Likewise.
1146 (vreinterpret_p16_s16): Likewise.
1147 (vreinterpret_p16_s32): Likewise.
1148 (vreinterpret_p16_s64): Likewise.
1149 (vreinterpret_p16_f32): Likewise.
1150 (vreinterpret_p16_u8): Likewise.
1151 (vreinterpret_p16_u16): Likewise.
1152 (vreinterpret_p16_u32): Likewise.
1153 (vreinterpret_p16_u64): Likewise.
1154 (vreinterpret_p16_p8): Likewise.
1155 (vreinterpretq_p16_s8): Likewise.
1156 (vreinterpretq_p16_s16): Likewise.
1157 (vreinterpretq_p16_s32): Likewise.
1158 (vreinterpretq_p16_s64): Likewise.
1159 (vreinterpretq_p16_f32): Likewise.
1160 (vreinterpretq_p16_u8): Likewise.
1161 (vreinterpretq_p16_u16): Likewise.
1162 (vreinterpretq_p16_u32): Likewise.
1163 (vreinterpretq_p16_u64): Likewise.
1164 (vreinterpretq_p16_p8): Likewise.
1165 (vreinterpret_f32_s8): Likewise.
1166 (vreinterpret_f32_s16): Likewise.
1167 (vreinterpret_f32_s32): Likewise.
1168 (vreinterpret_f32_s64): Likewise.
1169 (vreinterpret_f32_u8): Likewise.
1170 (vreinterpret_f32_u16): Likewise.
1171 (vreinterpret_f32_u32): Likewise.
1172 (vreinterpret_f32_u64): Likewise.
1173 (vreinterpret_f32_p8): Likewise.
1174 (vreinterpret_f32_p16): Likewise.
1175 (vreinterpretq_f32_s8): Likewise.
1176 (vreinterpretq_f32_s16): Likewise.
1177 (vreinterpretq_f32_s32): Likewise.
1178 (vreinterpretq_f32_s64): Likewise.
1179 (vreinterpretq_f32_u8): Likewise.
1180 (vreinterpretq_f32_u16): Likewise.
1181 (vreinterpretq_f32_u32): Likewise.
1182 (vreinterpretq_f32_u64): Likewise.
1183 (vreinterpretq_f32_p8): Likewise.
1184 (vreinterpretq_f32_p16): Likewise.
1185 (vreinterpret_s64_s8): Likewise.
1186 (vreinterpret_s64_s16): Likewise.
1187 (vreinterpret_s64_s32): Likewise.
1188 (vreinterpret_s64_f32): Likewise.
1189 (vreinterpret_s64_u8): Likewise.
1190 (vreinterpret_s64_u16): Likewise.
1191 (vreinterpret_s64_u32): Likewise.
1192 (vreinterpret_s64_u64): Likewise.
1193 (vreinterpret_s64_p8): Likewise.
1194 (vreinterpret_s64_p16): Likewise.
1195 (vreinterpretq_s64_s8): Likewise.
1196 (vreinterpretq_s64_s16): Likewise.
1197 (vreinterpretq_s64_s32): Likewise.
1198 (vreinterpretq_s64_f32): Likewise.
1199 (vreinterpretq_s64_u8): Likewise.
1200 (vreinterpretq_s64_u16): Likewise.
1201 (vreinterpretq_s64_u32): Likewise.
1202 (vreinterpretq_s64_u64): Likewise.
1203 (vreinterpretq_s64_p8): Likewise.
1204 (vreinterpretq_s64_p16): Likewise.
1205 (vreinterpret_u64_s8): Likewise.
1206 (vreinterpret_u64_s16): Likewise.
1207 (vreinterpret_u64_s32): Likewise.
1208 (vreinterpret_u64_s64): Likewise.
1209 (vreinterpret_u64_f32): Likewise.
1210 (vreinterpret_u64_u8): Likewise.
1211 (vreinterpret_u64_u16): Likewise.
1212 (vreinterpret_u64_u32): Likewise.
1213 (vreinterpret_u64_p8): Likewise.
1214 (vreinterpret_u64_p16): Likewise.
1215 (vreinterpretq_u64_s8): Likewise.
1216 (vreinterpretq_u64_s16): Likewise.
1217 (vreinterpretq_u64_s32): Likewise.
1218 (vreinterpretq_u64_s64): Likewise.
1219 (vreinterpretq_u64_f32): Likewise.
1220 (vreinterpretq_u64_u8): Likewise.
1221 (vreinterpretq_u64_u16): Likewise.
1222 (vreinterpretq_u64_u32): Likewise.
1223 (vreinterpretq_u64_p8): Likewise.
1224 (vreinterpretq_u64_p16): Likewise.
1225 (vreinterpret_s8_s16): Likewise.
1226 (vreinterpret_s8_s32): Likewise.
1227 (vreinterpret_s8_s64): Likewise.
1228 (vreinterpret_s8_f32): Likewise.
1229 (vreinterpret_s8_u8): Likewise.
1230 (vreinterpret_s8_u16): Likewise.
1231 (vreinterpret_s8_u32): Likewise.
1232 (vreinterpret_s8_u64): Likewise.
1233 (vreinterpret_s8_p8): Likewise.
1234 (vreinterpret_s8_p16): Likewise.
1235 (vreinterpretq_s8_s16): Likewise.
1236 (vreinterpretq_s8_s32): Likewise.
1237 (vreinterpretq_s8_s64): Likewise.
1238 (vreinterpretq_s8_f32): Likewise.
1239 (vreinterpretq_s8_u8): Likewise.
1240 (vreinterpretq_s8_u16): Likewise.
1241 (vreinterpretq_s8_u32): Likewise.
1242 (vreinterpretq_s8_u64): Likewise.
1243 (vreinterpretq_s8_p8): Likewise.
1244 (vreinterpretq_s8_p16): Likewise.
1245 (vreinterpret_s16_s8): Likewise.
1246 (vreinterpret_s16_s32): Likewise.
1247 (vreinterpret_s16_s64): Likewise.
1248 (vreinterpret_s16_f32): Likewise.
1249 (vreinterpret_s16_u8): Likewise.
1250 (vreinterpret_s16_u16): Likewise.
1251 (vreinterpret_s16_u32): Likewise.
1252 (vreinterpret_s16_u64): Likewise.
1253 (vreinterpret_s16_p8): Likewise.
1254 (vreinterpret_s16_p16): Likewise.
1255 (vreinterpretq_s16_s8): Likewise.
1256 (vreinterpretq_s16_s32): Likewise.
1257 (vreinterpretq_s16_s64): Likewise.
1258 (vreinterpretq_s16_f32): Likewise.
1259 (vreinterpretq_s16_u8): Likewise.
1260 (vreinterpretq_s16_u16): Likewise.
1261 (vreinterpretq_s16_u32): Likewise.
1262 (vreinterpretq_s16_u64): Likewise.
1263 (vreinterpretq_s16_p8): Likewise.
1264 (vreinterpretq_s16_p16): Likewise.
1265 (vreinterpret_s32_s8): Likewise.
1266 (vreinterpret_s32_s16): Likewise.
1267 (vreinterpret_s32_s64): Likewise.
1268 (vreinterpret_s32_f32): Likewise.
1269 (vreinterpret_s32_u8): Likewise.
1270 (vreinterpret_s32_u16): Likewise.
1271 (vreinterpret_s32_u32): Likewise.
1272 (vreinterpret_s32_u64): Likewise.
1273 (vreinterpret_s32_p8): Likewise.
1274 (vreinterpret_s32_p16): Likewise.
1275 (vreinterpretq_s32_s8): Likewise.
1276 (vreinterpretq_s32_s16): Likewise.
1277 (vreinterpretq_s32_s64): Likewise.
1278 (vreinterpretq_s32_f32): Likewise.
1279 (vreinterpretq_s32_u8): Likewise.
1280 (vreinterpretq_s32_u16): Likewise.
1281 (vreinterpretq_s32_u32): Likewise.
1282 (vreinterpretq_s32_u64): Likewise.
1283 (vreinterpretq_s32_p8): Likewise.
1284 (vreinterpretq_s32_p16): Likewise.
1285 (vreinterpret_u8_s8): Likewise.
1286 (vreinterpret_u8_s16): Likewise.
1287 (vreinterpret_u8_s32): Likewise.
1288 (vreinterpret_u8_s64): Likewise.
1289 (vreinterpret_u8_f32): Likewise.
1290 (vreinterpret_u8_u16): Likewise.
1291 (vreinterpret_u8_u32): Likewise.
1292 (vreinterpret_u8_u64): Likewise.
1293 (vreinterpret_u8_p8): Likewise.
1294 (vreinterpret_u8_p16): Likewise.
1295 (vreinterpretq_u8_s8): Likewise.
1296 (vreinterpretq_u8_s16): Likewise.
1297 (vreinterpretq_u8_s32): Likewise.
1298 (vreinterpretq_u8_s64): Likewise.
1299 (vreinterpretq_u8_f32): Likewise.
1300 (vreinterpretq_u8_u16): Likewise.
1301 (vreinterpretq_u8_u32): Likewise.
1302 (vreinterpretq_u8_u64): Likewise.
1303 (vreinterpretq_u8_p8): Likewise.
1304 (vreinterpretq_u8_p16): Likewise.
1305 (vreinterpret_u16_s8): Likewise.
1306 (vreinterpret_u16_s16): Likewise.
1307 (vreinterpret_u16_s32): Likewise.
1308 (vreinterpret_u16_s64): Likewise.
1309 (vreinterpret_u16_f32): Likewise.
1310 (vreinterpret_u16_u8): Likewise.
1311 (vreinterpret_u16_u32): Likewise.
1312 (vreinterpret_u16_u64): Likewise.
1313 (vreinterpret_u16_p8): Likewise.
1314 (vreinterpret_u16_p16): Likewise.
1315 (vreinterpretq_u16_s8): Likewise.
1316 (vreinterpretq_u16_s16): Likewise.
1317 (vreinterpretq_u16_s32): Likewise.
1318 (vreinterpretq_u16_s64): Likewise.
1319 (vreinterpretq_u16_f32): Likewise.
1320 (vreinterpretq_u16_u8): Likewise.
1321 (vreinterpretq_u16_u32): Likewise.
1322 (vreinterpretq_u16_u64): Likewise.
1323 (vreinterpretq_u16_p8): Likewise.
1324 (vreinterpretq_u16_p16): Likewise.
1325 (vreinterpret_u32_s8): Likewise.
1326 (vreinterpret_u32_s16): Likewise.
1327 (vreinterpret_u32_s32): Likewise.
1328 (vreinterpret_u32_s64): Likewise.
1329 (vreinterpret_u32_f32): Likewise.
1330 (vreinterpret_u32_u8): Likewise.
1331 (vreinterpret_u32_u16): Likewise.
1332 (vreinterpret_u32_u64): Likewise.
1333 (vreinterpret_u32_p8): Likewise.
1334 (vreinterpret_u32_p16): Likewise.
1335 (vreinterpretq_u32_s8): Likewise.
1336 (vreinterpretq_u32_s16): Likewise.
1337 (vreinterpretq_u32_s32): Likewise.
1338 (vreinterpretq_u32_s64): Likewise.
1339 (vreinterpretq_u32_f32): Likewise.
1340 (vreinterpretq_u32_u8): Likewise.
1341 (vreinterpretq_u32_u16): Likewise.
1342 (vreinterpretq_u32_u64): Likewise.
1343 (vreinterpretq_u32_p8): Likewise.
1344 (vreinterpretq_u32_p16): Likewise.
1345
1346 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1347
1348 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1349 Pattern extended.
1350 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1351 extended.
1352 (sqabs): Likewise.
1353 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1354 (vqnegd_s64): Likewise.
1355 (vqabs_s64): Likewise.
1356 (vqabsd_s64): Likewise.
1357
1358 2014-04-22 Richard Henderson <rth@redhat.com>
1359
1360 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1361 computation to the top of the loop.
1362
1363 2014-04-22 Renlin <renlin.li@arm.com>
1364 Jiong Wang <jiong.wang@arm.com>
1365
1366 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1367 * config/aarch64/aarch64.c (aarch64_layout_frame)
1368 (aarch64_initial_elimination_offset): Likewise.
1369
1370 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1371
1372 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1373 Fix indentation.
1374
1375 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1376
1377 * machmode.h (bitwise_mode_for_mode): Declare.
1378 * stor-layout.h (bitwise_type_for_mode): Likewise.
1379 * stor-layout.c (bitwise_mode_for_mode): New function.
1380 (bitwise_type_for_mode): Likewise.
1381 * builtins.c (fold_builtin_memory_op): Use it instead of
1382 int_mode_for_mode and build_nonstandard_integer_type.
1383
1384 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1385
1386 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1387 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1388 (*-*-solaris2*): Simplify.
1389 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1390 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1391 *-*-solaris2.9* handling.
1392
1393 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1394 as bug.
1395 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1396 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1397 handling, simplify.
1398 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1399 * configure: Regenerate.
1400
1401 * config/i386/sol2-9.h: Remove.
1402
1403 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1404 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1405 Remove Solaris 9 references.
1406
1407 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1408
1409 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1410 (floatuns<GPI:mode><GPF:mode>2): Remove.
1411 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1412 and floatuns conversions.
1413 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1414 and floatuns conversions.
1415 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1416 (w1,w2): New mode attributes for inequal width conversions.
1417
1418 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1419
1420 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1421 the output asm format.
1422
1423 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1424
1425 * config/aarch64/aarch64-simd.md
1426 (aarch64_cm<optab>di): Always split.
1427 (*aarch64_cm<optab>di): New.
1428 (aarch64_cmtstdi): Always split.
1429 (*aarch64_cmtstdi): New.
1430
1431 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1432
1433 PR tree-optimization/60823
1434 * omp-low.c (ipa_simd_modify_function_body): Go through
1435 all SSA_NAMEs and for those refering to vector arguments
1436 which are going to be replaced adjust SSA_NAME_VAR and,
1437 if it is a default definition, change it into a non-default
1438 definition assigned at the beginning of function from new_decl.
1439 (ipa_simd_modify_stmt_ops): Rewritten.
1440 * tree-dfa.c (set_ssa_default_def): When removing default def,
1441 check for NULL loc instead of NULL *loc.
1442
1443 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1444
1445 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1446 restrictions on core registers for DImode values in Thumb2.
1447
1448 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1449
1450 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1451 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1452
1453 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1454
1455 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1456 (*iordi_notzesidi_di): Likewise.
1457 (*iordi_notsesidi_di): Likewise.
1458
1459 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1460
1461 * config/arm/arm-protos.h (tune_params): New struct members.
1462 * config/arm/arm.c: Initialise tune_params per processor.
1463 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1464 for speed, based on new tune_params.
1465
1466 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1467
1468 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1469 added.
1470 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1471 macro.
1472 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1473 corrected.
1474 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1475 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1476 (vrnda_f64): Likewise.
1477 (vrndi_f64): Likewise.
1478 (vrndm_f64): Likewise.
1479 (vrndn_f64): Likewise.
1480 (vrndp_f64): Likewise.
1481 (vrndx_f64): Likewise.
1482
1483 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1484
1485 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1486 GET_MODE_SIZE argument is enum machine_mode.
1487
1488 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1489
1490 PR target/60910
1491 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1492 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1493
1494 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1495
1496 PR middle-end/60281
1497 * asan.c (asan_emit_stack_protection): Force the base to align to
1498 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1499 appropriate bits if STRICT_ALIGNMENT.
1500 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1501 when asan is on.
1502 (expand_used_vars): Leave a space in the stack frame for alignment
1503 if STRICT_ALIGNMENT.
1504
1505 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1506
1507 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1508 than a gimple.
1509 (gimple_store_p): Likewise.
1510 (gimple_assign_load_p): Likewise.
1511 (gimple_assign_cast_p): Likewise.
1512 (gimple_clobber_p): Likewise.
1513
1514 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1515 rather than a gimple.
1516 (gimple_assign_cast_p): Likewise.
1517
1518 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1519
1520 PR target/60735
1521 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1522 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1523
1524 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1525 more debug information for E500 if -mdebug=reg.
1526
1527 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1528
1529 PR target/60909
1530 * config/i386/i386.c (ix86_expand_builtin)
1531 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1532 register for target RTX.
1533 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1534
1535 2014-04-18 Cong Hou <congh@google.com>
1536
1537 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1538 the widen-mult pattern by handling two operands with different sizes,
1539 and operands whose size is smaller than half of the result type.
1540
1541 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1542
1543 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1544 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1545 (do_estimate_edge_time): Compute it.
1546 * ipa-inline.c (want_inline_small_function_p): Bypass
1547 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1548
1549 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1550
1551 * ipa-inline.c (spec_rem): New static variable.
1552 (dump_overall_stats): New function.
1553 (dump_inline_stats): New function.
1554
1555 2014-04-18 Richard Henderson <rth@redhat.com>
1556
1557 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1558 to GET_MODE_SIZE, not a reg_class_t.
1559
1560 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1561
1562 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1563 (vsx_xxmrglw_<mode>): Likewise.
1564
1565 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1566
1567 PR target/60876
1568 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1569 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1570 (rs6000_init_hard_regno_mode_ok): Likewise.
1571
1572 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1573
1574 * ipa-inline.c (inline_small_functions): Account only non-cold
1575 functions.
1576 * doc/invoke.texi (inline-unit-growth): Update documentation.
1577
1578 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1579
1580 * config/rs6000/rs6000.md (addti3, subti3): New.
1581
1582 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1583
1584 PR target/60863
1585 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1586 comment. Check optimize_insn_for_size_p instead of
1587 optimize_insn_for_speed_p.
1588
1589 2014-04-17 Martin Jambor <mjambor@suse.cz>
1590
1591 * gimple-iterator.c (gsi_start_edge): New function.
1592 * gimple-iterator.h (gsi_start_edge): Declare.
1593 * tree-sra.c (single_non_eh_succ): New function.
1594 (disqualify_ops_if_throwing_stmt): Renamed to
1595 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1596 having one non-EH successor BB.
1597 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1598 generate loads into replacements.
1599 (sra_modify_assign): Likewise and and also use the simple path for
1600 such statements.
1601 (sra_modify_function_body): Commit statements on edges.
1602
1603 2014-04-17 Richard Biener <rguenther@suse.de>
1604
1605 PR middle-end/60849
1606 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1607 comparison results and add clarifying comment.
1608
1609 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1610
1611 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1612 (blank_mode): Initialize it.
1613 (emit_mode_size_inline, emit_mode_nunits_inline,
1614 emit_mode_inner_inline): New functions.
1615 (emit_insn_modes_h): Call them and surround their output with
1616 #if GCC_VERSION >= 4001 ... #endif.
1617 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1618 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1619 mode_* arrays if the argument is __builtin_constant_p.
1620 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1621 is enum machine_mode.
1622
1623 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1624
1625 * passes.c (opt_pass::execute): Adjust.
1626 (pass_manager::execute_pass_mode_switching): Likewise.
1627 (early_local_passes::execute): Likewise.
1628 (execute_one_pass): Pass cfun to the pass's execute method.
1629 * tree-pass.h (opt_pass::execute): Add function * argument.
1630 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1631 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1632 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1633 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1634 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1635 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1636 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1637 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1638 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1639 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1640 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1641 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1642 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1643 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1644 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1645 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1646 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1647 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1648 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1649 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1650 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1651 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1652 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1653 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1654 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1655 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1656 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1657 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1658 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1659 Adjust.
1660
1661 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1662
1663 * passes.c (opt_pass::gate): Take function * argument.
1664 (gate_all_early_local_passes): Merge into
1665 (early_local_passes::gate): this.
1666 (gate_all_early_optimizations): Merge into
1667 (all_early_optimizations::gate): this.
1668 (gate_all_optimizations): Mege into
1669 (all_optimizations::gate): this.
1670 (gate_all_optimizations_g): Merge into
1671 (all_optimizations_g::gate): this.
1672 (gate_rest_of_compilation): Mege into
1673 (rest_of_compilation::gate): this.
1674 (gate_postreload): Merge into
1675 (postreload::gate): this.
1676 (dump_one_pass): Pass cfun to the pass's gate method.
1677 (execute_ipa_summary_passes): Likewise.
1678 (execute_one_pass): Likewise.
1679 (ipa_write_summaries_2): Likewise.
1680 (ipa_write_optimization_summaries_1): Likewise.
1681 (ipa_read_summaries_1): Likewise.
1682 (ipa_read_optimization_summaries_1): Likewise.
1683 (execute_ipa_stmt_fixups): Likewise.
1684 * tree-pass.h (opt_pass::gate): Add function * argument.
1685 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1686 combine-stack-adj.c, combine.c, compare-elim.c,
1687 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1688 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1689 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1690 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1691 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1692 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1693 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1694 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1695 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1696 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1697 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1698 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1699 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1700 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1701 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1702 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1703 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1704 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1705 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1706 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1707 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1708 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1709 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1710 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1711 var-tracking.c, vtable-verify.c, web.c: Adjust.
1712
1713 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1714
1715 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1716 * configure: Regenerate.
1717
1718 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1719
1720 * passes.c (dump_one_pass): don't check pass->has_gate.
1721 (execute_ipa_summary_passes): Likewise.
1722 (execute_one_pass): Likewise.
1723 (ipa_write_summaries_2): Likewise.
1724 (ipa_write_optimization_summaries_1): Likewise.
1725 (ipa_read_optimization_summaries_1): Likewise.
1726 (execute_ipa_stmt_fixups): Likewise.
1727 * tree-pass.h (pass_data::has_gate): Remove.
1728 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1729 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1730 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1731 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1732 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1733 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1734 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1735 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1736 gimple-low.c, gimple-ssa-isolate-paths.c,
1737 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1738 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1739 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1740 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1741 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1742 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1743 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1744 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1745 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1746 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1747 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1748 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1749 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1750 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1751 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1752 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1753 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1754 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1755 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1756 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1757 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1758 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1759 Adjust.
1760
1761 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1762
1763 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1764 declaration.
1765 * passes.c (pass_manager::register_dump_files_1): Merge into
1766 (pass_manager::register_dump_files): this, and remove its handling of
1767 properties since the pass always has the properties anyway.
1768 (pass_manager::pass_manager): Adjust.
1769
1770 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1771
1772 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1773 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1774 dealing with properties.
1775 (pass_manager::register_dump_files): Adjust.
1776
1777 2014-03-20 Mark Wielaard <mjw@redhat.com>
1778
1779 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1780 then represent the bound as normal constant value.
1781
1782 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1783
1784 PR target/60847
1785 Forward port from 4.8 branch
1786 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1787
1788 * config/i386/bmiintrin.h (_blsi_u32): New.
1789 (_blsi_u64): Ditto.
1790 (_blsr_u32): Ditto.
1791 (_blsr_u64): Ditto.
1792 (_blsmsk_u32): Ditto.
1793 (_blsmsk_u64): Ditto.
1794 (_tzcnt_u32): Ditto.
1795 (_tzcnt_u64): Ditto.
1796
1797 2014-04-17 Kito Cheng <kito@0xlab.org>
1798
1799 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1800
1801 2014-04-17 Richard Biener <rguenther@suse.de>
1802
1803 PR middle-end/60849
1804 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1805 boolean results for comparisons.
1806
1807 2014-04-17 Richard Biener <rguenther@suse.de>
1808
1809 PR tree-optimization/60836
1810 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1811 initial PHI args to be gimple values.
1812
1813 2014-04-17 Richard Biener <rguenther@suse.de>
1814
1815 PR tree-optimization/60841
1816 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1817 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1818 of stmts to SLP build.
1819 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1820 (vect_analyze_slp): Likewise.
1821 (vect_analyze_slp_instance): Likewise.
1822 (vect_build_slp_tree): Limit overall SLP tree growth.
1823 * tree-vectorizer.h (vect_analyze_data_refs,
1824 vect_analyze_slp): Adjust prototypes.
1825
1826 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1827
1828 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1829 Silvermont.
1830
1831 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1832
1833 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1834 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1835 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1836 for TARGET_SLOW_PSHUFB
1837
1838 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1839
1840 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1841 * config/i386/i386.c (intel_cost): Ditto.
1842
1843 2014-04-17 Joey Ye <joey.ye@arm.com>
1844
1845 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1846
1847 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1848
1849 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1850 with -fuse-profile.
1851
1852 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1853
1854 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1855 (type_all_derivations_known_p): New predicate.
1856 (type_all_ctors_visible_p): New predicate.
1857 (type_possibly_instantiated_p): New predicate.
1858 (get_odr_type): Compute all_derivations_known.
1859 (dump_odr_type): Dump the flag.
1860 (maybe_record_type): Cleanup.
1861 (record_target_from_binfo): Add bases_to_consider array;
1862 record bases for types w/o instances and skip CXX destructor.
1863 (possible_polymorphic_call_targets_1): Add bases_to_consider
1864 and consider_construction parameters; check if type may have instance.
1865 (get_polymorphic_call_info): Set maybe_in_construction to true
1866 when we know nothing.
1867 (record_targets_from_bases): Skip CXX destructors; they are
1868 never called for types in construction.
1869 (possible_polymorphic_call_targets): Do not record target when
1870 type may not have instance.
1871
1872 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1873
1874 PR ipa/60854
1875 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1876 external aliases alive, too.
1877
1878 2014-04-16 Andrew Pinski <apinski@cavium.com>
1879
1880 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1881 definition.
1882
1883 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1884
1885 * final.c (compute_alignments): Do not apply loop alignment to a block
1886 falling through to the exit.
1887
1888 2014-04-16 Catherine Moore <clm@codesourcery.com>
1889
1890 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1891 Adjust constraints for microMIPS store patterns.
1892
1893 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1894
1895 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1896
1897 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1898
1899 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1900 (append_use): Run at -O0.
1901 (append_vdef): Likewise.
1902 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1903 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1904
1905 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1906
1907 PR tree-optimization/60844
1908 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1909 (propagate_op_to_single_use, remove_visited_stmt_chain,
1910 linearize_expr, repropagate_negates, reassociate_bb): Use it
1911 instead of gsi_remove.
1912
1913 2014-04-16 Martin Jambor <mjambor@suse.cz>
1914
1915 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1916 ipa_transforms_to_apply.
1917 (cgraph_function_versioning): Assert that old_node has empty
1918 ipa_transforms_to_apply.
1919 * trans-mem.c (ipa_tm_create_version): Likewise.
1920 * tree-inline.c (tree_function_versioning): Do not duplicate
1921 ipa_transforms_to_apply.
1922
1923 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1924
1925 PR target/60817
1926 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
1927 x86_64-*-* cases.
1928 Pass necessary as flags on 64-bit Solaris/x86.
1929 Use lowercase relocs for x86_64-*-*.
1930 * configure: Regenerate.
1931
1932 2014-04-15 Jan Hubicka <jh@suse.cz>
1933
1934 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
1935 (maybe_record_node, likely_target_p): Use it.
1936
1937 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1938
1939 PR target/60839
1940 Revert following patch
1941
1942 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1943
1944 PR target/60735
1945 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
1946 software floating point or no floating point registers, do not
1947 allow any type in the FPRs. Eliminate a test for SPE SIMD types
1948 in GPRs that occurs after we tested for GPRs that would never be
1949 true.
1950
1951 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
1952 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
1953 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
1954 specifically allow DDmode, since that does not use the SPE SIMD
1955 instructions.
1956
1957 2014-03-21 Mark Wielaard <mjw@redhat.com>
1958
1959 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
1960 as unsigned or int depending on type and value used.
1961
1962 2014-04-15 Richard Biener <rguenther@suse.de>
1963
1964 PR rtl-optimization/56965
1965 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
1966 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
1967 ... here.
1968 * alias.c (true_dependence_1): Do not call
1969 nonoverlapping_component_refs_p.
1970 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
1971 nonoverlapping_component_refs_p.
1972 (indirect_refs_may_alias_p): Likewise.
1973
1974 2014-04-15 Teresa Johnson <tejohnson@google.com>
1975
1976 * cfg.c (dump_bb_info): Fix flags check.
1977 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
1978
1979 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1980
1981 PR rtl-optimization/60663
1982 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
1983 avoid 0 cost.
1984
1985 2014-04-15 Richard Biener <rguenther@suse.de>
1986
1987 * lto-streamer.h (LTO_major_version): Bump to 4.
1988
1989 2014-04-15 Richard Biener <rguenther@suse.de>
1990
1991 * common.opt (lto_partition_model): New enum.
1992 (flto-partition=): Merge separate options with a single with argument,
1993 add -flto-partition=one support.
1994 * flag-types.h (enum lto_partition_model): Declare.
1995 * opts.c (finish_options): Remove duplicate -flto-partition=
1996 option check.
1997 * lto-wrapper.c (run_gcc): Adjust.
1998
1999 2014-04-15 Richard Biener <rguenther@suse.de>
2000
2001 * alias.c (ncr_compar): New function.
2002 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2003
2004 2014-04-15 Richard Biener <rguenther@suse.de>
2005
2006 * alias.c (record_component_aliases): Do not walk BINFOs.
2007
2008 2014-04-15 Richard Biener <rguenther@suse.de>
2009
2010 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2011 Add struct function argument and adjust.
2012 (find_func_aliases_for_call): Likewise.
2013 (find_func_aliases): Likewise.
2014 (find_func_clobbers): Likewise.
2015 (intra_create_variable_infos): Likewise.
2016 (compute_points_to_sets): Likewise.
2017 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2018
2019 2014-04-15 Richard Biener <rguenther@suse.de>
2020
2021 * tree.c (iterative_hash_expr): Use enum tree_code_class
2022 to store TREE_CODE_CLASS.
2023 (tree_block): Likewise.
2024 (tree_set_block): Likewise.
2025 * tree.h (fold_build_pointer_plus_loc): Use
2026 convert_to_ptrofftype_loc.
2027
2028 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2029
2030 PR plugins/59335
2031 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2032 added in 4.9.
2033
2034 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2035
2036 * cfgloop.h (struct loop): Move force_vectorize down.
2037 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2038 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2039 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2040 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2041 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2042 * tree-core.h (enum annot_expr_kind): Add new kind values.
2043 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2044 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2045 kinds.
2046 * tree.def (ANNOTATE_EXPR): Tweak comment.
2047
2048 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2049
2050 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2051 cxa_pure_virtual).
2052
2053 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2054
2055 * tree.h (TYPE_IDENTIFIER): Declare.
2056 * tree.c (subrange_type_for_debug_p): Use it.
2057 * godump.c (go_format_type): Likewise.
2058 * dwarf2out.c (is_cxx_auto, modified_type_die,
2059 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2060 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2061
2062 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2063
2064 PR lto/60820
2065 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2066
2067 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2068
2069 * config/i386/i386.c (examine_argument): Return bool. Return true if
2070 parameter should be passed in memory.
2071 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2072 (construct_container): Update calls to examine_argument.
2073 (function_arg_advance_64): Ditto.
2074 (return_in_memory_32): Merge with ix86_return_in_memory.
2075 (return_in_memory_64): Ditto.
2076 (return_in_memory_ms_64): Ditto.
2077
2078 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2079
2080 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2081 * coverage.c (coverage_compute_profile_id): Handle externally visible
2082 symbols.
2083
2084 2014-04-14 Martin Jambor <mjambor@suse.cz>
2085
2086 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2087 DECL_DISREGARD_INLINE_LIMITS functions.
2088
2089 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2090
2091 PR target/60827
2092 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2093
2094 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2095
2096 PR target/60827
2097 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2098 optimize_insn_for_speed_p instead of
2099 optimize_function_for_speed_p.
2100
2101 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2102
2103 * doc/invoke.texi (free): Document AArch64.
2104
2105 2014-04-14 Richard Biener <rguenther@suse.de>
2106
2107 PR tree-optimization/60042
2108 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2109 (insert_into_preds_of_block): Do not prevent PHI insertion
2110 for REFERENCE exprs here ...
2111 (eliminate_dom_walker::before_dom_children): ... but prevent
2112 their use here under similar conditions when applied to the
2113 IL after PRE optimizations.
2114
2115 2014-04-14 Richard Biener <rguenther@suse.de>
2116
2117 * passes.def: Move early points-to after early SRA.
2118
2119 2014-04-14 Richard Biener <rguenther@suse.de>
2120
2121 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2122 check for which sign-changes we allow when forwarding
2123 a converted value into a switch.
2124
2125 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2126
2127 * stor-layout.c (place_field): Finalize non-constant offset for the
2128 field, if any.
2129
2130 2014-04-14 Richard Biener <rguenther@suse.de>
2131
2132 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2133 as argument.
2134 (expand_switch_using_bit_tests_p): Likewise.
2135 (process_switch): Compute and pass on speed_p based on the
2136 switch stmt.
2137 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2138 optimize_bb_for_speed_p.
2139
2140 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2141
2142 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2143 * function.h (struct function): Rename has_force_vect_loops into
2144 has_force_vectorize_loops.
2145 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2146 (input_struct_function_base): Likewise.
2147 * lto-streamer-out.c (output_cfg): Likewise.
2148 (output_struct_function_base): Likewise.
2149 * omp-low.c (expand_omp_simd): Likewise.
2150 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2151 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2152 (version_loop_for_if_conversion): Likewise.
2153 (tree_if_conversion): Likewise.
2154 (main_tree_if_conversion): Likewise.
2155 (gate_tree_if_conversion): Likewise.
2156 * tree-inline.c (copy_loops): Likewise.
2157 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2158 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2159 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2160 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2161 * tree-vectorizer.c (vectorize_loops): Likewise.
2162 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2163
2164 2014-04-14 Richard Biener <rguenther@suse.de>
2165
2166 PR lto/60720
2167 * lto-streamer-out.c (wrap_refs): New function.
2168 (lto_output): Wrap symbol references in global initializes in
2169 type-preserving MEM_REFs.
2170
2171 2014-04-14 Christian Bruel <christian.bruel@st.com>
2172
2173 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2174
2175 2014-04-14 Christian Bruel <christian.bruel@st.com>
2176
2177 * config/sh/sh.md (setmemqi): New expand pattern.
2178 * config/sh/sh.h (CLEAR_RATIO): Define.
2179 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2180 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2181
2182 2014-04-14 Richard Biener <rguenther@suse.de>
2183
2184 PR middle-end/55022
2185 * fold-const.c (negate_expr_p): Don't negate directional rounding
2186 division.
2187 (fold_negate_expr): Likewise.
2188
2189 2014-04-14 Richard Biener <rguenther@suse.de>
2190
2191 PR tree-optimization/59817
2192 PR tree-optimization/60453
2193 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2194 recursion to catch all CHRECs in the scalar evolution and restrict
2195 the predicate for the remains appropriately.
2196
2197 2014-04-12 Catherine Moore <clm@codesourcery.com>
2198
2199 * config/mips/constraints.md: Add new register constraint "kb".
2200 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2201 (*movhi_internal): Likewise.
2202 (*movqi_internal): Likewise.
2203 * config/mips/mips.h (M16_STORE_REGS): New register class.
2204 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2205 (REG_CLASS_CONTENTS): Likewise.
2206 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2207
2208 2014-04-11 Tobias Burnus <burnus@net-b.de>
2209
2210 PR c/60194
2211 * doc/invoke.texi (-Wformat-signedness): Document it.
2212 (Wformat=2): Mention that this enables -Wformat-signedness.
2213
2214 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2215
2216 * common/config/epiphany/epiphany-common.c
2217 (epiphany_option_optimization_table): Enable section anchors by
2218 default at -O1 or higher.
2219 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2220 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2221 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2222 carries no extra cost.
2223 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2224 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2225 * config/epiphany/predicates.md (memclob_operand): New predicate.
2226 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2227 Use memclob_operand predicate and X constraint for operand 3.
2228
2229 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2230
2231 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2232 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2233 its operands.
2234
2235 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2236
2237 PR rtl-optimization/60651
2238 * mode-switching.c (optimize_mode_switching): Make sure to emit
2239 sets of a lower numbered entity before sets of a higher numbered
2240 entity to a mode of the same or lower priority.
2241 When creating a seginfo for a basic block that starts with a code
2242 label, move the insertion point past the code label.
2243 (new_seginfo): Document and enforce requirement that
2244 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2245 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2246 * doc/tm.texi: Regenerate.
2247
2248 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2249
2250 PR target/60811
2251 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2252
2253 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2254
2255 * BASE-VER: Set to 4.10.0.
2256
2257 2014-04-11 Tobias Burnus <burnus@net-b.de>
2258
2259 PR other/59055
2260 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2261 * doc/gcc.texi (Service): Update description in the @menu
2262 * doc/invoke.texi (Option Summary): Remove misplaced and
2263 duplicated @menu.
2264
2265 2014-04-11 Steve Ellcey <sellcey@mips.com>
2266 Jakub Jelinek <jakub@redhat.com>
2267
2268 PR middle-end/60556
2269 * expr.c (convert_move): Use emit_store_flag_force instead of
2270 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2271 argument to it.
2272
2273 2014-04-11 Richard Biener <rguenther@suse.de>
2274
2275 PR middle-end/60797
2276 * varasm.c (assemble_alias): Avoid endless error reporting
2277 recursion by setting TREE_ASM_WRITTEN.
2278
2279 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2280
2281 * config/s390/s390.md: Add a splitter for NOT rtx.
2282
2283 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2284
2285 PR rtl-optimization/60663
2286 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2287
2288 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2289 Jakub Jelinek <jakub@redhat.com>
2290
2291 PR lto/60567
2292 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2293 flag from decl_node to node.
2294
2295 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2296
2297 PR debug/60655
2298 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2299 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2300 ameliorating the cases where it can be.
2301
2302 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2303
2304 Revert
2305 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2306
2307 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2308 (loadsync_<mode>): Change mode.
2309 (load_quadpti, store_quadpti): New.
2310 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2311 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2312 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2313
2314 2014-04-09 Cong Hou <congh@google.com>
2315
2316 PR testsuite/60773
2317 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2318 documentation.
2319
2320 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2321
2322 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2323 instead of vnor to exploit possible fusion opportunity in the
2324 future.
2325 (altivec_expand_vec_perm_const_le): Likewise.
2326
2327 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2328
2329 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2330 (loadsync_<mode>): Change mode.
2331 (load_quadpti, store_quadpti): New.
2332 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2333 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2334
2335 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2336
2337 PR target/60763
2338 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2339 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2340 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2341
2342 2014-04-08 Richard Biener <rguenther@suse.de>
2343
2344 PR middle-end/60706
2345 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2346 a 64bit widest int print double-int similar to on HWI64 hosts.
2347
2348 2014-04-08 Richard Biener <rguenther@suse.de>
2349
2350 PR tree-optimization/60785
2351 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2352 default defs properly.
2353
2354 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2355
2356 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2357 (Weffc++): Likewise.
2358
2359 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2360
2361 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2362 set completep to false rather than true.
2363
2364 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2365
2366 PR target/60504
2367 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2368 ARM_TARGET2_DWARF_FORMAT.
2369
2370 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2371
2372 PR target/60609
2373 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2374 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2375 ADDR_DIFF_VEC.
2376
2377 2014-04-07 Richard Biener <rguenther@suse.de>
2378
2379 PR tree-optimization/60766
2380 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2381 (may_eliminate_iv): Convert cand_value_at result to desired type.
2382
2383 2014-04-07 Jason Merrill <jason@redhat.com>
2384
2385 PR c++/60731
2386 * common.opt (-fno-gnu-unique): Add.
2387 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2388
2389 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2390
2391 * haifa-sched.c: Fix outdated function reference and minor
2392 grammar errors in introductory comment.
2393
2394 2014-04-07 Richard Biener <rguenther@suse.de>
2395
2396 PR middle-end/60750
2397 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2398 for noreturn calls.
2399 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2400
2401 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2402
2403 PR debug/55794
2404 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2405 size accounting for thunks.
2406 (pa_asm_output_mi_thunk): Use final_start_function() and
2407 final_end_function() to output function start and end directives.
2408
2409 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2410
2411 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2412 device specific ISA/ feature information. Remove short_sp and
2413 errata_skip ds. Add avr_device_specific_features enum to have device
2414 specific info.
2415 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2416 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2417 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2418 updated device specific info.
2419 * config/avr/avr-mcus.def: Merge device specific details to
2420 dev_attribute field.
2421 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2422 errata_skip.
2423 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2424 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2425 assembler if RMW isa supported by current device.
2426 * config/avr/genmultilib.awk: Update as device info structure changed.
2427 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2428
2429 2014-04-04 Cong Hou <congh@google.com>
2430
2431 PR tree-optimization/60656
2432 * tree-vect-stmts.c (supportable_widening_operation):
2433 Fix a bug that elements in a vector with vect_used_by_reduction
2434 property are incorrectly reordered when the operation on it is not
2435 consistant with the one in reduction operation.
2436
2437 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2438
2439 PR rtl-optimization/60155
2440 * gcse.c (record_set_data): New function.
2441 (single_set_gcse): New function.
2442 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2443 (hoist_code): Likewise.
2444 (get_pressure_class_and_nregs): Likewise.
2445
2446 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2447
2448 * explow.c (probe_stack_range): Emit a final optimization blockage.
2449
2450 2014-04-04 Anthony Green <green@moxielogic.com>
2451
2452 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2453 typos.
2454
2455 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2456
2457 PR ipa/59626
2458 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2459 flags are set only during streaming.
2460 * ipa.c (process_references, walk_polymorphic_call_targets,
2461 symtab_remove_unreachable_nodes): Drop bodies of always inline
2462 after early inlining.
2463 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2464
2465 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2466 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2467
2468 PR debug/60655
2469 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2470 containing a NOT.
2471
2472 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2473
2474 PR bootstrap/60743
2475 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2476 duration.
2477 (cortex_a53_fdivd): Likewise.
2478
2479 2014-04-04 Martin Jambor <mjambor@suse.cz>
2480
2481 PR ipa/60640
2482 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2483 Adjust all callers.
2484 * cgraph.c (clone_of_p): Also return true if thunks match.
2485 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2486 cgraph_function_or_thunk_node and an obsolete comment.
2487 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2488 file.
2489 (build_function_decl_skip_args): Likewise.
2490 (set_new_clone_decl_and_node_flags): New function.
2491 (duplicate_thunk_for_node): Likewise.
2492 (redirect_edge_duplicating_thunks): Likewise.
2493 (cgraph_clone_node): New parameter args_to_skip, pass it to
2494 redirect_edge_duplicating_thunks which is called instead of
2495 cgraph_redirect_edge_callee.
2496 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2497 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2498
2499 2014-04-04 Jeff Law <law@redhat.com>
2500
2501 PR target/60657
2502 * config/arm/predicates.md (const_int_I_operand): New predicate.
2503 (const_int_M_operand): Similarly.
2504 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2505 const_int_operand.
2506 (insv_t2, extv_reg, extzv_t2): Likewise.
2507 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2508 (pop_multiple_with_writeback_and_return): Likewise.
2509 (vfp_pop_multiple_with_writeback): Likewise
2510
2511 2014-04-04 Richard Biener <rguenther@suse.de>
2512
2513 PR ipa/60746
2514 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2515 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2516 non-GIMPLE_LABELs.
2517 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2518 * gimplify.c (gimple_add_tmp_var_fn): New function.
2519 * gimple-expr.h (create_tmp_reg_fn): Declare.
2520 * gimple-expr.c (create_tmp_reg_fn): New function.
2521 * gimple-low.c (record_vars_into): Don't change cfun.
2522 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2523 code generation without cfun.
2524
2525 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2526
2527 PR bootstrap/60719
2528 * Makefile.in (install-driver): Fix shell scripting.
2529
2530 2014-04-03 Cong Hou <congh@google.com>
2531
2532 PR tree-optimization/60505
2533 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2534 threshold of number of iterations below which no vectorization
2535 will be done.
2536 * tree-vect-loop.c (new_loop_vec_info):
2537 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2538 * tree-vect-loop.c (vect_analyze_loop_operations):
2539 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2540 * tree-vect-loop.c (vect_transform_loop):
2541 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2542 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2543 of iterations of the loop and see if we should build the epilogue.
2544
2545 2014-04-03 Richard Biener <rguenther@suse.de>
2546
2547 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2548 (streamer_tree_cache_create): Adjust.
2549 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2550 to allow optional nodes array.
2551 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2552 (streamer_tree_cache_append): Likewise.
2553 (streamer_tree_cache_create): Create nodes array optionally
2554 as specified by parameter.
2555 * lto-streamer-out.c (create_output_block): Avoid maintaining
2556 the node array in the writer cache.
2557 (DFS_write_tree): Remove assertion.
2558 (produce_asm_for_decls): Free the out decl state hash table early.
2559 * lto-streamer-in.c (lto_data_in_create): Adjust for
2560 streamer_tree_cache_create prototype change.
2561
2562 2014-04-03 Richard Biener <rguenther@suse.de>
2563
2564 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2565 set TREE_CHAIN to NULL_TREE.
2566
2567 2014-04-03 Richard Biener <rguenther@suse.de>
2568
2569 PR tree-optimization/60740
2570 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2571 over all GIMPLE_COND operands.
2572
2573 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2574
2575 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2576 (Weffc++): Remove Scott's numbering, merge lists and reference
2577 Wnon-virtual-dtor.
2578
2579 2014-04-03 Nick Clifton <nickc@redhat.com>
2580
2581 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2582 properly.
2583
2584 2014-04-03 Martin Jambor <mjambor@suse.cz>
2585
2586 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2587 mention gcc_unreachable before failing.
2588 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2589 removed symbols.
2590
2591 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2592
2593 PR ipa/60659
2594 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2595 inconsistent code and instead mark the context inconsistent.
2596 (possible_polymorphic_call_targets): For inconsistent contexts
2597 return empty complete list.
2598
2599 2014-04-02 Anthony Green <green@moxielogic.com>
2600
2601 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2602 (extendqisi2, extendhisi2): Define.
2603 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2604 (WCHAR_TYPE): Change to unsigned int.
2605
2606 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2607
2608 PR tree-optimization/60733
2609 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2610 insertion point for PHI candidates to be the end of the feeding
2611 block for the PHI argument.
2612
2613 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2614
2615 PR rtl-optimization/60650
2616 * lra-constraints.c (process_alt_operands): Decrease reject for
2617 earlyclobber matching.
2618
2619 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2620
2621 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2622
2623 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2624
2625 * config/spu/spu.c (pad_bb): Do not crash when the last
2626 insn is CODE_FOR_blockage.
2627
2628 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2629
2630 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2631 lies outside the target mode.
2632
2633 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2634
2635 PR target/60735
2636 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2637 software floating point or no floating point registers, do not
2638 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2639 in GPRs that occurs after we tested for GPRs that would never be
2640 true.
2641
2642 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2643 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2644 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2645 specifically allow DDmode, since that does not use the SPE SIMD
2646 instructions.
2647
2648 2014-04-02 Richard Biener <rguenther@suse.de>
2649
2650 PR middle-end/60729
2651 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2652 MODE_INTs. Properly use negv_optab.
2653 (expand_abs): Likewise.
2654
2655 2014-04-02 Richard Biener <rguenther@suse.de>
2656
2657 PR bootstrap/60719
2658 * Makefile.in (install-driver): Guard extra installs with special
2659 names properly.
2660
2661 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2662
2663 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2664 Document vec_vgbbd.
2665
2666 2014-04-01 Richard Henderson <rth@redhat.com>
2667
2668 PR target/60704
2669 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2670 alternative enabled before register allocation.
2671
2672 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2673
2674 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2675 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2676 typo.
2677 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2678 (nios2_got_address): Update nios2_large_got_address call site.
2679 (nios2_delegitimize_address): New function.
2680 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2681 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2682 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2683
2684 2014-04-01 Martin Husemann <martin@duskware.de>
2685
2686 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2687 for -mabi=32.
2688
2689 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2690
2691 PR rtl-optimization/60604
2692 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2693 check from register_operand.
2694 (register_operand): Redefine in terms of general_operand.
2695 (nonmemory_operand): Use register_operand for the non-constant cases.
2696
2697 2014-04-01 Richard Biener <rguenther@suse.de>
2698
2699 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2700
2701 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2702
2703 * doc/invoke.texi (mapp-regs): Clarify.
2704
2705 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2706
2707 * config/i386/avx512fintrin.h (__v32hi): Define type.
2708 (__v64qi): Likewise.
2709 (_mm512_set1_epi8): Define.
2710 (_mm512_set1_epi16): Define.
2711 (_mm512_set4_epi32): Define.
2712 (_mm512_set4_epi64): Define.
2713 (_mm512_set4_pd): Define.
2714 (_mm512_set4_ps): Define.
2715 (_mm512_setr4_epi64): Define.
2716 (_mm512_setr4_epi32): Define.
2717 (_mm512_setr4_pd): Define.
2718 (_mm512_setr4_ps): Define.
2719 (_mm512_setzero_epi32): Define.
2720
2721 2014-03-31 Martin Jambor <mjambor@suse.cz>
2722
2723 PR middle-end/60647
2724 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2725 callsite_arguments_match_p. Updated all callers. Also check types of
2726 corresponding formal parameters and actual arguments.
2727 (not_all_callers_have_enough_arguments_p) Renamed to
2728 some_callers_have_mismatched_arguments_p.
2729
2730 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2731
2732 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2733
2734 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2735
2736 PR target/60034
2737 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2738 section anchor.
2739
2740 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2741
2742 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2743 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2744 Split out
2745 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2746 Use FMAMODE_NOVF512 mode iterator.
2747 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2748 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2749 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2750 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2751 Split out
2752 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2753 Use VF_128_256 mode iterator.
2754 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2755 Ditto.
2756
2757 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2758
2759 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2760 static chain if needed.
2761
2762 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2763
2764 PR target/60697
2765 * lra-constraints.c (index_part_to_reg): New.
2766 (process_address): Use it.
2767
2768 2014-03-27 Jeff Law <law@redhat.com>
2769 Jakub Jelinek <jakub@redhat.com>
2770
2771 PR target/60648
2772 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2773 gen_rtx_{PLUS,MULT} to build up the address expression.
2774
2775 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2776 creating non-canonical RTL.
2777
2778 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2779
2780 PR ipa/60243
2781 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2782 functions; reorganize to make cheap checks first.
2783 (inline_small_functions): Do not estimate growth when dumping;
2784 it is expensive.
2785 * ipa-inline.h (inline_summary): Add min_size.
2786 (growth_likely_positive): New function.
2787 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2788 (set_cond_stmt_execution_predicate): Cleanup.
2789 (estimate_edge_size_and_time): Compute min_size.
2790 (estimate_calls_size_and_time): Likewise.
2791 (estimate_node_size_and_time): Likewise.
2792 (inline_update_overall_summary): Update min_size.
2793 (do_estimate_edge_time): Likewise.
2794 (do_estimate_edge_size): Update.
2795 (do_estimate_edge_hints): Update.
2796 (growth_likely_positive): New function.
2797
2798 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2799
2800 PR target/60693
2801 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2802 also if addr has VOIDmode.
2803
2804 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2805
2806 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2807 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2808 Declare extern.
2809 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2810 instructions as well as AdvancedSIMD loads.
2811
2812 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2813
2814 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2815 Use crypto_aese type.
2816 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2817 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2818 crypto_aese, crypto_aesmc. Move to types.md.
2819 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2820 crypto_aesmc.
2821 * config/arm/iterators.md (crypto_type): Likewise.
2822
2823 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2824
2825 * cgraph.c: Include expr.h and tree-dfa.h.
2826 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2827 remove LHS.
2828
2829 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2830
2831 PR target/60675
2832 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2833 regs from checking multi-reg pseudos.
2834
2835 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2836
2837 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2838
2839 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2840
2841 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2842 if it would clobber the stack pointer, even temporarily.
2843
2844 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2845
2846 * mode-switching.c: Make small adjustments to the top comment.
2847
2848 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2849
2850 * config/rs6000/constraints.md (wD constraint): New constraint to
2851 match the constant integer to get the top DImode/DFmode out of a
2852 vector in a VSX register.
2853
2854 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2855 match the constant integer to get the top DImode/DFmode out of a
2856 vector in a VSX register.
2857
2858 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2859 for ISA 2.07.
2860
2861 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2862 vbpermq builtins.
2863
2864 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2865 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2866
2867 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2868 Optimize vec_extract of 64-bit values, where the value being
2869 extracted is in the top word, where we can use scalar
2870 instructions. Add direct move and store support. Combine the big
2871 endian/little endian vector select load support into a single insn.
2872 (vsx_extract_<mode>_internal1): Likewise.
2873 (vsx_extract_<mode>_internal2): Likewise.
2874 (vsx_extract_<mode>_load): Likewise.
2875 (vsx_extract_<mode>_store): Likewise.
2876 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2877 combined into vsx_extract_<mode>_load.
2878 (vsx_extract_<mode>_one_le): Likewise.
2879
2880 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2881 define the top 64-bit vector element.
2882
2883 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2884 constraint.
2885
2886 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2887 Document vec_vbpermq builtin.
2888
2889 PR target/60672
2890 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2891 enable use of xxsldwi and xxpermdi builtin functions.
2892 (vec_xxpermdi): Likewise.
2893
2894 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2895 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2896
2897 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2898
2899 PR rtl-optimization/60650
2900 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2901 first_p. Use it.
2902 (find_spills_for): New.
2903 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2904 Spill all pseudos on the second iteration.
2905
2906 2014-03-27 Marek Polacek <polacek@redhat.com>
2907
2908 PR c/50347
2909 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2910 types.
2911
2912 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2913
2914 * config/s390/s390.c (s390_can_use_return_insn): Check for
2915 call-saved FPRs on 31 bit.
2916
2917 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2918
2919 PR middle-end/60682
2920 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2921 if they need regimplification, just drop them instead of
2922 calling gimple_regimplify_operands on them.
2923
2924 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2925
2926 PR target/60580
2927 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
2928 (aarch64_frame_pointer_required): Adjust logic.
2929 (aarch64_can_eliminate): Adjust logic.
2930 (aarch64_override_options_after_change): Adjust logic.
2931
2932 2014-03-27 Dehao Chen <dehao@google.com>
2933
2934 * ipa-inline.c (early_inliner): Update node's inline info.
2935
2936 2014-03-26 Dehao Chen <dehao@google.com>
2937
2938 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
2939 compiler inserted conditional jumps for NAN float check.
2940
2941 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2942
2943 * ubsan.h (ubsan_create_data): Change second argument's type
2944 to const location_t *.
2945 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
2946 _("<unknown>").
2947 (ubsan_create_data): Change second argument to const location_t *PLOC.
2948 Create Loc field whenever PLOC is non-NULL.
2949 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
2950 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
2951 callers.
2952
2953 PR other/59545
2954 * real.c (real_to_integer2): Change type of low to UHWI.
2955
2956 2014-03-26 Tobias Burnus <burnus@net-b.de>
2957
2958 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
2959 (CILK_SELF_SPECS): New define.
2960 (driver_self_specs): Use it.
2961
2962 2014-03-26 Richard Biener <rguenther@suse.de>
2963
2964 * tree-pretty-print.c (percent_K_format): Implement special
2965 case for LTO and its stripped down BLOCK tree.
2966
2967 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2968
2969 PR sanitizer/60636
2970 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
2971
2972 * tree-vrp.c (simplify_internal_call_using_ranges): If only
2973 one range is range_int_cst_p, but not both, at least optimize
2974 addition/subtraction of 0 and multiplication by 0 or 1.
2975 * gimple-fold.c (gimple_fold_call): Fold
2976 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
2977 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
2978 INTEGER_CSTs, try to fold at least x * 0 and y - y.
2979
2980 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
2981
2982 PR rtl-optimization/60452
2983 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
2984 <case REG>: Return 1 for invalid offsets from the frame pointer.
2985
2986 2014-03-26 Marek Polacek <polacek@redhat.com>
2987
2988 PR c/37428
2989 * doc/extend.texi (C Extensions): Mention variable-length arrays in
2990 a structure/union.
2991
2992 2014-03-26 Marek Polacek <polacek@redhat.com>
2993
2994 PR c/39525
2995 * doc/extend.texi (Designated Inits): Describe what happens to omitted
2996 field members.
2997
2998 2014-03-26 Marek Polacek <polacek@redhat.com>
2999
3000 PR other/59545
3001 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3002 multiplication in unsigned type.
3003
3004 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3005
3006 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3007
3008 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3009
3010 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3011
3012 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3013
3014 PR ipa/60315
3015 * cif-code.def (UNREACHABLE) New code.
3016 * ipa-inline.c (inline_small_functions): Skip edges to
3017 __builtlin_unreachable.
3018 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3019 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3020 predicate to __bulitin_unreachable.
3021 (set_cond_stmt_execution_predicate): Fix issue when
3022 invert_tree_comparison returns ERROR_MARK.
3023 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3024 propagate to inline clones.
3025 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3026 to unreachable.
3027 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3028 * cgraphclones.c (cgraph_clone_node): If call destination is already
3029 ureachable, do not redirect it back.
3030 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3031 unreachable.
3032
3033 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3034
3035 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3036 Do not modify inline clones.
3037
3038 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3039
3040 * config/i386/i386.md (general_sext_operand): New mode attr.
3041 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3042 don't generate (sign_extend (const_int)).
3043 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3044 operands[2]. Use We constraint instead of <i> and
3045 <general_sext_operand> predicate instead of <general_operand>.
3046 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3047 * config/i386/constraints.md (We): New constraint.
3048 * config/i386/predicates.md (x86_64_sext_operand,
3049 sext_operand): New predicates.
3050
3051 2014-03-25 Martin Jambor <mjambor@suse.cz>
3052
3053 PR ipa/60600
3054 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3055 inconsistent devirtualizations to __builtin_unreachable.
3056
3057 2014-03-25 Marek Polacek <polacek@redhat.com>
3058
3059 PR c/35449
3060 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3061
3062 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3063
3064 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3065 order of elements for big-endian.
3066
3067 2014-03-25 Richard Biener <rguenther@suse.de>
3068
3069 PR middle-end/60635
3070 * gimplify-me.c (gimple_regimplify_operands): Update the
3071 re-gimplifed stmt.
3072
3073 2014-03-25 Martin Jambor <mjambor@suse.cz>
3074
3075 PR ipa/59176
3076 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3077 (lto_output_varpool_node): Likewise.
3078 (input_overwrite_node): Likewise.
3079 (input_varpool_node): Likewise.
3080
3081 2014-03-25 Richard Biener <rguenther@suse.de>
3082
3083 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3084 (run_gcc): Likewise.
3085
3086 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3087
3088 * combine.c (simplify_compare_const): Add MODE argument.
3089 Handle mode_width 0 as very large mode_width.
3090 (try_combine, simplify_comparison): Adjust callers.
3091
3092 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3093 type to avoid signed integer overflow.
3094 * explow.c (plus_constant): Likewise.
3095
3096 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3097
3098 * doc/generic.texi: Correct typos.
3099
3100 2014-03-24 Tobias Burnus <burnus@net-b.de>
3101
3102 * doc/invoke.texi (-flto): Expand section about
3103 using static libraries with LTO.
3104
3105 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3106
3107 PR rtl-optimization/60501
3108 * optabs.def (addptr3_optab): New optab.
3109 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3110 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3111 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3112
3113 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3114
3115 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3116
3117 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3118
3119 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3120 _mm512_set1_pd.
3121
3122 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3123 (_mm256_undefined_ps): Define.
3124 (_mm256_undefined_pd): Define.
3125 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3126 (_mm_undefined_pd): Define.
3127 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3128 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3129 (_mm512_undefined_ps): Define.
3130 (_mm512_undefined_pd): Define.
3131 Use _mm*_undefined_*.
3132 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3133
3134 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3135
3136 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3137 (lshr_simd): DI mode added.
3138 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3139 (aarch64_ushr_simddi): Likewise.
3140 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3141 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3142 (vshrd_n_u64): Likewise.
3143
3144 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3145
3146 * Makefile.in (s-macro_list): Depend on cc1.
3147
3148 2014-03-23 Teresa Johnson <tejohnson@google.com>
3149
3150 * ipa-utils.c (ipa_print_order): Use specified dump file.
3151
3152 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3153
3154 PR rtl-optimization/60601
3155 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3156
3157 * gcc.c (eval_spec_function): Initialize save_growing_value.
3158
3159 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3160
3161 PR sanitizer/60613
3162 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3163 code == MINUS_EXPR, never swap op0 with op1.
3164
3165 * toplev.c (init_local_tick): Avoid signed integer multiplication
3166 overflow.
3167 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3168 shift by first operand's bitsize.
3169
3170 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3171
3172 PR target/60610
3173 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3174 redefine to 1 or 0.
3175 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3176 TARGET_ISA_64BIT_P(x).
3177
3178 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3179
3180 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3181 pattern for vector nor instead of subtract from splat(-1).
3182 (altivec_expand_vec_perm_const_le): Likewise.
3183
3184 2014-03-21 Richard Henderson <rth@twiddle.net>
3185
3186 PR target/60598
3187 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3188 related insns after epilogue_completed.
3189
3190 2014-03-21 Martin Jambor <mjambor@suse.cz>
3191
3192 PR ipa/59176
3193 * cgraph.h (symtab_node): New flag body_removed.
3194 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3195 when removing bodies.
3196 * symtab.c (dump_symtab_base): Dump body_removed flag.
3197 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3198 had their bodies removed.
3199
3200 2014-03-21 Martin Jambor <mjambor@suse.cz>
3201
3202 PR ipa/60419
3203 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3204 in the border.
3205
3206 2014-03-21 Richard Biener <rguenther@suse.de>
3207
3208 PR tree-optimization/60577
3209 * tree-core.h (struct tree_base): Document nothrow_flag use
3210 in VAR_DECL_NONALIASED.
3211 * tree.h (VAR_DECL_NONALIASED): New.
3212 (may_be_aliased): Adjust.
3213 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3214
3215 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3216
3217 * expr.c (expand_expr_real_1): Remove outdated comment.
3218
3219 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3220
3221 PR middle-end/60597
3222 * ira.c (adjust_cleared_regs): Call copy_rtx on
3223 *reg_equiv[REGNO (loc)].src_p before passing it to
3224 simplify_replace_fn_rtx.
3225
3226 PR target/60568
3227 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3228 into CONST, put pic register as first operand of PLUS. Use
3229 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3230
3231 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3232
3233 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3234
3235 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3236
3237 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3238 around for store forwarding issue in the FPU on the UT699.
3239 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3240 loads and operations if -mfix-ut699 is specified.
3241 (divtf3_hq): Tweak attribute.
3242 (sqrttf2_hq): Likewise.
3243
3244 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3245
3246 * calls.c (store_one_arg): Remove incorrect const qualification on the
3247 type of the temporary.
3248 * cfgexpand.c (expand_return): Likewise.
3249 * expr.c (expand_constructor): Likewise.
3250 (expand_expr_real_1): Likewise.
3251
3252 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3253
3254 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3255 of parts.
3256
3257 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3258
3259 PR target/60039
3260 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3261
3262 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3263
3264 * config/arm/aarch-common-protos.h
3265 (alu_cost_table): Fix spelling of "extend".
3266 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3267
3268 2014-03-19 Richard Biener <rguenther@suse.de>
3269
3270 PR middle-end/60553
3271 * tree-core.h (tree_type_common): Re-order pointer members
3272 to reduce recursion depth during GC walks.
3273
3274 2014-03-19 Marek Polacek <polacek@redhat.com>
3275
3276 PR sanitizer/60569
3277 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3278 before accessing it.
3279
3280 2014-03-19 Richard Biener <rguenther@suse.de>
3281
3282 PR lto/59543
3283 * lto-streamer-in.c (input_function): In WPA stage do not drop
3284 debug stmts.
3285
3286 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3287
3288 PR tree-optimization/60559
3289 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3290 with build_zero_cst assignment.
3291
3292 2014-03-18 Kai Tietz <ktietz@redhat.com>
3293
3294 PR rtl-optimization/56356
3295 * sdbout.c (sdbout_parms): Verify that parms'
3296 incoming argument is valid.
3297 (sdbout_reg_parms): Likewise.
3298
3299 2014-03-18 Richard Henderson <rth@redhat.com>
3300
3301 PR target/60562
3302 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3303 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3304 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3305
3306 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3307
3308 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3309 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3310 Italicize plugin event names in description. Explain that
3311 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3312 Remind that no GCC functions should be called after PLUGIN_FINISH.
3313 Explain what pragmas with expansion are.
3314
3315 2014-03-18 Martin Liska <mliska@suse.cz>
3316
3317 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3318 gimple call statement is update.
3319 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3320 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3321
3322 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3323
3324 PR sanitizer/60557
3325 * ubsan.c (ubsan_instrument_unreachable): Call
3326 initialize_sanitizer_builtins.
3327 (ubsan_pass): Likewise.
3328
3329 PR sanitizer/60535
3330 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3331 varpool_finalize_decl instead of rest_of_decl_compilation.
3332
3333 2014-03-18 Richard Biener <rguenther@suse.de>
3334
3335 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3336 by using bitmap_and_compl instead of bitmap_and_compl_into.
3337 (df_rd_transfer_function): Likewise.
3338
3339 2014-03-18 Richard Biener <rguenther@suse.de>
3340
3341 * doc/lto.texi (fresolution): Fix typo.
3342
3343 2014-03-18 Richard Biener <rguenther@suse.de>
3344
3345 * doc/invoke.texi (flto): Update for changes in 4.9.
3346
3347 2014-03-18 Richard Biener <rguenther@suse.de>
3348
3349 * doc/loop.texi: Remove section on the removed lambda framework.
3350 Update loop docs with recent changes in preserving loop structure.
3351
3352 2014-03-18 Richard Biener <rguenther@suse.de>
3353
3354 * doc/lto.texi (-fresolution): Document.
3355
3356 2014-03-18 Richard Biener <rguenther@suse.de>
3357
3358 * doc/contrib.texi: Adjust my name.
3359
3360 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3361
3362 PR ipa/58721
3363 * internal-fn.c: Include diagnostic-core.h.
3364 (expand_BUILTIN_EXPECT): New function.
3365 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3366 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3367 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3368 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3369 IFN_BUILTIN_EXPECT.
3370 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3371 Revert 3 argument __builtin_expect code.
3372 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3373 * gimple-fold.c (gimple_fold_call): Likewise.
3374 * tree.h (fold_builtin_expect): New prototype.
3375 * builtins.c (build_builtin_expect_predicate): Add predictor
3376 argument, if non-NULL, create 3 argument __builtin_expect.
3377 (fold_builtin_expect): No longer static. Add ARG2 argument,
3378 pass it through to build_builtin_expect_predicate.
3379 (fold_builtin_2): Adjust caller.
3380 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3381 * internal-fn.def (BUILTIN_EXPECT): New.
3382
3383 2014-03-18 Tobias Burnus <burnus@net-b.de>
3384
3385 PR ipa/58721
3386 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3387 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3388 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3389
3390 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3391
3392 PR ipa/58721
3393 * predict.c (combine_predictions_for_bb): Fix up formatting.
3394 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3395 fill what it points to if non-NULL.
3396 (tree_predict_by_opcode): Adjust caller, use the predictor.
3397 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3398
3399 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3400
3401 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3402 proper constant for the store mode.
3403
3404 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3405
3406 * symtab.c (change_decl_assembler_name): Fix transparent alias
3407 chain construction.
3408
3409 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3410
3411 * config/aarch64/aarch64.c: Correct the comments about the
3412 aarch64 stack layout.
3413
3414 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3415
3416 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3417 check for GF_OMP_FOR_KIND_FOR.
3418
3419 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3420
3421 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3422 ymm and zmm register names.
3423
3424 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3425
3426 PR target/60516
3427 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3428 note creation for the 2010-08-31 changes.
3429
3430 2014-03-17 Marek Polacek <polacek@redhat.com>
3431
3432 PR middle-end/60534
3433 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3434 as -fno-tree-loop-vectorize.
3435 (expand_omp_simd): Likewise.
3436
3437 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3438
3439 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3440 (eligible_for_call_delay): New prototype.
3441 * config/sparc/sparc.c (tls_call_delay): Rename into...
3442 (eligible_for_call_delay): ...this. Return false if the instruction
3443 cannot be put in the delay slot of a branch.
3444 (eligible_for_restore_insn): Simplify.
3445 (eligible_for_return_delay): Return false if the instruction cannot be
3446 put in the delay slot of a branch and simplify.
3447 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3448 put in the delay slot of a branch.
3449 * config/sparc/sparc.md (fix_ut699): New attribute.
3450 (tls_call_delay): Delete.
3451 (in_call_delay): Reimplement.
3452 (eligible_for_sibcall_delay): Rename into...
3453 (in_sibcall_delay): ...this.
3454 (eligible_for_return_delay): Rename into...
3455 (in_return_delay): ...this.
3456 (in_branch_delay): Reimplement.
3457 (in_uncond_branch_delay): Delete.
3458 (in_annul_branch_delay): Delete.
3459
3460 2014-03-14 Richard Henderson <rth@redhat.com>
3461
3462 PR target/60525
3463 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3464 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3465 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3466 (floathi splitters): Remove.
3467 (float<SWI48x>xf2): New pattern.
3468 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3469 code that tried to handle DImode for 32-bit, but which was excluded
3470 by the pattern's condition. Drop allocation of stack temporary.
3471 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3472 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3473 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3474 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3475 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3476 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3477 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3478 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3479 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3480 (*float<SWI48x><X87MODEF>2_i387): Remove.
3481 (all float _with_temp splitters): Remove.
3482 (*float<SWI48x><MODEF>2_i387): New pattern.
3483 (*float<SWI48><MODEF>2_sse): New pattern.
3484 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3485 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3486
3487 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3488 Marek Polacek <polacek@redhat.com>
3489
3490 PR middle-end/60484
3491 * common.opt (dump_base_name_prefixed): New Variable.
3492 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3493 if x_dump_base_name_prefixed is already set, set it at the end.
3494
3495 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3496
3497 PR rtl-optimization/60508
3498 * lra-constraints.c (get_reload_reg): Add new parameter
3499 in_subreg_p.
3500 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3501 Pass the new parameter values.
3502
3503 2014-03-14 Richard Biener <rguenther@suse.de>
3504
3505 * common.opt: Revert unintented changes from r205065.
3506 * opts.c: Likewise.
3507
3508 2014-03-14 Richard Biener <rguenther@suse.de>
3509
3510 PR middle-end/60518
3511 * cfghooks.c (split_block): Properly adjust all loops the
3512 block was a latch of.
3513
3514 2014-03-14 Martin Jambor <mjambor@suse.cz>
3515
3516 PR lto/60461
3517 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3518 and simplify it.
3519
3520 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3521
3522 PR target/59396
3523 * config/avr/avr.c (avr_set_current_function): Pass function name
3524 through default_strip_name_encoding before sanity checking instead
3525 of skipping the first char of the assembler name.
3526
3527 2014-03-13 Richard Henderson <rth@redhat.com>
3528
3529 PR debug/60438
3530 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3531 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3532 * config/i386/i386-protos.h: Likewise.
3533 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3534 in the expander instead of a splitter.
3535 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3536 any possibility of requiring a memory.
3537 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3538 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3539 (fp branch splitters): Update for ix86_split_fp_branch.
3540 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3541 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3542 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3543 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3544 (*fop_<MODEF>_3_i387): Likewise.
3545 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3546 (splitters for the fop_* register patterns): Remove.
3547 (fscalexf4_i387): Rename from *fscalexf4_i387.
3548 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3549
3550 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3551
3552 PR tree-optimization/59779
3553 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3554 type for bitsize and maxsize instead of HOST_WIDE_INT.
3555
3556 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3557
3558 PR rtl-optimization/57320
3559 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3560 the CFG after thread_prologue_and_epilogue_insns.
3561
3562 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3563
3564 PR rtl-optimization/57189
3565 * lra-constraints.c (process_alt_operands): Disfavor spilling
3566 vector pseudos.
3567
3568 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3569
3570 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3571
3572 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3573
3574 PR tree-optimization/59025
3575 PR middle-end/60418
3576 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3577 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3578
3579 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3580
3581 PR target/60486
3582 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3583 calls of avr_out_plus_1.
3584
3585 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3586
3587 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3588 BB's single pred and update the father loop's latch info later.
3589
3590 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3591
3592 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3593 (VEC_M): Likewise.
3594 (VEC_N): Likewise.
3595 (VEC_R): Likewise.
3596 (VEC_base): Likewise.
3597 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3598 registers, we need to swap double words in little endian mode.
3599
3600 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3601 to be a container mode for 128-bit integer operations added in ISA
3602 2.07. Unlike TImode and PTImode, the preferred register set is
3603 the Altivec/VMX registers for the 128-bit operations.
3604
3605 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3606 declarations.
3607 (rs6000_split_128bit_ok_p): Likewise.
3608
3609 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3610 macros for creating ISA 2.07 normal and overloaded builtin
3611 functions with 3 arguments.
3612 (BU_P8V_OVERLOAD_3): Likewise.
3613 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3614 for use as overloaded functions.
3615 (VPERM_1TI_UNS): Likewise.
3616 (VSEL_1TI): Likewise.
3617 (VSEL_1TI_UNS): Likewise.
3618 (ST_INTERNAL_1ti): Likewise.
3619 (LD_INTERNAL_1ti): Likewise.
3620 (XXSEL_1TI): Likewise.
3621 (XXSEL_1TI_UNS): Likewise.
3622 (VPERM_1TI): Likewise.
3623 (VPERM_1TI_UNS): Likewise.
3624 (XXPERMDI_1TI): Likewise.
3625 (SET_1TI): Likewise.
3626 (LXVD2X_V1TI): Likewise.
3627 (STXVD2X_V1TI): Likewise.
3628 (VEC_INIT_V1TI): Likewise.
3629 (VEC_SET_V1TI): Likewise.
3630 (VEC_EXT_V1TI): Likewise.
3631 (EQV_V1TI): Likewise.
3632 (NAND_V1TI): Likewise.
3633 (ORC_V1TI): Likewise.
3634 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3635 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3636 overloaded builtin.
3637 (VADDUQM): Likewise.
3638 (VSUBCUQ): Likewise.
3639 (VADDEUQM): Likewise.
3640 (VADDECUQ): Likewise.
3641 (VSUBEUQM): Likewise.
3642 (VSUBECUQ): Likewise.
3643
3644 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3645 __int128_t and __uint128_t types.
3646 (__uint128_type): Likewise.
3647 (altivec_categorize_keyword): Add support for vector __int128_t,
3648 vector __uint128_t, vector __int128, and vector unsigned __int128
3649 as a container type for TImode operations that need to be done in
3650 VSX/Altivec registers.
3651 (rs6000_macro_to_expand): Likewise.
3652 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3653 to support 128-bit integer instructions vaddcuq, vadduqm,
3654 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3655 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3656
3657 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3658 for V1TImode, and set up preferences to use VSX/Altivec registers.
3659 Setup VSX reload handlers.
3660 (rs6000_debug_reg_global): Likewise.
3661 (rs6000_init_hard_regno_mode_ok): Likewise.
3662 (rs6000_preferred_simd_mode): Likewise.
3663 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3664 (easy_altivec_constant): Likewise.
3665 (output_vec_const_move): Likewise.
3666 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3667 simple move.
3668 (rs6000_expand_vector_extract): Likewise.
3669 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3670 addressing.
3671 (rs6000_const_vec): Add support for V1TImode.
3672 (rs6000_emit_le_vsx_load): Swap double words when loading or
3673 storing TImode/V1TImode.
3674 (rs6000_emit_le_vsx_store): Likewise.
3675 (rs6000_emit_le_vsx_move): Likewise.
3676 (rs6000_emit_move): Add support for V1TImode.
3677 (altivec_expand_ld_builtin): Likewise.
3678 (altivec_expand_st_builtin): Likewise.
3679 (altivec_expand_vec_init_builtin): Likewise.
3680 (altivec_expand_builtin): Likewise.
3681 (rs6000_init_builtins): Add support for V1TImode type. Add
3682 support for ISA 2.07 128-bit integer builtins. Define type names
3683 for the VSX/Altivec vector types.
3684 (altivec_init_builtins): Add support for overloaded vector
3685 functions with V1TImode type.
3686 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3687 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3688 external function.
3689 (rs6000_split_128bit_ok_p): Likewise.
3690 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3691 __int128_t and vector __uint128_t.
3692
3693 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3694 and mode attributes.
3695 (VSX_M): Likewise.
3696 (VSX_M2): Likewise.
3697 (VSm): Likewise.
3698 (VSs): Likewise.
3699 (VSr): Likewise.
3700 (VSv): Likewise.
3701 (VS_scalar): Likewise.
3702 (VS_double): Likewise.
3703 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3704
3705 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3706 we support the ISA 2.07 128-bit integer arithmetic instructions.
3707 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3708 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3709 and TImode types for use with the builtin functions.
3710 (V1TI_type_node): Likewise.
3711 (unsigned_V1TI_type_node): Likewise.
3712 (intTI_type_internal_node): Likewise.
3713 (uintTI_type_internal_node): Likewise.
3714
3715 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3716 128-bit builtin functions.
3717 (UNSPEC_VADDEUQM): Likewise.
3718 (UNSPEC_VADDECUQ): Likewise.
3719 (UNSPEC_VSUBCUQ): Likewise.
3720 (UNSPEC_VSUBEUQM): Likewise.
3721 (UNSPEC_VSUBECUQ): Likewise.
3722 (VM): Add V1TImode to vector mode iterators.
3723 (VM2): Likewise.
3724 (VI_unit): Likewise.
3725 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3726 (altivec_vaddcuq): Likewise.
3727 (altivec_vsubuqm): Likewise.
3728 (altivec_vsubcuq): Likewise.
3729 (altivec_vaddeuqm): Likewise.
3730 (altivec_vaddecuq): Likewise.
3731 (altivec_vsubeuqm): Likewise.
3732 (altivec_vsubecuq): Likewise.
3733
3734 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3735 mode iterators.
3736 (BOOL_128): Likewise.
3737 (BOOL_REGS_OUTPUT): Likewise.
3738 (BOOL_REGS_OP1): Likewise.
3739 (BOOL_REGS_OP2): Likewise.
3740 (BOOL_REGS_UNARY): Likewise.
3741 (BOOL_REGS_AND_CR0): Likewise.
3742
3743 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3744 128-bit integer builtin support.
3745 (vec_vadduqm): Likewise.
3746 (vec_vaddecuq): Likewise.
3747 (vec_vaddeuqm): Likewise.
3748 (vec_vsubecuq): Likewise.
3749 (vec_vsubeuqm): Likewise.
3750 (vec_vsubcuq): Likewise.
3751 (vec_vsubuqm): Likewise.
3752
3753 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3754 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3755 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3756 128-bit integer add/subtract to ISA 2.07.
3757
3758 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3759
3760 * config/arc/arc.c (arc_predicate_delay_insns):
3761 Fix third argument passed to conditionalize_nonjump.
3762
3763 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3764
3765 * config/aarch64/aarch64-builtins.c
3766 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3767 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3768 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3769 instead of __builtin_lfloor.
3770 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3771
3772 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3773
3774 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3775 (tree_ssa_ifcombine_bb_1): New function.
3776 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3777 is an empty forwarder block to then_bb or vice versa and then_bb
3778 and else_bb are effectively swapped.
3779
3780 2014-03-12 Christian Bruel <christian.bruel@st.com>
3781
3782 PR target/60264
3783 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3784 REG_CFA_DEF_CFA note.
3785 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3786 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3787
3788 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3789
3790 PR tree-optimization/60454
3791 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3792
3793 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3794
3795 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3796 Do not define target_cpu_default2 to generic.
3797 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3798 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3799 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3800
3801 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3802 Marc Glisse <marc.glisse@inria.fr>
3803
3804 PR tree-optimization/60502
3805 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3806 instead of build_low_bits_mask.
3807
3808 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3809
3810 PR middle-end/60482
3811 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3812 if there are multiple uses, but op doesn't live on E edge.
3813 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3814 clobber stmts before __builtin_unreachable.
3815
3816 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3817
3818 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3819 hard_frame_pointer_rtx.
3820 * cse.c (cse_insn): Remove volatile check.
3821 * cselib.c (cselib_process_insn): Likewise.
3822 * dse.c (scan_insn): Likewise.
3823
3824 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3825
3826 * config/arc/arc.c (conditionalize_nonjump): New function,
3827 broken out of ...
3828 (arc_ifcvt): ... this.
3829 (arc_predicate_delay_insns): Use it.
3830
3831 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3832
3833 * config/arc/predicates.md (extend_operand): During/after reload,
3834 allow const_int_operand.
3835 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3836 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3837 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3838 to "i".
3839 (umulsi3_highpart_i): Likewise.
3840
3841 2014-03-11 Richard Biener <rguenther@suse.de>
3842
3843 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3844 Add asserts to guard possible wrong-code bugs.
3845
3846 2014-03-11 Richard Biener <rguenther@suse.de>
3847
3848 PR tree-optimization/60429
3849 PR tree-optimization/60485
3850 * tree-ssa-structalias.c (set_union_with_increment): Properly
3851 take into account all fields that overlap the shifted vars.
3852 (do_sd_constraint): Likewise.
3853 (do_ds_constraint): Likewise.
3854 (get_constraint_for_ptr_offset): Likewise.
3855
3856 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3857
3858 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3859 (nios2_compute_frame_layout):
3860 Add calculation of cfun->machine->fp_save_offset.
3861 (nios2_expand_prologue): Correct setting of frame pointer register
3862 in prologue.
3863 (nios2_expand_epilogue): Update recovery of stack pointer from
3864 frame pointer accordingly.
3865 (nios2_initial_elimination_offset): Update calculation of offset
3866 for eliminating to HARD_FRAME_POINTER_REGNUM.
3867
3868 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3869
3870 PR ipa/60457
3871 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3872 cgraph_get_create_node on VAR_DECLs.
3873
3874 2014-03-10 Richard Biener <rguenther@suse.de>
3875
3876 PR middle-end/60474
3877 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3878
3879 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3880
3881 * config/vms/vms.opt (vms_float_format): New variable.
3882
3883 2014-03-08 Tobias Burnus <burnus@net-b.de>
3884
3885 * doc/invoke.texi (-fcilkplus): Update implementation status.
3886
3887 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3888 Richard Biener <rguenther@suse.de>
3889
3890 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3891 consistently accross all TUs.
3892 (run_gcc): Enable -fshort-double automatically at link at link-time
3893 and disallow override.
3894
3895 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3896
3897 PR target/58271
3898 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3899 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3900 if they can't be used.
3901
3902 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3903
3904 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3905 for Solaris 11/x86 ld.
3906 * configure: Regenerate.
3907
3908 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3909
3910 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3911 (LIB_TLS_SPEC): Save as ld_tls_libs.
3912 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3913 (HAVE_AS_IX86_TLSLDM): New test.
3914 * configure, config.in: Regenerate.
3915 * config/i386/i386.c (legitimize_tls_address): Fall back to
3916 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3917 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3918 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3919 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3920
3921 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3922
3923 * common.opt (fira-loop-pressure): Mark as optimization.
3924
3925 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3926
3927 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
3928 an OpenMP mappable type.
3929
3930 2014-03-06 Matthias Klose <doko@ubuntu.com>
3931
3932 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
3933 MULTILIB_OSDIRNAMES is not defined.
3934
3935 2014-03-06 Jakub Jelinek <jakub@redhat.com>
3936 Meador Inge <meadori@codesourcery.com>
3937
3938 PR target/58595
3939 * config/arm/arm.c (arm_tls_symbol_p): Remove.
3940 (arm_legitimize_address): Call legitimize_tls_address for any
3941 arm_tls_referenced_p expression, handle constant addend. Call it
3942 before testing for !TARGET_ARM.
3943 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
3944
3945 2014-03-06 Richard Biener <rguenther@suse.de>
3946
3947 PR middle-end/60445
3948 PR lto/60424
3949 PR lto/60427
3950 Revert
3951 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3952
3953 * tree-streamer.c (record_common_node): Assert we don't record
3954 nodes with type double.
3955 (preload_common_node): Skip type double, complex double and double
3956 pointer since it is now frontend dependent due to fshort-double option.
3957
3958 2014-03-06 Richard Biener <rguenther@suse.de>
3959
3960 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
3961 or -fno-lto is specified and the linker has full plugin support.
3962 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
3963 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
3964 * lto-wrapper.c (merge_and_complain): Merge compile-time
3965 optimization levels.
3966 (run_gcc): And pass it through to the link options.
3967
3968 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
3969
3970 PR debug/60381
3971 Revert:
3972 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3973 PR debug/59992
3974 * cselib.c (remove_useless_values): Skip to avoid quadratic
3975 behavior if the condition moved from...
3976 (cselib_process_insn): ... here holds.
3977
3978 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3979
3980 PR plugins/59335
3981 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
3982 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
3983
3984 PR plugins/59335
3985 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
3986 (TM_H): Add x86-tune.def.
3987
3988 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3989
3990 * config/aarch64/aarch64.c (generic_tunings):
3991 Use cortexa57_extra_costs.
3992
3993 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3994
3995 PR lto/60404
3996 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
3997 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
3998 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
3999 cost for in_lto_p.
4000
4001 2014-03-04 Heiher <r@hev.cc>
4002
4003 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4004 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4005
4006 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4007
4008 * config/i386/predicates.md (const2356_operand): Change to ...
4009 (const2367_operand): ... this.
4010 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4011 const2367_operand.
4012 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4013 (*avx512pf_scatterpf<mode>sf): Ditto.
4014 (avx512pf_scatterpf<mode>df): Ditto.
4015 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4016 (*avx512pf_scatterpf<mode>df): Ditto.
4017 * config/i386/i386.c (ix86_expand_builtin): Update
4018 incorrect hint operand error message.
4019
4020 2014-03-04 Richard Biener <rguenther@suse.de>
4021
4022 * lto-section-in.c (lto_get_section_data): Fix const cast.
4023
4024 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4025
4026 * tree-streamer.c (record_common_node): Assert we don't record
4027 nodes with type double.
4028 (preload_common_node): Skip type double, complex double and double
4029 pointer since it is now frontend dependent due to fshort-double option.
4030
4031 2014-03-04 Richard Biener <rguenther@suse.de>
4032
4033 PR lto/60405
4034 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4035 (lto_input_toplevel_asms): Likewise.
4036 * lto-section-in.c (lto_get_section_data): Instead do it here
4037 for every section.
4038
4039 2014-03-04 Richard Biener <rguenther@suse.de>
4040
4041 PR tree-optimization/60382
4042 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4043 dead PHIs a reduction.
4044
4045 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4046
4047 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4048 hint value.
4049 (_mm_prefetch): Move out of GCC target("sse") pragma.
4050 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4051 GCC target("prfchw") pragma.
4052 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4053 for locality <= 2.
4054 * config/i386/i386.c (ix86_option_override_internal): Enable
4055 -mprfchw with -mprefetchwt1.
4056
4057 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4058
4059 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4060 Mark as varying.
4061
4062 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4063
4064 * opts.h (CL_PCH_IGNORE): Define.
4065 * targhooks.c (option_affects_pch_p):
4066 Return false for options that have CL_PCH_IGNORE set.
4067 * opt-functions.awk: Process PchIgnore.
4068 * doc/options.texi: Document PchIgnore.
4069
4070 * config/arc/arc.opt (misize): Add PchIgnore property.
4071
4072 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4073
4074 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4075 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4076 constraint on constants to permit them being loaded into
4077 GENERAL_REGS or BASE_REGS.
4078
4079 2014-03-03 Nick Clifton <nickc@redhat.com>
4080
4081 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4082 anti-cacnonical alternatives.
4083 (negandhi3_real): New pattern.
4084 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4085
4086 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4087
4088 * config/avr/avr-mcus.def: Remove atxmega16x1.
4089 * config/avr/avr-tables.opt: Regenerate.
4090 * config/avr/t-multilib: Regenerate.
4091 * doc/avr-mmcu.texi: Regenerate.
4092
4093 2014-03-03 Tobias Grosser <tobias@grosser.es>
4094 Mircea Namolaru <mircea.namolaru@inria.fr>
4095
4096 PR tree-optimization/58028
4097 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4098 scalar dimensions.
4099
4100 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4101
4102 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4103 not handled by recognizers.
4104
4105 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4106
4107 PR middle-end/60175
4108 * function.c (expand_function_end): Don't emit
4109 clobber_return_register sequence if clobber_after is a BARRIER.
4110 * cfgexpand.c (construct_exit_block): Append instructions before
4111 return_label to prev_bb.
4112
4113 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4114
4115 * config/rs6000/constraints.md: Document reserved use of "wc".
4116
4117 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4118
4119 PR ipa/60150
4120 * ipa.c (function_and_variable_visibility): When dissolving comdat
4121 group, also set all symbols to local.
4122
4123 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4124
4125 PR ipa/60306
4126
4127 Revert:
4128 2013-12-14 Jan Hubicka <jh@suse.cz>
4129 PR middle-end/58477
4130 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4131
4132 2014-03-02 Jon Beniston <jon@beniston.com>
4133
4134 PR bootstrap/48230
4135 PR bootstrap/50927
4136 PR bootstrap/52466
4137 PR target/46898
4138 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4139 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4140 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4141 (simple_return, *simple_return): New patterns
4142 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4143 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4144
4145 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4146
4147 * dwarf2out.c (gen_subprogram_die): Tidy.
4148
4149 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4150
4151 PR target/60071
4152 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4153 (*mov_t_msb_neg_negc): ... this new insn.
4154
4155 2014-02-28 Jason Merrill <jason@redhat.com>
4156
4157 PR c++/58678
4158 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4159 function.
4160
4161 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4162
4163 PR c++/60314
4164 * dwarf2out.c (decltype_auto_die): New static.
4165 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4166 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4167 (is_cxx_auto): Likewise.
4168
4169 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4170
4171 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4172 we are not using general regs only.
4173
4174 2014-02-28 Richard Biener <rguenther@suse.de>
4175
4176 PR target/60280
4177 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4178 previous fix and only allow to remove trivial pre-headers
4179 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4180 (remove_forwarder_block): Properly update the latch of a loop.
4181
4182 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4183
4184 PR debug/59992
4185 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4186 (cselib_preserved_hash_table): New.
4187 (preserve_constants_and_equivs): Move preserved vals to it.
4188 (cselib_find_slot): Look it up first.
4189 (cselib_init): Initialize it.
4190 (cselib_finish): Release it.
4191 (dump_cselib_table): Dump it.
4192
4193 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4194
4195 PR debug/59992
4196 * cselib.c (remove_useless_values): Skip to avoid quadratic
4197 behavior if the condition moved from...
4198 (cselib_process_insn): ... here holds.
4199
4200 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4201
4202 PR debug/57232
4203 * var-tracking.c (vt_initialize): Apply the same condition to
4204 preserve the CFA base value.
4205
4206 2014-02-28 Joey Ye <joey.ye@arm.com>
4207
4208 PR target/PR60169
4209 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4210 if reload in progress or completed.
4211
4212 2014-02-28 Tobias Burnus <burnus@net-b.de>
4213
4214 PR middle-end/60147
4215 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4216 NAMELIST_DECL.
4217
4218 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4219
4220 * doc/tm.texi.in (Condition Code Status): Update documention for
4221 relative locations of cc0-setter and cc0-user.
4222
4223 2014-02-27 Jeff Law <law@redhat.com>
4224
4225 PR rtl-optimization/52714
4226 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4227 into two independent simple sets, if I3 is a jump, ensure the
4228 pattern we place into I3 is a (set (pc) ...).
4229
4230 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4231 Jeff Law <law@redhat.com>
4232
4233 PR rtl-optimization/49847
4234 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4235 are in different blocks.
4236 * doc/tm.texi (Condition Code Status): Update documention for
4237 relative locations of cc0-setter and cc0-user.
4238
4239 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4240
4241 PR target/59222
4242 * lra.c (lra_emit_add): Check SUBREG too.
4243
4244 2014-02-27 Andreas Schwab <schwab@suse.de>
4245
4246 * config/m68k/m68k.c (m68k_option_override): Disable
4247 -flive-range-shrinkage for classic m68k.
4248 (m68k_override_options_after_change): Likewise.
4249
4250 2014-02-27 Marek Polacek <polacek@redhat.com>
4251
4252 PR middle-end/59223
4253 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4254 -Wmaybe-uninitialized.
4255
4256 2014-02-27 Alan Modra <amodra@gmail.com>
4257
4258 PR target/57936
4259 * reload1.c (emit_input_reload_insns): When reload_override_in,
4260 set old to rl->in_reg when rl->in_reg is a subreg.
4261
4262 2014-02-26 Richard Biener <rguenther@suse.de>
4263
4264 PR bootstrap/60343
4265 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4266
4267 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4268
4269 * common/config/i386/predicates.md (const1256_operand): Remove.
4270 (const2356_operand): New.
4271 (const_1_to_2_operand): Remove.
4272 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4273 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4274 (*avx512pf_gatherpf<mode>sf): Ditto.
4275 (avx512pf_gatherpf<mode>df): Ditto.
4276 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4277 (*avx512pf_gatherpf<mode>df): Ditto.
4278 (avx512pf_scatterpf<mode>sf): Ditto.
4279 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4280 (*avx512pf_scatterpf<mode>sf): Ditto.
4281 (avx512pf_scatterpf<mode>df): Ditto.
4282 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4283 (*avx512pf_scatterpf<mode>df): Ditto.
4284 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4285
4286 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4287
4288 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4289 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4290 (_mm512_mask_testn_epi64_mask): Move to ...
4291 * config/i386/avx512cdintrin.h: Here.
4292 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4293 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4294 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4295 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4296 TARGET_AVX512F from TARGET_AVX512CD.
4297
4298 2014-02-26 Richard Biener <rguenther@suse.de>
4299
4300 PR ipa/60327
4301 * ipa.c (walk_polymorphic_call_targets): Properly guard
4302 call to inline_update_overall_summary.
4303
4304 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4305
4306 PR target/60280
4307 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4308 and latches only if requested. Fix latch if it is removed.
4309 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4310 LOOPS_HAVE_PREHEADERS.
4311
4312 2014-02-25 Andrew Pinski <apinski@cavium.com>
4313
4314 * builtins.c (expand_builtin_thread_pointer): Create a new target
4315 when the target is NULL.
4316
4317 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4318
4319 PR rtl-optimization/60317
4320 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4321 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4322 * lra-assigns.c: Include params.h.
4323 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4324 other reload pseudos considerations.
4325
4326 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4327
4328 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4329 to use canonical form for nor<mode>3.
4330
4331 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4332
4333 PR target/55426
4334 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4335 conversions.
4336
4337 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4338
4339 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4340 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4341 (ix86_handle_option): Handle OPT_mprefetchwt1.
4342 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4343 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4344 PREFETCHWT1 CPUID.
4345 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4346 OPTION_MASK_ISA_PREFETCHWT1.
4347 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4348 (PTA_PREFETCHWT1): New.
4349 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4350 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4351 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4352 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4353 (*prefetch_avx512pf_<mode>_: Change into ...
4354 (*prefetch_prefetchwt1_<mode>: This.
4355 * config/i386/i386.opt (mprefetchwt1): New.
4356 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4357 (_mm_prefetch): Handle intent to write.
4358 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4359
4360 2014-02-25 Richard Biener <rguenther@suse.de>
4361
4362 PR middle-end/60291
4363 * emit-rtl.c (mem_attrs_htab): Remove.
4364 (mem_attrs_htab_hash): Likewise.
4365 (mem_attrs_htab_eq): Likewise.
4366 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4367 (init_emit_once): Do not allocate mem_attrs_htab.
4368
4369 2014-02-25 Richard Biener <rguenther@suse.de>
4370
4371 PR lto/60319
4372 * lto-opts.c (lto_write_options): Output non-explicit conservative
4373 -fwrapv, -fno-trapv and -fno-strict-overflow.
4374 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4375 (run_gcc): And pass them through.
4376
4377 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4378
4379 * sel-sched.c (calculate_new_fences): New parameter ptime.
4380 Calculate it as a maximum over all fence cycles.
4381 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4382 Print the final schedule timing when sched_verbose.
4383
4384 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4385
4386 PR rtl-optimization/60292
4387 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4388 bit fot the fence instruction.
4389
4390 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4391
4392 * calls.h: Fix typo in comment.
4393
4394 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4395
4396 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4397 adjusting offsetable addresses.
4398
4399 2014-02-24 Guozhi Wei <carrot@google.com>
4400
4401 * sparseset.h (sparseset_pop): Fix the wrong index.
4402
4403 2014-02-24 Walter Lee <walt@tilera.com>
4404
4405 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4406 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4407 triplet.
4408 * common/config/tilegx/tilegx-common.c
4409 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4410 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4411 (LINK_SPEC): Ditto.
4412 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4413 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4414 (tilegx_gimplify_va_arg_expr): Handle big endian.
4415 (tilegx_expand_unaligned_load): Ditto.
4416 (tilegx_expand_unaligned_store): Ditto.
4417 (TARGET_RETURN_IN_MSB): New.
4418 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4419 (TARGET_ENDIAN_DEFAULT): New.
4420 (TARGET_BIG_ENDIAN): Handle big endian.
4421 (BYTES_BIG_ENDIAN): Ditto.
4422 (WORDS_BIG_ENDIAN): Ditto.
4423 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4424 (ENDIAN_SPEC): New.
4425 (EXTRA_SPECS): New.
4426 * config/tilegx/tilegx.md (extv): Handle big endian.
4427 (extzv): Ditto.
4428 (insn_st<n>): Ditto.
4429 (insn_st<n>_add<bitsuffix>): Ditto.
4430 (insn_stnt<n>): Ditto.
4431 (insn_stnt<n>_add<bitsuffix>):Ditto.
4432 (vec_interleave_highv8qi): Handle big endian.
4433 (vec_interleave_highv8qi_be): New.
4434 (vec_interleave_highv8qi_le): New.
4435 (insn_v1int_h): Handle big endian.
4436 (vec_interleave_lowv8qi): Handle big endian.
4437 (vec_interleave_lowv8qi_be): New.
4438 (vec_interleave_lowv8qi_le): New.
4439 (insn_v1int_l): Handle big endian.
4440 (vec_interleave_highv4hi): Handle big endian.
4441 (vec_interleave_highv4hi_be): New.
4442 (vec_interleave_highv4hi_le): New.
4443 (insn_v2int_h): Handle big endian.
4444 (vec_interleave_lowv4hi): Handle big endian.
4445 (vec_interleave_lowv4hi_be): New.
4446 (vec_interleave_lowv4hi_le): New.
4447 (insn_v2int_l): Handle big endian.
4448 (vec_interleave_highv2si): Handle big endian.
4449 (vec_interleave_highv2si_be): New.
4450 (vec_interleave_highv2si_le): New.
4451 (insn_v4int_h): Handle big endian.
4452 (vec_interleave_lowv2si): Handle big endian.
4453 (vec_interleave_lowv2si_be): New.
4454 (vec_interleave_lowv2si_le): New.
4455 (insn_v4int_l): Handle big endian.
4456 * config/tilegx/tilegx.opt (mbig-endian): New option.
4457 (mlittle-endian): New option.
4458 * doc/install.texi: Document tilegxbe-linux.
4459 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4460
4461 2014-02-24 Martin Jambor <mjambor@suse.cz>
4462
4463 PR ipa/60266
4464 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4465 there are no parameter descriptors.
4466
4467 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4468
4469 PR rtl-optimization/60268
4470 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4471 initialization to ...
4472 (sched_rgn_init): ... here.
4473 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4474
4475 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4476
4477 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4478 names.
4479
4480 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4481
4482 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4483 definition.
4484
4485 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4486
4487 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4488 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4489
4490 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4491
4492 * config/microblaze/predicates.md: Add cmp_op predicate.
4493 * config/microblaze/microblaze.md: Add branch_compare instruction
4494 which uses cmp_op predicate and emits cmp insn before branch.
4495 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4496 to microblaze_expand_conditional_branch and consolidate logic.
4497 (microblaze_expand_conditional_branch): emit branch_compare
4498 insn instead of handling cmp op separate from branch insn.
4499
4500 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4501
4502 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4503 to permit subregs.
4504
4505 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4506
4507 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4508 define_insn with define_expand and new define_insn
4509 *altivec_lve<VI_char>x_internal.
4510 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4511 and new define_insn *altivec_stve<VI_char>x_internal.
4512 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4513 prototype.
4514 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4515 lve*x built-ins.
4516 (altivec_expand_stvex_be): New function.
4517
4518 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4519
4520 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4521 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4522 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4523 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4524
4525 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4526
4527 PR target/60298
4528 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4529 instead of emit_move_insn.
4530
4531 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4532
4533 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4534 vspltw with vsldoi.
4535 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4536 gen_altivec_vsumsws.
4537
4538 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4539
4540 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4541 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4542 (altivec_lvxl_<mode>): New define_expand incorporating
4543 -maltivec=be semantics where needed.
4544 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4545 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4546 semantics where needed.
4547 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4548 (altivec_stvx_<mode>): New define_expand incorporating
4549 -maltivec=be semantics where needed.
4550 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4551 VM2 iterator instead of V4SI.
4552 (altivec_stvxl_<mode>): New define_expand incorporating
4553 -maltivec=be semantics where needed.
4554 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4555 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4556 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4557 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4558 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4559 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4560 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4561 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4562 ALTIVEC_BUILTIN_STVXL.
4563 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4564 (altivec_expand_stvx_be): Likewise.
4565 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4566 (altivec_expand_lvx_be): Likewise.
4567 (altivec_expand_stvx_be): Likewise.
4568 (altivec_expand_builtin): Add cases for
4569 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4570 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4571 (altivec_init_builtins): Add definitions for
4572 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4573 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4574
4575 2014-02-21 Catherine Moore <clm@codesourcery.com>
4576
4577 * doc/invoke.texi (mvirt, mno-virt): Document.
4578 * config/mips/mips.opt (mvirt): New option.
4579 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4580
4581 2014-02-21 Richard Biener <rguenther@suse.de>
4582
4583 PR tree-optimization/60276
4584 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4585 (STMT_VINFO_MIN_NEG_DIST): New macro.
4586 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4587 STMT_VINFO_MIN_NEG_DIST.
4588 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4589 made for negative dependence distances still hold.
4590
4591 2014-02-21 Richard Biener <rguenther@suse.de>
4592
4593 PR middle-end/60291
4594 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4595 DECL_INITIAL for globals not in the current function context.
4596
4597 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4598
4599 PR tree-optimization/56490
4600 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4601 * tree-ssa-uninit.c: Include params.h.
4602 (compute_control_dep_chain): Add num_calls argument, return false
4603 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4604 num_calls to recursive call.
4605 (find_predicates): Change dep_chain into normal array,
4606 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4607 variable and adjust compute_control_dep_chain caller.
4608 (find_def_preds): Likewise.
4609
4610 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4611
4612 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4613 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4614
4615 2014-02-21 Nick Clifton <nickc@redhat.com>
4616
4617 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4618 (pushhi1): Likewise.
4619 (popqi1): Add mode to pre_dec.
4620 (pophi1): Likewise.
4621
4622 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4623
4624 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4625 mode for mask of V8SFmode permutation.
4626
4627 2014-02-20 Richard Henderson <rth@redhat.com>
4628
4629 PR c++/60272
4630 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4631 a new pseudo for OLDVAL.
4632
4633 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4634
4635 PR target/57896
4636 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4637 gen_reg_rtx if d->testing_p.
4638 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4639 if d->testing_p and we will certainly return true.
4640 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4641 if d->testing_p.
4642
4643 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4644
4645 * emit-rtl.c (gen_reg_rtx): Assert that
4646 crtl->emit.regno_pointer_align_length is non-zero.
4647
4648 2014-02-20 Richard Henderson <rth@redhat.com>
4649
4650 PR c++/60272
4651 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4652 on failure the store back into EXPECT.
4653
4654 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4655 Sandra Loosemore <sandra@codesourcery.com>
4656
4657 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4658 * config/nios2/nios2.c (nios2_function_profiler): Add
4659 -fPIC (flag_pic == 2) support.
4660 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4661 (nios2_large_offset_p): New function.
4662 (nios2_unspec_reloc_p): Move up position, update to use
4663 nios2_large_offset_p.
4664 (nios2_unspec_address): Remove function.
4665 (nios2_unspec_offset): New function.
4666 (nios2_large_got_address): New function.
4667 (nios2_got_address): Add large offset support.
4668 (nios2_legitimize_tls_address): Update usage of removed and new
4669 functions.
4670 (nios2_symbol_binds_local_p): New function.
4671 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4672 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4673 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4674 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4675 processing for (const (unspec ...)).
4676 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4677
4678 2014-02-20 Richard Biener <rguenther@suse.de>
4679
4680 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4681 doing the substitution.
4682 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4683
4684 2014-02-20 Martin Jambor <mjambor@suse.cz>
4685
4686 PR ipa/55260
4687 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4688 info when checking whether lattices are bottom.
4689
4690 2014-02-20 Richard Biener <rguenther@suse.de>
4691
4692 PR middle-end/60221
4693 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4694 regions at -O0.
4695
4696 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4697
4698 PR ipa/58555
4699 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4700 parameter specifying the scaling.
4701 (inline_call): Update.
4702 (want_inline_recursively): Guard division by zero.
4703 (recursive_inlining): Update.
4704 * ipa-inline.h (clone_inlined_nodes): Update.
4705
4706 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4707
4708 PR target/60204
4709 * config/i386/i386.c (classify_argument): Pass structures of size
4710 64 bytes or less in register.
4711
4712 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4713 Kirill Yukhin <kirill.yukhin@intel.com>
4714
4715 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4716 (_mm_rcp28_round_ss): Ditto.
4717 (_mm_rsqrt28_round_sd): Ditto.
4718 (_mm_rsqrt28_round_ss): Ditto.
4719 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4720 (_mm_rcp14_round_ss): Ditto.
4721 (_mm_rsqrt14_round_sd): Ditto.
4722 (_mm_rsqrt14_round_ss): Ditto.
4723 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4724 the first input operand, get rid of match_dup.
4725 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4726 attribute to sse.
4727 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4728 Ditto.
4729 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4730 operand as the first input operand, set type attribute.
4731 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4732 Set type attribute.
4733 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4734 operand as the first input operand, set type attribute.
4735
4736 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4737
4738 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4739 bit of zero.
4740
4741 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4742
4743 PR target/60207
4744 * config/i386/i386.c (construct_container): Remove TFmode check
4745 for X86_64_INTEGER_CLASS.
4746
4747 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4748
4749 PR target/59794
4750 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4751 only when -Wpsabi is enabled.
4752
4753 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4754
4755 PR target/59799
4756 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4757 passing arrays in registers are the same as for structs, so remove the
4758 special case for them.
4759
4760 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4761
4762 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4763 destination type, extract only the valid bits if the source type is not
4764 integral and has a different mode.
4765
4766 2014-02-19 Richard Biener <rguenther@suse.de>
4767
4768 PR ipa/60243
4769 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4770 for all calls.
4771
4772 2014-02-19 Richard Biener <rguenther@suse.de>
4773
4774 PR ipa/60243
4775 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4776 (ipa_modify_call_arguments): Emit an argument load explicitely and
4777 preserve virtual SSA form there and for the replacement call.
4778 Do not update SSA form nor free dominance info.
4779
4780 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4781
4782 * ipa.c (function_and_variable_visibility): Also clear WEAK
4783 flag when disolving COMDAT_GROUP.
4784
4785 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4786
4787 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4788 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4789 not devirtualizing.
4790 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4791 do more sanity checks.
4792 (detect_type_change): Return true when giving up early.
4793 (compute_complex_assign_jump_func): Fix type parameter of
4794 ipa_set_ancestor_jf.
4795 (compute_complex_ancestor_jump_func): Likewise.
4796 (update_jump_functions_after_inlining): Fix updating of
4797 ancestor function.
4798 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4799
4800 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4801
4802 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4803 inline clones when edge disappears.
4804
4805 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4806
4807 PR target/60203
4808 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4809 Split 64-bit moves into 2 patterns. Do not allow the use of
4810 direct move for TDmode in little endian, since the decimal value
4811 has little endian bytes within a word, but the 64-bit pieces are
4812 ordered in a big endian fashion, and normal subreg's of TDmode are
4813 not allowed.
4814 (mov<mode>_64bit_dm): Likewise.
4815 (movtd_64bit_nodm): Likewise.
4816
4817 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4818
4819 PR tree-optimization/60174
4820 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4821 statement of an SSA_NAME that occurs in an abnormal PHI node.
4822
4823 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4824
4825 PR sanitizer/60142
4826 * final.c (SEEN_BB): Remove.
4827 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4828 (final_scan_insn): Don't force_source_line on second
4829 NOTE_INSN_BASIC_BLOCK.
4830
4831 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4832
4833 PR target/60205
4834 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4835 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4836 (type_natural_mode): Warn ABI change when %zmm register is not
4837 available for AVX512F vector value passing.
4838
4839 2014-02-18 Kai Tietz <ktietz@redhat.com>
4840
4841 PR target/60193
4842 * config/i386/i386.c (ix86_expand_prologue): Use value in
4843 rax register as displacement when restoring %r10 or %rax.
4844 Fix wrong offset when restoring both registers.
4845
4846 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4847
4848 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4849 assertion with conditional return.
4850
4851 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4852 Uros Bizjak <ubizjak@gmail.com>
4853
4854 PR driver/60233
4855 * config/i386/driver-i386.c (host_detect_local_cpu): If
4856 YMM state is not saved by the OS, also clear has_f16c. Move
4857 CPUID 0x80000001 handling before YMM state saving checking.
4858
4859 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4860
4861 PR rtl-optimization/58960
4862 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4863 factored out from ...
4864 (sched_init): ... here.
4865 (free_global_sched_pressure_data): New, factored out from ...
4866 (sched_finish): ... here.
4867 * sched-int.h (free_global_sched_pressure_data): Declare.
4868 * sched-rgn.c (nr_regions_initial): New static global.
4869 (haifa_find_rgns): Initialize it.
4870 (schedule_region): Disable sched-pressure for the newly
4871 generated regions.
4872
4873 2014-02-17 Richard Biener <rguenther@suse.de>
4874
4875 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4876 release SSA defs of pattern stmts.
4877
4878 2014-02-17 Richard Biener <rguenther@suse.de>
4879
4880 * tree-inline.c (expand_call_inline): Release the virtual
4881 operand defined by the call we are about to inline.
4882
4883 2014-02-17 Richard Biener <rguenther@suse.de>
4884
4885 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4886
4887 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4888 Ilya Tocar <ilya.tocar@intel.com>
4889
4890 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4891 arguments order in builtin.
4892 (_mm512_permutexvar_epi64): Ditto.
4893 (_mm512_mask_permutexvar_epi64): Ditto
4894 (_mm512_maskz_permutexvar_epi32): Ditto
4895 (_mm512_permutexvar_epi32): Ditto
4896 (_mm512_mask_permutexvar_epi32): Ditto
4897
4898 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4899
4900 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4901 (p8_vmrgow): Likewise.
4902
4903 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4904
4905 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4906 endian targets.
4907
4908 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4909
4910 PR target/60203
4911 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4912 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4913 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4914 using direct move instructions on ISA 2.07. Also adjust
4915 instruction length for 64-bit.
4916 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4917 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4918
4919 2014-02-15 Alan Modra <amodra@gmail.com>
4920
4921 PR target/58675
4922 PR target/57935
4923 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4924 find_replacement on parts of insn rtl that might be reloaded.
4925
4926 2014-02-15 Richard Biener <rguenther@suse.de>
4927
4928 PR tree-optimization/60183
4929 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
4930 (tree_ssa_phiprop): Calculate and free post-dominators.
4931
4932 2014-02-14 Jeff Law <law@redhat.com>
4933
4934 PR rtl-optimization/60131
4935 * ree.c (get_extended_src_reg): New function.
4936 (combine_reaching_defs): Use it rather than assuming location of REG.
4937 (find_and_remove_re): Verify first operand of extension is
4938 a REG before adding the insns to the copy list.
4939
4940 2014-02-14 Roland McGrath <mcgrathr@google.com>
4941
4942 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
4943 * configure: Regenerated.
4944 * config.in: Regenerated.
4945 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
4946 instead of ASM_SHORT.
4947
4948 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
4949 Richard Earnshaw <rearnsha@arm.com>
4950
4951 PR rtl-optimization/59535
4952 * lra-constraints.c (process_alt_operands): Encourage alternative
4953 when unassigned pseudo class is superset of the alternative class.
4954 (inherit_reload_reg): Don't inherit when optimizing for code size.
4955 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
4956 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
4957 modes not less than 4 for Thumb1.
4958
4959 2014-02-14 Kyle McMartin <kyle@redhat.com>
4960
4961 PR pch/60010
4962 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
4963
4964 2014-02-14 Richard Biener <rguenther@suse.de>
4965
4966 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
4967 (get_frame_arg): Drop the assert with langhook types_compatible_p.
4968 Do not strip INDIRECT_REFs.
4969
4970 2014-02-14 Richard Biener <rguenther@suse.de>
4971
4972 PR lto/60179
4973 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
4974 DECL_FUNCTION_SPECIFIC_TARGET.
4975 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
4976 * tree-streamer-out.c (pack_ts_target_option): Remove.
4977 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
4978 (write_ts_function_decl_tree_pointers): Do not stream
4979 DECL_FUNCTION_SPECIFIC_TARGET.
4980 * tree-streamer-in.c (unpack_ts_target_option): Remove.
4981 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
4982 (lto_input_ts_function_decl_tree_pointers): Do not stream
4983 DECL_FUNCTION_SPECIFIC_TARGET.
4984
4985 2014-02-14 Jakub Jelinek <jakub@redhat.com>
4986
4987 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
4988 (get_initial_def_for_induction, vectorizable_induction): Ignore
4989 debug stmts when looking for exit_phi.
4990 (vectorizable_live_operation): Fix up condition.
4991
4992 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4993
4994 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
4995 nreverse() because it changes the content of original tree list.
4996
4997 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4998
4999 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5000 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5001
5002 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5003
5004 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5005 GNU coding standards.
5006
5007 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5008
5009 PR debug/60152
5010 * dwarf2out.c (gen_subprogram_die): Don't call
5011 add_calling_convention_attribute if subr_die is old_die.
5012
5013 2014-02-13 Sharad Singhai <singhai@google.com>
5014
5015 * doc/optinfo.texi: Fix order of nodes.
5016
5017 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5018
5019 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5020 operands[2], not operands[3].
5021
5022 2014-02-13 Richard Biener <rguenther@suse.de>
5023
5024 PR bootstrap/59878
5025 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5026 mention the possibility of an in-tree build.
5027 (CLooG): Update recommended version to 0.18.1, mention the
5028 possibility of an in-tree build and clarify that the ISL
5029 bundled with CLooG does not work.
5030
5031 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5032
5033 PR target/43546
5034 * expr.c (compress_float_constant): If x is a hard register,
5035 extend into a pseudo and then move to x.
5036
5037 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5038
5039 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5040 caused by bad second argument to warning_at() with -mhotpatch and
5041 nested functions (e.g. with gfortran).
5042
5043 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5044
5045 * opts.c (option_name): Remove "enabled by default" rider.
5046
5047 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5048
5049 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5050
5051 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5052 Uros Bizjak <ubizjak@gmail.com>
5053
5054 PR target/60151
5055 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5056 * configure: Regenerated.
5057
5058 2014-02-12 Richard Biener <rguenther@suse.de>
5059
5060 * vec.c (vec_prefix::calculate_allocation): Move as
5061 inline variant to vec.h.
5062 (vec_prefix::calculate_allocation_1): New out-of-line version.
5063 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5064 (vec_prefix::m_has_auto_buf): Rename to ...
5065 (vec_prefix::m_using_auto_storage): ... this.
5066 (vec_prefix::calculate_allocation): Inline the easy cases
5067 and dispatch to calculate_allocation_1 which doesn't need the
5068 prefix address.
5069 (va_heap::reserve): Use gcc_checking_assert.
5070 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5071 m_using_auto_storage.
5072 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5073 member and adjust.
5074 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5075 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5076 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5077
5078 2014-02-12 Richard Biener <rguenther@suse.de>
5079
5080 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5081 when we found a dependence.
5082
5083 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5084
5085 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5086 common code...
5087 (maybe_fold_stmt): ... into this new function.
5088 * omp-low.c (lower_omp): Update comment.
5089
5090 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5091 last use.
5092
5093 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5094 dereference.
5095
5096 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5097
5098 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5099 identifiers in comments.
5100 (cortexa53_extra_costs): Likewise.
5101 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5102 (cortexa7_extra_costs): Likewise.
5103 (cortexa12_extra_costs): Likewise.
5104 (cortexa15_extra_costs): Likewise.
5105 (v7m_extra_costs): Likewise.
5106
5107 2014-02-12 Richard Biener <rguenther@suse.de>
5108
5109 PR middle-end/60092
5110 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5111 of posix_memalign being successful.
5112 (lower_stmt): Restrict lowering of posix_memalign to when
5113 -ftree-bit-ccp is enabled.
5114
5115 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5116
5117 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5118 arg_loc.
5119 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5120
5121 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5122
5123 PR rtl-optimization/60116
5124 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5125 other_insn once the combination has been validated.
5126
5127 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5128
5129 PR lto/59468
5130 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5131 and wrapper.
5132 * ipa-devirt.c: Include demangle.h
5133 (odr_violation_reported): New static variable.
5134 (add_type_duplicate): Update odr_violations.
5135 (maybe_record_node): Add completep parameter; update it.
5136 (record_target_from_binfo): Add COMPLETEP parameter;
5137 update it as needed.
5138 (possible_polymorphic_call_targets_1): Likewise.
5139 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5140 rename FINAL to COMPLETE.
5141 (record_targets_from_bases): Sanity check we found the binfo;
5142 fix COMPLETEP updating.
5143 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5144 parameter, fix computing of COMPLETEP.
5145 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5146 at LTO time do demangling.
5147 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5148 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5149 parameter.
5150 (gimple_get_virt_method_for_binfo): Likewise.
5151 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5152 gimple_get_virt_method_for_vtable): Update prototypes.
5153
5154 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5155
5156 PR target/49008
5157 * genautomata.c (add_presence_absence): Fix typo with
5158 {final_}presence_list.
5159
5160 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5161
5162 PR target/60137
5163 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5164 for VSX/Altivec vectors that land in GPR registers.
5165
5166 2014-02-11 Richard Henderson <rth@redhat.com>
5167 Jakub Jelinek <jakub@redhat.com>
5168
5169 PR debug/59776
5170 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5171 around drhs if type conversion to lacc->type is not useless.
5172
5173 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5174
5175 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5176 tuning struct.
5177 (cortex-a57.cortex-a53): Likewise.
5178 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5179
5180 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5181
5182 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5183 arm_restrict_it.
5184
5185 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5186
5187 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5188 add_options_for_arm_vfp3.
5189
5190 2014-02-11 Jeff Law <law@redhat.com>
5191
5192 PR middle-end/54041
5193 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5194 object with an undesirable mode.
5195
5196 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5197
5198 PR libgomp/60107
5199 * config/i386/sol2-9.h: New file.
5200 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5201 *-*-solaris2.9*): Use it.
5202
5203 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5204
5205 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5206 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5207
5208 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5209
5210 * config/microblaze/microblaze.c: Extend mcpu version format
5211
5212 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5213
5214 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5215
5216 2014-02-10 Richard Henderson <rth@redhat.com>
5217
5218 PR target/59927
5219 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5220 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5221 ms-abi vs -mno-accumulate-outgoing-args.
5222 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5223 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5224 respect to ms-abi.
5225
5226 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5227
5228 PR middle-end/60080
5229 * cfgexpand.c (expand_asm_operands): Attach source location to
5230 ASM_INPUT rtx objects.
5231 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5232
5233 2014-02-10 Nick Clifton <nickc@redhat.com>
5234
5235 * config/mn10300/mn10300.c (popcount): New function.
5236 (mn10300_expand_prologue): Include saved registers in stack usage
5237 count.
5238
5239 2014-02-10 Jeff Law <law@redhat.com>
5240
5241 PR middle-end/52306
5242 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5243 when changing the SET_DEST of a prior insn to avoid an input reload.
5244
5245 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5246
5247 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5248 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5249 -mcall-openbsd, or -mcall-linux.
5250 (CC1_ENDIAN_BIG_SPEC): Remove.
5251 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5252 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5253 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5254 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5255 and %cc1_endian_default.
5256 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5257
5258 2014-02-10 Richard Biener <rguenther@suse.de>
5259
5260 PR tree-optimization/60115
5261 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5262 MEM_REF handling. Properly verify that the accesses are not
5263 out of the objects bound.
5264
5265 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5266
5267 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5268 coretex to cortex.
5269
5270 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5271
5272 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5273 proper constants and fix formatting.
5274 (possible_polymorphic_call_targets): Fix formatting.
5275
5276 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5277 Ilya Tocar <ilya.tocar@intel.com>
5278
5279 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5280 (_mm512_loadu_epi32): Renamed into...
5281 (_mm512_loadu_si512): This.
5282 (_mm512_storeu_epi32): Renamed into...
5283 (_mm512_storeu_si512): This.
5284 (_mm512_maskz_ceil_ps): Removed.
5285 (_mm512_maskz_ceil_pd): Ditto.
5286 (_mm512_maskz_floor_ps): Ditto.
5287 (_mm512_maskz_floor_pd): Ditto.
5288 (_mm512_floor_round_ps): Ditto.
5289 (_mm512_floor_round_pd): Ditto.
5290 (_mm512_ceil_round_ps): Ditto.
5291 (_mm512_ceil_round_pd): Ditto.
5292 (_mm512_mask_floor_round_ps): Ditto.
5293 (_mm512_mask_floor_round_pd): Ditto.
5294 (_mm512_mask_ceil_round_ps): Ditto.
5295 (_mm512_mask_ceil_round_pd): Ditto.
5296 (_mm512_maskz_floor_round_ps): Ditto.
5297 (_mm512_maskz_floor_round_pd): Ditto.
5298 (_mm512_maskz_ceil_round_ps): Ditto.
5299 (_mm512_maskz_ceil_round_pd): Ditto.
5300 (_mm512_expand_pd): Ditto.
5301 (_mm512_expand_ps): Ditto.
5302 * config/i386/i386.c (ix86_builtins): Remove
5303 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5304 (bdesc_args): Ditto.
5305 * config/i386/predicates.md (const1256_operand): New.
5306 (const_1_to_2_operand): Ditto.
5307 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5308 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5309 (*avx512pf_gatherpf<mode>sf): Ditto.
5310 (avx512pf_gatherpf<mode>df): Ditto.
5311 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5312 (*avx512pf_gatherpf<mode>df): Ditto.
5313 (avx512pf_scatterpf<mode>sf): Ditto.
5314 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5315 (*avx512pf_scatterpf<mode>sf): Ditto.
5316 (avx512pf_scatterpf<mode>df): Ditto.
5317 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5318 (*avx512pf_scatterpf<mode>df): Ditto.
5319 (avx512f_expand<mode>): Removed.
5320 (<shift_insn><mode>3<mask_name>): Change predicate type.
5321
5322 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5323
5324 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5325 not at the end of datarefs vector use ordered_remove to avoid
5326 reordering datarefs vector.
5327
5328 PR c/59984
5329 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5330 mark local addressable non-static vars as GOVD_PRIVATE
5331 instead of GOVD_LOCAL.
5332 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5333 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5334 than copying them.
5335
5336 PR middle-end/60092
5337 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5338 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5339 assume_aligned or alloc_align attributes.
5340 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5341 arguments. Handle also assume_aligned and alloc_align attributes.
5342 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5343 calls to functions with assume_aligned or alloc_align attributes.
5344 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5345
5346 2014-02-08 Terry Guo <terry.guo@arm.com>
5347
5348 * doc/invoke.texi: Document ARM -march=armv7e-m.
5349
5350 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5351
5352 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5353 flag on __cilkrts_rethrow builtin.
5354
5355 PR ipa/60026
5356 * ipa-cp.c (determine_versionability): Fail at -O0
5357 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5358 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5359
5360 Revert:
5361 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5362
5363 PR ipa/60026
5364 * tree-inline.c (copy_forbidden): Fail for
5365 __attribute__((optimize (0))) functions.
5366
5367 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5368
5369 * varpool.c: Include pointer-set.h.
5370 (varpool_remove_unreferenced_decls): Variables in other partitions
5371 will not be output; be however careful to not lose information
5372 about partitioning.
5373
5374 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5375
5376 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5377 lookup in the vtable constructor.
5378
5379 2014-02-07 Jeff Law <law@redhat.com>
5380
5381 PR target/40977
5382 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5383 define_insn_and_split.
5384
5385 * ipa-inline.c (inline_small_functions): Fix typos.
5386
5387 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5388
5389 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5390 (s390_can_use_return_insn): Declare.
5391 * config/s390/s390.h (EPILOGUE_USES): Define.
5392 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5393 instructions.
5394 (s390_chunkify_start): Handle return JUMP_LABELs.
5395 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5396 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5397 (s390_can_use_return_insn): New functions.
5398 (s390_fix_long_loop_prediction): Handle conditional returns.
5399 (TARGET_SET_UP_BY_PROLOGUE): Define.
5400 * config/s390/s390.md (ANY_RETURN): New code iterator.
5401 (*creturn, *csimple_return, return, simple_return): New patterns.
5402
5403 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5404
5405 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5406 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5407 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5408 REG_CFA_RESTORE list when deciding not to restore a register.
5409
5410 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5411
5412 * config/s390/s390.c: Include tree-pass.h and context.h.
5413 (s390_early_mach): New function, split out from...
5414 (s390_emit_prologue): ...here.
5415 (pass_data_s390_early_mach): New pass structure.
5416 (pass_s390_early_mach): New class.
5417 (s390_option_override): Create and register early_mach pass.
5418 Move to end of file.
5419
5420 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5421
5422 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5423 to match for the exit block.
5424
5425 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5426
5427 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5428 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5429 Reject misaligned operands.
5430
5431 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5432
5433 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5434
5435 2014-02-07 Richard Biener <rguenther@suse.de>
5436
5437 PR middle-end/60092
5438 * gimple-low.c (lower_builtin_posix_memalign): New function.
5439 (lower_stmt): Call it to lower posix_memalign in a way
5440 to make alignment info accessible.
5441
5442 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5443
5444 PR c++/60082
5445 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5446 __builtin_setjmp_receiver.
5447
5448 2014-02-07 Richard Biener <rguenther@suse.de>
5449
5450 PR middle-end/60092
5451 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5452 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5453 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5454 Handle BUILT_IN_POSIX_MEMALIGN.
5455 (find_func_clobbers): Likewise.
5456 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5457 (call_may_clobber_ref_p_1): Likewise.
5458
5459 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5460
5461 PR ipa/59918
5462 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5463 sanity check.
5464
5465 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5466
5467 PR ipa/59469
5468 * lto-cgraph.c (lto_output_node): Use
5469 symtab_get_symbol_partitioning_class.
5470 (lto_output_varpool_node): likewise.
5471 (symtab_get_symbol_partitioning_class): Move here from
5472 lto/lto-partition.c
5473 * cgraph.h (symbol_partitioning_class): Likewise.
5474 (symtab_get_symbol_partitioning_class): Declare.
5475
5476 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5477
5478 * ggc.h (ggc_internal_cleared_alloc): New macro.
5479 * vec.h (vec_safe_copy): Handle memory stats.
5480 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5481 * target-globals.c (save_target_globals): Likewise.
5482
5483 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5484
5485 PR target/60077
5486 * expr.c (emit_move_resolve_push): Export; be bit more selective
5487 on when to clear alias set.
5488 * expr.h (emit_move_resolve_push): Declare.
5489 * function.h (struct function): Add tail_call_marked.
5490 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5491 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5492 * config/i386/i386.md (TImode move expander): De not call
5493 ix86_expand_push.
5494 (FP push expanders): Preserve memory attributes.
5495 * config/i386/sse.md (push<mode>1): Remove.
5496 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5497 (ix86_expand_push): Remove.
5498 * config/i386/mmx.md (push<mode>1): Remove.
5499
5500 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5501
5502 PR rtl-optimization/60030
5503 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5504 lopart with paradoxical subreg before shifting it up by hprec.
5505
5506 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5507
5508 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5509 Remove extra newline at end of file.
5510 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5511 (arm_issue_rate): Handle cortexa57.
5512 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5513 (cortex-a57.cortex-a53): Likewise.
5514
5515 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5516
5517 PR target/59575
5518 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5519 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5520 bitmask.
5521 (arm_expand_prologue): Adjust all callers.
5522 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5523 info, registers also at the lowest numbered registers side. Use
5524 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5525 XEXP.
5526
5527 PR debug/59992
5528 * var-tracking.c (adjust_mems): Before adding a SET to
5529 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5530
5531 2014-02-06 Alan Modra <amodra@gmail.com>
5532
5533 PR target/60032
5534 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5535 change SDmode to DDmode when lra_in_progress.
5536
5537 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5538
5539 PR middle-end/59150
5540 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5541 free_data_ref on the dr first, and before goto again also set dr
5542 to the next dr. For simd_lane_access, free old datarefs[i] before
5543 overwriting it. For get_vectype_for_scalar_type failure, don't
5544 free_data_ref if simd_lane_access.
5545
5546 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5547
5548 PR target/60062
5549 * tree.h (opts_for_fn): New inline function.
5550 (opt_for_fn): Define.
5551 * config/i386/i386.c (ix86_function_regparm): Use
5552 opt_for_fn (decl, optimize) instead of optimize.
5553
5554 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5555
5556 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5557 for SYMBOL_REF in large memory model.
5558
5559 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5560
5561 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5562 and crypto support.
5563 (cortex-a57): Likewise.
5564 (cortex-a57.cortex-a53): Likewise.
5565
5566 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5567 Kugan Vivekanandarajah <kuganv@linaro.org>
5568
5569 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5570 unaligned_access.
5571 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5572
5573 2014-02-06 Richard Biener <rguenther@suse.de>
5574
5575 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5576 set_loop_copy and initialize_original_copy_tables.
5577
5578 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5579
5580 * config/aarch64/aarch64-simd.md
5581 (aarch64_ashr_simddi): Change QI to SI.
5582
5583 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5584 Jakub Jelinek <jakub@redhat.com>
5585
5586 PR middle-end/60013
5587 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5588 of the dataflow.
5589
5590 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5591
5592 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5593 CODE_FOR_altivec_vpku[hw]um to
5594 CODE_FOR_altivec_vpku[hw]um_direct.
5595 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5596 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5597 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5598 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5599
5600 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5601
5602 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5603 generation for -maltivec=be.
5604 (altivec_vsumsws): Simplify redundant test.
5605
5606 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5607
5608 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5609 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5610 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5611 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5612 gen_altivec_vpkuwum.
5613 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5614 BYTES_BIG_ENDIAN.
5615 (altivec_vpks<VI_char>ss): Likewise.
5616 (altivec_vpks<VI_char>us): Likewise.
5617 (altivec_vpku<VI_char>us): Likewise.
5618 (altivec_vpku<VI_char>um): Likewise.
5619 (altivec_vpku<VI_char>um_direct): New (copy of
5620 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5621 internal use).
5622 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5623 target is little endian and -maltivec=be is not specified.
5624 (*altivec_vupkhs<VU_char>_direct): New (copy of
5625 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5626 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5627 target is little endian and -maltivec=be is not specified.
5628 (*altivec_vupkls<VU_char>_direct): New (copy of
5629 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5630 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5631 little endian and -maltivec=be is not specified.
5632 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5633 little endian and -maltivec=be is not specified.
5634
5635 2014-02-05 Richard Henderson <rth@redhat.com>
5636
5637 PR debug/52727
5638 * combine-stack-adj.c: Revert r206943.
5639 * sched-int.h (struct deps_desc): Add last_args_size.
5640 * sched-deps.c (init_deps): Initialize it.
5641 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5642 contain REG_ARGS_SIZE notes.
5643
5644 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5645
5646 * lto-cgraph.c (asm_nodes_output): Make global.
5647 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5648 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5649 (driver_handle_option): Handle OPT_fwpa.
5650
5651 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5652
5653 PR ipa/59947
5654 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5655 a comment typo and formatting issue. If odr_hash hasn't been
5656 created, return vNULL and set *completep to false.
5657
5658 PR middle-end/57499
5659 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5660 bb with no successors.
5661
5662 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5663
5664 PR target/59718
5665 * doc/invoke.texi (-march): Clarify documentation for ARM.
5666 (-mtune): Likewise.
5667 (-mcpu): Likewise.
5668
5669 2014-02-05 Richard Biener <rguenther@suse.de>
5670
5671 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5672 when not vectorizing because of too many alias checks.
5673 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5674 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5675
5676 2014-02-05 Nick Clifton <nickc@redhat.com>
5677
5678 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5679 accept extended registers in any mode when compiling for the MN10300.
5680
5681 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5682
5683 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5684 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5685 sanitization attributes.
5686 (can_inline_edge_p): Likewise.
5687 (sanitize_attrs_match_for_inline_p): New function.
5688
5689 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5690
5691 * ipa-prop.c (detect_type_change): Shor circuit testing of
5692 type changes on THIS pointer.
5693
5694 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5695
5696 PR target/59777
5697 * config/pa/pa.c (legitimize_tls_address): Return original address
5698 if not passed a SYMBOL_REF rtx.
5699 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5700 addresses.
5701 (pa_emit_move_sequence): Simplify TLS source operands.
5702 (pa_legitimate_constant_p): Reject all TLS constants.
5703 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5704 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5705
5706 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5707
5708 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5709 groups when we know they are controlled by LTO.
5710 * varasm.c (default_binds_local_p_1): If object is in other partition,
5711 it will be resolved locally.
5712
5713 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5714
5715 * config/host-linux.c (linux_gt_pch_use_address): Don't
5716 use SSIZE_MAX because it is not always defined.
5717
5718 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5719
5720 PR bootstrap/59913
5721 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5722 threshold for pseudo splitting.
5723 (update_ebb_live_info): Process call argument hard registers and
5724 hard registers from insn definition too.
5725 (max_small_class_regs_num): New constant.
5726 (inherit_in_ebb): Update live hard regs through EBBs. Update
5727 reloads_num only for small register classes. Don't split for
5728 outputs of jumps.
5729
5730 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5731
5732 PR ipa/60058
5733 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5734 is non-null.
5735
5736 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5737
5738 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5739 visibility is safe.
5740
5741 2014-02-04 Marek Polacek <polacek@redhat.com>
5742
5743 * gdbinit.in (pel): Define.
5744
5745 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5746
5747 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5748 behavior.
5749
5750 2014-02-04 Richard Biener <rguenther@suse.de>
5751
5752 PR lto/59723
5753 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5754 in function context local.
5755 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5756 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5757 similar to LTO_imported_decl_ref.
5758
5759 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5760
5761 PR tree-optimization/60002
5762 * cgraphclones.c (build_function_decl_skip_args): Clear
5763 DECL_LANG_SPECIFIC.
5764
5765 PR tree-optimization/60023
5766 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5767 false to gsi_replace.
5768 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5769 has been in some EH region and vec_stmt could throw, add
5770 vec_stmt into the same EH region.
5771 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5772 has no lhs, ignore it.
5773 * internal-fn.c (expand_MASK_LOAD): Likewise.
5774
5775 PR ipa/60026
5776 * tree-inline.c (copy_forbidden): Fail for
5777 __attribute__((optimize (0))) functions.
5778
5779 PR other/58712
5780 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5781 is set, copy one less argument.
5782 (expand_simd_clones): Don't subtract clone_info->inbranch
5783 from simd_clone_struct_alloc argument.
5784
5785 PR rtl-optimization/57915
5786 * recog.c (simplify_while_replacing): If all unary/binary/relational
5787 operation arguments are constant, attempt to simplify those.
5788
5789 PR middle-end/59261
5790 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5791 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5792
5793 2014-02-04 Richard Biener <rguenther@suse.de>
5794
5795 PR tree-optimization/60012
5796 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5797 TBAA disambiguation to all DDRs.
5798
5799 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5800
5801 PR target/59788
5802 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5803 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5804
5805 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5806
5807 PR ipa/59882
5808 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5809
5810 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5811
5812 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5813 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5814
5815 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5816
5817 PR ipa/59831
5818 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5819 to figure out targets of polymorphic calls with known decl.
5820 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5821 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5822 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5823 (get_polymorphic_call_info): ... here.
5824 (get_polymorphic_call_info_from_invariant): New function.
5825
5826 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5827
5828 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5829 lookup via vtable pointer; check for type consistency
5830 and turn inconsitent facts into UNREACHABLE.
5831 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5832 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5833 type inconsistent querries; return UNREACHABLE instead.
5834
5835 2014-02-03 Richard Henderson <rth@twiddle.net>
5836
5837 PR tree-opt/59924
5838 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5839 already processed this node.
5840 (normalize_one_pred_1): Pass along mark_set.
5841 (normalize_one_pred): Create and destroy a pointer_set_t.
5842 (normalize_one_pred_chain): Likewise.
5843
5844 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5845
5846 PR gcov-profile/58602
5847 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5848
5849 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5850
5851 PR ipa/59831
5852 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5853 -fno-devirtualize; try to devirtualize by the knowledge of
5854 virtual table pointer given by aggregate propagation.
5855 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5856 (ipa_print_node_jump_functions): Dump also offset that
5857 is relevant for polymorphic calls.
5858 (determine_known_aggregate_parts): Add arg_type parameter; use it
5859 instead of determining the type from pointer type.
5860 (ipa_compute_jump_functions_for_edge): Update call of
5861 determine_known_aggregate_parts.
5862 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5863 (gimple_get_virt_method_for_binfo): ... here; simplify using
5864 vtable_pointer_value_to_vtable.
5865 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5866 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5867 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5868 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5869 POINTER_PLUS_EXPR.
5870 (vtable_pointer_value_to_binfo): ... here.
5871 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5872
5873 2014-02-03 Teresa Johnson <tejohnson@google.com>
5874
5875 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5876 redef of outer loop index variable.
5877
5878 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5879
5880 PR c++/53017
5881 PR c++/59211
5882 * doc/extend.texi (Function Attributes): Typo.
5883
5884 2014-02-03 Cong Hou <congh@google.com>
5885
5886 PR tree-optimization/60000
5887 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5888 if the vectorized statement is a store. A store statement can only
5889 appear at the end of pattern statements.
5890
5891 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5892
5893 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5894 (ix86_option_override_internal): Default long double to 64-bit for
5895 32-bit Bionic and to 128-bit for 64-bit Bionic.
5896
5897 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5898 TARGET_LONG_DOUBLE_128 is true.
5899 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5900
5901 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5902 (mlong-double-64): Negate -mlong-double-128.
5903 (mlong-double-128): New option.
5904
5905 * config/i386/i386-c.c (ix86_target_macros): Define
5906 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5907
5908 * doc/invoke.texi: Document -mlong-double-128.
5909
5910 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5911
5912 PR rtl-optimization/60024
5913 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5914
5915 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5916
5917 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5918
5919 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5920
5921 PR rtl-optimization/57662
5922 * sel-sched.c (code_motion_path_driver): Do not mark already not
5923 existing blocks in the visiting bitmap.
5924
5925 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5926
5927 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
5928 on the insn being emitted.
5929
5930 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
5931 Will Deacon <will.deacon@arm.com>
5932
5933 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
5934
5935 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5936
5937 * config/arm/arm-tables.opt: Regenerate.
5938
5939 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5940
5941 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
5942 for vector types other than V16QImode.
5943 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
5944 define_expand, and call altivec_expand_vec_perm_le when producing
5945 code with little endian element order.
5946 (*altivec_vperm_<mode>_internal): New insn having previous
5947 behavior of altivec_vperm_<mode>.
5948 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
5949 altivec_expand_vec_perm_le when producing code with little endian
5950 element order.
5951 (*altivec_vperm_<mode>_uns_internal): New insn having previous
5952 behavior of altivec_vperm_<mode>_uns.
5953
5954 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5955
5956 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
5957 (altivec_vsumsws): Add handling for -maltivec=be with a little
5958 endian target.
5959 (altivec_vsumsws_direct): New.
5960 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
5961 gen_altivec_vsumsws.
5962
5963 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5964
5965 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
5966 vtable_pointer_value_to_binfo): New functions.
5967 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
5968 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
5969
5970 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
5971
5972 * config/nios2/nios2.md (load_got_register): Initialize GOT
5973 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
5974 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
5975
5976 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5977
5978 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
5979 preserverd by passthrough, do not propagate the type.
5980
5981 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5982
5983 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
5984 (mips_atomic_assign_expand_fenv): New function.
5985 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
5986
5987 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5988
5989 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
5990 (__builtin_mips_set_fcsr): Likewise.
5991 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
5992 MIPS_USI_FTYPE_VOID.
5993 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
5994 (mips16_expand_set_fcsr): Likewise.
5995 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
5996 (mips16_set_fcsr_stub): Likewise.
5997 (mips16_get_fcsr_one_only_stub): New class.
5998 (mips16_set_fcsr_one_only_stub): Likewise.
5999 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6000 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6001 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6002 (hard_float): New availability predicate.
6003 (mips_builtins): Add get_fcsr and set_fcsr.
6004 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6005 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6006 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6007 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6008 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6009 patterns.
6010
6011 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6012
6013 * config/mips/mips.c (mips_one_only_stub): New class.
6014 (mips_need_mips16_rdhwr_p): Replace with...
6015 (mips16_rdhwr_stub): ...this new variable.
6016 (mips16_stub_call_address): New function.
6017 (mips16_rdhwr_one_only_stub): New class.
6018 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6019 (mips_output_mips16_rdhwr): Delete.
6020 (mips_finish_stub): New function.
6021 (mips_code_end): Use it to handle rdhwr stubs.
6022
6023 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6024
6025 PR target/60017
6026 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6027 when calculating size of integer atomic types.
6028
6029 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6030
6031 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6032
6033 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6034
6035 PR tree-optimization/60003
6036 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6037 * profile.c (branch_prob): Use gimple_call_builtin_p
6038 to check for BUILT_IN_SETJMP_RECEIVER.
6039 * tree-inline.c (copy_bb): Call notice_special_calls.
6040
6041 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6042
6043 PR bootstrap/59985
6044 * lra-constraints.c (process_alt_operands): Update reload_sum only
6045 on the first pass.
6046
6047 2014-01-31 Richard Henderson <rth@redhat.com>
6048
6049 PR middle-end/60004
6050 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6051 until after else_eh is processed.
6052
6053 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6054
6055 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6056 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6057 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6058 in smmintrin.h, remove them.
6059 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6060 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6061 * config/i386/i386.md (ROUND_SAE): Fix value.
6062 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6063 (const48_operand): New.
6064 * config/i386/subst.md (round), (round_expand): Use
6065 const_4_or_8_to_11_operand.
6066 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6067
6068 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6069
6070 * config/i386/constraints.md (Yk): Swap meaning with k.
6071 * config/i386/i386.md (movhi_internal): Change Yk to k.
6072 (movqi_internal): Ditto.
6073 (*k<logic><mode>): Ditto.
6074 (*andhi_1): Ditto.
6075 (*andqi_1): Ditto.
6076 (kandn<mode>): Ditto.
6077 (*<code>hi_1): Ditto.
6078 (*<code>qi_1): Ditto.
6079 (kxnor<mode>): Ditto.
6080 (kortestzhi): Ditto.
6081 (kortestchi): Ditto.
6082 (kunpckhi): Ditto.
6083 (*one_cmplhi2_1): Ditto.
6084 (*one_cmplqi2_1): Ditto.
6085 * config/i386/sse.md (): Change k to Yk.
6086 (avx512f_load<mode>_mask): Ditto.
6087 (avx512f_blendm<mode>): Ditto.
6088 (avx512f_store<mode>_mask): Ditto.
6089 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6090 (avx512f_storedqu<mode>_mask): Ditto.
6091 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6092 Ditto.
6093 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6094 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6095 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6096 (avx512f_maskcmp<mode>3): Ditto.
6097 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6098 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6099 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6100 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6101 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6102 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6103 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6104 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6105 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6106 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6107 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6108 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6109 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6110 (vec_extract_lo_<mode>_maskm): Ditto.
6111 (vec_extract_hi_<mode>_maskm): Ditto.
6112 (avx512f_vternlog<mode>_mask): Ditto.
6113 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6114 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6115 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6116 (avx512f_<code>v8div16qi2_mask): Ditto.
6117 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6118 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6119 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6120 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6121 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6122 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6123 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6124 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6125 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6126 (avx512cd_maskb_vec_dupv8di): Ditto.
6127 (avx512cd_maskw_vec_dupv16si): Ditto.
6128 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6129 (avx512f_vpermi2var<mode>3_mask): Ditto.
6130 (avx512f_vpermi2var<mode>3_mask): Ditto.
6131 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6132 (*avx512f_gathersi<mode>): Ditto.
6133 (*avx512f_gathersi<mode>_2): Ditto.
6134 (*avx512f_gatherdi<mode>): Ditto.
6135 (*avx512f_gatherdi<mode>_2): Ditto.
6136 (*avx512f_scattersi<mode>): Ditto.
6137 (*avx512f_scatterdi<mode>): Ditto.
6138 (avx512f_compress<mode>_mask): Ditto.
6139 (avx512f_compressstore<mode>_mask): Ditto.
6140 (avx512f_expand<mode>_mask): Ditto.
6141 * config/i386/subst.md (mask): Change k to Yk.
6142 (mask_scalar_merge): Ditto.
6143 (sd): Ditto.
6144
6145 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6146
6147 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6148
6149 2014-01-31 Richard Biener <rguenther@suse.de>
6150
6151 PR middle-end/59990
6152 * builtins.c (fold_builtin_memory_op): Make sure to not
6153 use a floating-point mode or a boolean or enumeral type for
6154 the copy operation.
6155
6156 2014-01-30 DJ Delorie <dj@redhat.com>
6157
6158 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6159 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6160 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6161 whenever main() has an epilogue.
6162
6163 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6164
6165 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6166 unused variable "field".
6167 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6168 (vsx_mergeh_<mode>): Likewise.
6169 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6170 (altivec_vmrghh): Likewise.
6171 (altivec_vmrghw): Likewise.
6172 (altivec_vmrglb): Likewise.
6173 (altivec_vmrglh): Likewise.
6174 (altivec_vmrglw): Likewise.
6175 (altivec_vspltb): Add missing uses.
6176 (altivec_vsplth): Likewise.
6177 (altivec_vspltw): Likewise.
6178 (altivec_vspltsf): Likewise.
6179
6180 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6181
6182 PR target/59923
6183 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6184 frame related instructions.
6185
6186 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6187
6188 PR rtl-optimization/59959
6189 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6190 any reload of register whose subreg is invalid.
6191
6192 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6193
6194 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6195 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6196 Add missing return type - void.
6197
6198 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6199
6200 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6201 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6202 remove element index adjustment for endian (now handled in vsx.md
6203 and altivec.md).
6204 (altivec_expand_vec_perm_const): Use
6205 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6206 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6207 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6208 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6209 define_expand and a new define_insn *altivec_vspltb_internal;
6210 adjust for -maltivec=be on a little endian target.
6211 (altivec_vspltb_direct): New.
6212 (altivec_vsplth): Divide into a define_expand and a new
6213 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6214 little endian target.
6215 (altivec_vsplth_direct): New.
6216 (altivec_vspltw): Divide into a define_expand and a new
6217 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6218 little endian target.
6219 (altivec_vspltw_direct): New.
6220 (altivec_vspltsf): Divide into a define_expand and a new
6221 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6222 a little endian target.
6223
6224 2014-01-30 Richard Biener <rguenther@suse.de>
6225
6226 PR tree-optimization/59993
6227 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6228 can propagate form the earlier stmt and avoid the transform
6229 when the intermediate result is needed.
6230
6231 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6232
6233 * README.Portability: Fix typo.
6234
6235 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6236
6237 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6238 comparison_operator with ordered_comparison_operator.
6239
6240 2014-01-30 Nick Clifton <nickc@redhat.com>
6241
6242 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6243 Rename to mn10300_store_multiple_regs.
6244 * config/mn10300/mn10300.c: Likewise.
6245 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6246 store_multiple_regs.
6247 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6248 Call mn10300_store_multiple_regs.
6249
6250 2014-01-30 Nick Clifton <nickc@redhat.com>
6251 DJ Delorie <dj@redhat.com>
6252
6253 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6254 %fp 2 to keep registers after it properly word-aligned.
6255 (rl78_alloc_physical_registers_umul): Handle the case where both
6256 input operands are the same.
6257
6258 2014-01-30 Richard Biener <rguenther@suse.de>
6259
6260 PR tree-optimization/59903
6261 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6262 check properly.
6263
6264 2014-01-30 Jason Merrill <jason@redhat.com>
6265
6266 PR c++/59633
6267 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6268
6269 PR c++/59645
6270 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6271
6272 2014-01-30 Richard Biener <rguenther@suse.de>
6273
6274 PR tree-optimization/59951
6275 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6276
6277 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6278
6279 PR target/59784
6280 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6281 SFmode to DFmode case.
6282
6283 2014-01-29 DJ Delorie <dj@redhat.com>
6284
6285 * config/msp430/msp430.opt (-minrt): New.
6286 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6287 if -minrt given.
6288 (ENDFILE_SPEC): Likewise.
6289
6290 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6291
6292 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6293 (estimate_function_body_sizes): Use it.
6294
6295 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6296
6297 PR c++/58561
6298 * dwarf2out.c (is_cxx_auto): New.
6299 (is_base_type): Use it.
6300 (gen_type_die_with_usage): Likewise.
6301
6302 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6303
6304 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6305 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6306 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6307 -maltivec=be with LE targets.
6308 (vsx_mergeh_<mode>): Likewise.
6309 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6310 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6311 (altivec_vmrghb): Replace with define_expand and new
6312 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6313 (altivec_vmrghb_direct): New define_insn.
6314 (altivec_vmrghh): Replace with define_expand and new
6315 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6316 (altivec_vmrghh_direct): New define_insn.
6317 (altivec_vmrghw): Replace with define_expand and new
6318 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6319 (altivec_vmrghw_direct): New define_insn.
6320 (*altivec_vmrghsf): Adjust for endianness.
6321 (altivec_vmrglb): Replace with define_expand and new
6322 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6323 (altivec_vmrglb_direct): New define_insn.
6324 (altivec_vmrglh): Replace with define_expand and new
6325 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6326 (altivec_vmrglh_direct): New define_insn.
6327 (altivec_vmrglw): Replace with define_expand and new
6328 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6329 (altivec_vmrglw_direct): New define_insn.
6330 (*altivec_vmrglsf): Adjust for endianness.
6331 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6332 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6333 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6334 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6335 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6336 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6337 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6338 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6339
6340 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6341
6342 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6343 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6344 whitespace.
6345
6346 2014-01-29 Richard Biener <rguenther@suse.de>
6347
6348 PR tree-optimization/58742
6349 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6350 associate_pointerplus_align.
6351 (associate_pointerplus_diff): New function.
6352 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6353 and associate_pointerplus_diff.
6354
6355 2014-01-29 Richard Biener <rguenther@suse.de>
6356
6357 * lto-streamer.h (LTO_major_version): Bump to 3.
6358 (LTO_minor_version): Reset to 0.
6359
6360 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6361
6362 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6363 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6364 (arm_file_start): Generate correct asm header for armv7ve.
6365 * config/arm/bpabi.h: Add multilib support for armv7ve.
6366 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6367 and cortex-a15 to armv7ve.
6368 * config/arm/t-aprofile: Add multilib support for armv7ve.
6369 * doc/invoke.texi: Document -march=armv7ve.
6370
6371 2014-01-29 Richard Biener <rguenther@suse.de>
6372
6373 PR tree-optimization/58742
6374 * tree-ssa-forwprop.c (associate_plusminus): Return true
6375 if we changed sth, defer EH cleanup to ...
6376 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6377 (simplify_mult): New function.
6378
6379 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6380
6381 PR middle-end/59917
6382 PR tree-optimization/59920
6383 * tree.c (build_common_builtin_nodes): Remove
6384 __builtin_setjmp_dispatcher initialization.
6385 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6386 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6387 instead of gsi_after_labels + manually skipping debug stmts.
6388 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6389 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6390 * tree-inline.c (copy_edges_for_bb): Remove
6391 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6392 argument. Ignore computed_goto_p stmts. Don't call
6393 make_abnormal_goto_edges. If a call might need abnormal edges
6394 for non-local gotos, see if it already has an edge to
6395 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6396 with true argument, don't do anything then, otherwise add
6397 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6398 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6399 caller.
6400 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6401 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6402 (lower_stmt): Don't set data->calls_builtin_setjmp.
6403 (lower_builtin_setjmp): Adjust comment.
6404 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6405 * tree-cfg.c (found_computed_goto): Remove.
6406 (factor_computed_gotos): Remove.
6407 (make_goto_expr_edges): Return bool, true for computed gotos.
6408 Don't call make_abnormal_goto_edges.
6409 (build_gimple_cfg): Don't set found_computed_goto, don't call
6410 factor_computed_gotos.
6411 (computed_goto_p): No longer static.
6412 (make_blocks): Don't set found_computed_goto.
6413 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6414 (make_edges): If make_goto_expr_edges returns true, push bb
6415 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6416 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6417 vector. Record mapping between bbs and OpenMP regions if there
6418 are any, adjust make_gimple_omp_edges caller. Call
6419 handle_abnormal_edges.
6420 (make_abnormal_goto_edges): Remove.
6421 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6422 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6423 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6424 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6425 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6426 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6427 filling *region also set *region_idx to (*region)->entry->index.
6428
6429 PR other/58712
6430 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6431 For REGs set ORIGINAL_REGNO.
6432
6433 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6434
6435 * doc/md.texi: Mention that a target shouldn't implement
6436 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6437 than hi/lo pair.
6438
6439 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6440
6441 PR tree-optimization/59594
6442 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6443 a copy of the datarefs vector rather than the vector itself.
6444
6445 2014-01-28 Jason Merrill <jason@redhat.com>
6446
6447 PR c++/53756
6448 * dwarf2out.c (auto_die): New static.
6449 (gen_type_die_with_usage): Handle C++1y 'auto'.
6450 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6451 on definition.
6452
6453 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6454
6455 PR target/59672
6456 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6457 (SPEC_X32): Likewise.
6458 (SPEC_64): Likewise.
6459 * config/i386/i386.c (ix86_option_override_internal): Turn off
6460 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6461 for TARGET_16BIT.
6462 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6463 * config/i386/i386.h (TARGET_16BIT): New macro.
6464 (TARGET_16BIT_P): Likewise.
6465 * config/i386/i386.opt: Add m16.
6466 * doc/invoke.texi: Document -m16.
6467
6468 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6469
6470 PR preprocessor/59935
6471 * input.c (location_get_source_line): Bail out on when line number
6472 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6473
6474 2014-01-28 Richard Biener <rguenther@suse.de>
6475
6476 PR tree-optimization/58742
6477 * tree-ssa-forwprop.c (associate_plusminus): Handle
6478 pointer subtraction of the form (T)(P + A) - (T)P.
6479
6480 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6481
6482 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6483 at const_int_cost.
6484
6485 2014-01-28 Richard Biener <rguenther@suse.de>
6486
6487 Revert
6488 2014-01-28 Richard Biener <rguenther@suse.de>
6489
6490 PR rtl-optimization/45364
6491 PR rtl-optimization/59890
6492 * var-tracking.c (local_get_addr_clear_given_value): Handle
6493 already cleared slot.
6494 (val_reset): Handle not allocated local_get_addr_cache.
6495 (vt_find_locations): Use post-order on the inverted CFG.
6496
6497 2014-01-28 Richard Biener <rguenther@suse.de>
6498
6499 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6500
6501 2014-01-28 Richard Biener <rguenther@suse.de>
6502
6503 PR rtl-optimization/45364
6504 PR rtl-optimization/59890
6505 * var-tracking.c (local_get_addr_clear_given_value): Handle
6506 already cleared slot.
6507 (val_reset): Handle not allocated local_get_addr_cache.
6508 (vt_find_locations): Use post-order on the inverted CFG.
6509
6510 2014-01-28 Alan Modra <amodra@gmail.com>
6511
6512 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6513 * configure.ac <recursive call for build != host>: Define
6514 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6515 and LD_FOR_BUILD too.
6516 * configure: Regenerate.
6517
6518 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6519
6520 * config/i386/i386.c (get_builtin_code_for_version): Separate
6521 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6522 Broadwell from Haswell.
6523
6524 2014-01-27 Steve Ellcey <sellcey@mips.com>
6525
6526 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6527 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6528 * config/mips/mips.c (mips_option_override): Change setting
6529 of TARGET_DSP.
6530 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6531 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6532 Change from Mask to Var.
6533
6534 2014-01-27 Jeff Law <law@redhat.com>
6535
6536 * ipa-inline.c (inline_small_functions): Fix typo.
6537
6538 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6539
6540 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6541 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6542 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6543 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6544 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6545 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6546 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6547 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6548 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6549 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6550 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6551 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6552 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6553 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6554 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6555 (_mm512_storeu_epi64): Ditto.
6556 (_mm512_cmpge_epi32_mask): Ditto.
6557 (_mm512_cmpge_epu32_mask): Ditto.
6558 (_mm512_cmpge_epi64_mask): Ditto.
6559 (_mm512_cmpge_epu64_mask): Ditto.
6560 (_mm512_cmple_epi32_mask): Ditto.
6561 (_mm512_cmple_epu32_mask): Ditto.
6562 (_mm512_cmple_epi64_mask): Ditto.
6563 (_mm512_cmple_epu64_mask): Ditto.
6564 (_mm512_cmplt_epi32_mask): Ditto.
6565 (_mm512_cmplt_epu32_mask): Ditto.
6566 (_mm512_cmplt_epi64_mask): Ditto.
6567 (_mm512_cmplt_epu64_mask): Ditto.
6568 (_mm512_cmpneq_epi32_mask): Ditto.
6569 (_mm512_cmpneq_epu32_mask): Ditto.
6570 (_mm512_cmpneq_epi64_mask): Ditto.
6571 (_mm512_cmpneq_epu64_mask): Ditto.
6572 (_mm512_expand_pd): Ditto.
6573 (_mm512_expand_ps): Ditto.
6574 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6575 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6576 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6577 * config/i386/i386.c (ix86_builtins): Add
6578 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6579 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6580 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6581 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6582 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6583 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6584 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6585 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6586 IX86_BUILTIN_PMOVUSQW512_MEM.
6587 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6588 __builtin_ia32_pmovsqd512mem_mask,
6589 __builtin_ia32_pmovqd512mem_mask,
6590 __builtin_ia32_pmovusqw512mem_mask,
6591 __builtin_ia32_pmovsqw512mem_mask,
6592 __builtin_ia32_pmovqw512mem_mask,
6593 __builtin_ia32_pmovusdw512mem_mask,
6594 __builtin_ia32_pmovsdw512mem_mask,
6595 __builtin_ia32_pmovdw512mem_mask,
6596 __builtin_ia32_pmovqb512mem_mask,
6597 __builtin_ia32_pmovusqb512mem_mask,
6598 __builtin_ia32_pmovsqb512mem_mask,
6599 __builtin_ia32_pmovusdb512mem_mask,
6600 __builtin_ia32_pmovsdb512mem_mask,
6601 __builtin_ia32_pmovdb512mem_mask.
6602 (bdesc_args): Add __builtin_ia32_expanddf512,
6603 __builtin_ia32_expandsf512.
6604 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6605 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6606 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6607 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6608 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6609 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6610 (avx512f_<code>v8div16qi2_mask_store): This.
6611 (avx512f_expand<mode>): New.
6612
6613 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6614
6615 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6616 New.
6617 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6618 (_mm512_prefetch_i32scatter_pd): Ditto.
6619 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6620 (_mm512_prefetch_i64scatter_pd): Ditto.
6621 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6622 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6623 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6624 (_mm512_prefetch_i32scatter_ps): Ditto.
6625 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6626 (_mm512_prefetch_i64scatter_ps): Ditto.
6627 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6628 * config/i386/i386-builtin-types.def: Define
6629 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6630 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6631 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6632 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6633 IX86_BUILTIN_SCATTERPFQPD.
6634 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6635 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6636 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6637 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6638 __builtin_ia32_scatterpfqps.
6639 (ix86_expand_builtin): Expand new built-ins.
6640 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6641 fix memory access data type.
6642 (*avx512pf_gatherpf<mode>_mask): Ditto.
6643 (*avx512pf_gatherpf<mode>): Ditto.
6644 (avx512pf_scatterpf<mode>): Ditto.
6645 (*avx512pf_scatterpf<mode>_mask): Ditto.
6646 (*avx512pf_scatterpf<mode>): Ditto.
6647 (GATHER_SCATTER_SF_MEM_MODE): New.
6648 (avx512pf_gatherpf<mode>df): Ditto.
6649 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6650 (*avx512pf_scatterpf<mode>df): Ditto.
6651
6652 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6653
6654 PR bootstrap/59934
6655 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6656 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6657 reached.
6658
6659 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6660
6661 * common/config/arm/arm-common.c
6662 (arm_rewrite_mcpu): Handle multiple names.
6663 * config/arm/arm.h
6664 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6665
6666 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6667
6668 * gimple-builder.h (create_gimple_tmp): Delete.
6669
6670 2014-01-27 Christian Bruel <christian.bruel@st.com>
6671
6672 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6673 words comparisons.
6674
6675 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6676
6677 * config/pa/pa.md (call): Generate indirect long calls to non-local
6678 functions when outputing 32-bit code.
6679 (call_value): Likewise except for special call to buggy powf function.
6680
6681 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6682 portable runtime and PIC indirect calls.
6683 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6684 and PIC call sequences. Use ldo instead of blr to set return register
6685 in PIC call sequence.
6686
6687 2014-01-25 Walter Lee <walt@tilera.com>
6688
6689 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6690 avoid clobbering a live register.
6691
6692 2014-01-25 Walter Lee <walt@tilera.com>
6693
6694 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6695 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6696 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6697 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6698
6699 2014-01-25 Walter Lee <walt@tilera.com>
6700
6701 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6702 arguments on even registers.
6703 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6704 STACK_BOUNDARY.
6705 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6706 (BIGGEST_ALIGNMENT): Ditto.
6707 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6708
6709 2014-01-25 Walter Lee <walt@tilera.com>
6710
6711 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6712 insns before bundling.
6713 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6714
6715 2014-01-25 Walter Lee <walt@tilera.com>
6716
6717 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6718 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6719 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6720
6721 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6722
6723 * config/mips/constraints.md (kl): Delete.
6724 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6725 define expands, using...
6726 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6727 instructions for MIPS16.
6728 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6729 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6730
6731 2014-01-25 Walter Lee <walt@tilera.com>
6732
6733 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6734 (clzdi2): Ditto.
6735 (ffsdi2): Ditto.
6736
6737 2014-01-25 Walter Lee <walt@tilera.com>
6738
6739 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6740 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6741
6742 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6743
6744 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6745 Handle XOR.
6746
6747 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6748
6749 * print-rtl.c (in_call_function_usage): New var.
6750 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6751 EXPR_LIST mode as mode and not as reg note name.
6752
6753 PR middle-end/59561
6754 * cfgloopmanip.c (copy_loop_info): If
6755 loop->warned_aggressive_loop_optimizations, make sure
6756 the flag is set in target loop too.
6757
6758 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6759
6760 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6761 flag_cilkplus.
6762 * builtins.def: Likewise.
6763 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6764 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6765 * ira.c (ira_setup_eliminable_regset): Likewise.
6766 * omp-low.c (gate_expand_omp): Likewise.
6767 (execute_lower_omp): Likewise.
6768 (diagnose_sb_0): Likewise.
6769 (gate_diagnose_omp_blocks): Likewise.
6770 (simd_clone_clauses_extract): Likewise.
6771 (gate): Likewise.
6772
6773 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6774
6775 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6776 correction for little endian...
6777 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6778 here.
6779
6780 2014-01-24 Jeff Law <law@redhat.com>
6781
6782 PR tree-optimization/59919
6783 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6784 for non-returning calls.
6785
6786 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6787
6788 * common/config/aarch64/aarch64-common.c
6789 (aarch64_rewrite_mcpu): Handle multiple names.
6790 * config/aarch64/aarch64.h
6791 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6792
6793 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6794
6795 * input.c (add_file_to_cache_tab): Handle the case where fopen
6796 returns NULL.
6797
6798 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6799
6800 PR target/59929
6801 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6802 from push operand if code of push isn't PRE_DEC.
6803
6804 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6805
6806 PR target/59909
6807 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6808 -mquad-memory-atomic. Update -mquad-memory documentation to say
6809 it is only used for non-atomic loads/stores.
6810
6811 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6812 -mquad-memory or -mquad-memory-atomic switches.
6813
6814 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6815 -mquad-memory-atomic to ISA 2.07 support.
6816
6817 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6818 to separate support of normal quad word memory operations (ldq, stq)
6819 from the atomic quad word memory operations.
6820
6821 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6822 support to separate non-atomic quad word operations from atomic
6823 quad word operations. Disable non-atomic quad word operations in
6824 little endian mode so that we don't have to swap words after the
6825 load and before the store.
6826 (quad_load_store_p): Add comment about atomic quad word support.
6827 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6828 options printed with -mdebug=reg.
6829
6830 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6831 -mquad-memory-atomic as the test for whether we have quad word
6832 atomic instructions.
6833 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6834 or -mp8-vector are used, allow byte/half-word atomic operations.
6835
6836 * config/rs6000/sync.md (load_lockedti): Insure that the address
6837 is a proper indexed or indirect address for the lqarx instruction.
6838 On little endian systems, swap the hi/lo registers after the lqarx
6839 instruction.
6840 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6841 insure the address is valid for the lqarx instruction.
6842 (store_conditionalti): Insure that the address is a proper indexed
6843 or indirect address for the stqcrx. instruction. On little endian
6844 systems, swap the hi/lo registers before doing the stqcrx.
6845 instruction.
6846 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6847 insure the address is valid for the stqcrx. instruction.
6848
6849 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6850 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6851 type of quad memory support is available.
6852
6853 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6854
6855 PR regression/59915
6856 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6857 there is a danger of looping.
6858
6859 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6860
6861 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6862 force flag_ira_loop_pressure if set via command line.
6863
6864 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6865
6866 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6867 (ashr_simd): New builtin handling DI mode.
6868 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6869 (aarch64_sshr_simddi): New match pattern.
6870 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6871 (vshrd_n_s64): Likewise.
6872 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6873
6874 2014-01-23 Nick Clifton <nickc@redhat.com>
6875
6876 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6877 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6878 favour of mcu specific scripts.
6879 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6880 430x multilibs.
6881
6882 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6883 Alex Velenko <Alex.Velenko@arm.com>
6884
6885 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6886 (vaddv_s16): Likewise.
6887 (vaddv_s32): Likewise.
6888 (vaddv_u8): Likewise.
6889 (vaddv_u16): Likewise.
6890 (vaddv_u32): Likewise.
6891 (vaddvq_s8): Likewise.
6892 (vaddvq_s16): Likewise.
6893 (vaddvq_s32): Likewise.
6894 (vaddvq_s64): Likewise.
6895 (vaddvq_u8): Likewise.
6896 (vaddvq_u16): Likewise.
6897 (vaddvq_u32): Likewise.
6898 (vaddvq_u64): Likewise.
6899 (vaddv_f32): Likewise.
6900 (vaddvq_f32): Likewise.
6901 (vaddvq_f64): Likewise.
6902 (vmaxv_f32): Likewise.
6903 (vmaxv_s8): Likewise.
6904 (vmaxv_s16): Likewise.
6905 (vmaxv_s32): Likewise.
6906 (vmaxv_u8): Likewise.
6907 (vmaxv_u16): Likewise.
6908 (vmaxv_u32): Likewise.
6909 (vmaxvq_f32): Likewise.
6910 (vmaxvq_f64): Likewise.
6911 (vmaxvq_s8): Likewise.
6912 (vmaxvq_s16): Likewise.
6913 (vmaxvq_s32): Likewise.
6914 (vmaxvq_u8): Likewise.
6915 (vmaxvq_u16): Likewise.
6916 (vmaxvq_u32): Likewise.
6917 (vmaxnmv_f32): Likewise.
6918 (vmaxnmvq_f32): Likewise.
6919 (vmaxnmvq_f64): Likewise.
6920 (vminv_f32): Likewise.
6921 (vminv_s8): Likewise.
6922 (vminv_s16): Likewise.
6923 (vminv_s32): Likewise.
6924 (vminv_u8): Likewise.
6925 (vminv_u16): Likewise.
6926 (vminv_u32): Likewise.
6927 (vminvq_f32): Likewise.
6928 (vminvq_f64): Likewise.
6929 (vminvq_s8): Likewise.
6930 (vminvq_s16): Likewise.
6931 (vminvq_s32): Likewise.
6932 (vminvq_u8): Likewise.
6933 (vminvq_u16): Likewise.
6934 (vminvq_u32): Likewise.
6935 (vminnmv_f32): Likewise.
6936 (vminnmvq_f32): Likewise.
6937 (vminnmvq_f64): Likewise.
6938
6939 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6940
6941 * config/aarch64/aarch64-simd.md
6942 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
6943 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
6944 (*aarch64_mul3_elt<mode>): Likewise.
6945 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
6946 (*aarch64_mul3_elt_to_64v2df): Likewise.
6947 (*aarch64_mla_elt<mode>): Likewise.
6948 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6949 (*aarch64_mls_elt<mode>): Likewise.
6950 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6951 (*aarch64_fma4_elt<mode>): Likewise.
6952 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6953 (*aarch64_fma4_elt_to_64v2df): Likewise.
6954 (*aarch64_fnma4_elt<mode>): Likewise.
6955 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6956 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6957 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6958 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6959 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6960 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6961 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6962 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6963 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6964
6965 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
6966
6967 * config/aarch64/aarch64-simd.md
6968 (aarch64_be_checked_get_lane<mode>): New define_expand.
6969 * config/aarch64/aarch64-simd-builtins.def
6970 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
6971 New builtin definition.
6972 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
6973 Use new safe be builtin.
6974
6975 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6976
6977 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
6978 New define_insn.
6979 (aarch64_be_st1<mode>): Likewise.
6980 (aarch_ld1<VALL:mode>): Define_expand modified.
6981 (aarch_st1<VALL:mode>): Likewise.
6982 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
6983 (UNSPEC_ST1): Likewise.
6984
6985 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
6986
6987 * config/microblaze/microblaze.md: Add trap insn and attribute
6988
6989 2014-01-23 Dodji Seketeli <dodji@redhat.com>
6990
6991 PR preprocessor/58580
6992 * input.h (location_get_source_line): Take an additional line_size
6993 parameter.
6994 (void diagnostics_file_cache_fini): Declare new function.
6995 * input.c (struct fcache): New type.
6996 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
6997 New static constants.
6998 (diagnostic_file_cache_init, total_lines_num)
6999 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7000 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7001 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7002 (get_next_line, read_next_line, goto_next_line, read_line_num):
7003 New static function definitions.
7004 (diagnostic_file_cache_fini): New function.
7005 (location_get_source_line): Take an additional output line_len
7006 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7007 read_line_num.
7008 * diagnostic.c (diagnostic_finish): Call
7009 diagnostic_file_cache_fini.
7010 (adjust_line): Take an additional input parameter for the length
7011 of the line, rather than calculating it with strlen.
7012 (diagnostic_show_locus): Adjust the use of
7013 location_get_source_line and adjust_line with respect to their new
7014 signature. While displaying a line now, do not stop at the first
7015 null byte. Rather, display the zero byte as a space and keep
7016 going until we reach the size of the line.
7017 * Makefile.in: Add vec.o to OBJS-libcommon
7018
7019 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7020 Ilya Tocar <ilya.tocar@intel.com>
7021
7022 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7023 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7024 (__builtin_ia32_kmov16): Ditto.
7025 * config/i386/i386.md (UNSPEC_KMOV): New.
7026 (kmovw): Ditto.
7027
7028 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7029
7030 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7031 (_mm512_storeu_si512): Ditto.
7032
7033 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7034
7035 PR target/52125
7036 * rtl.h (get_referenced_operands): Declare.
7037 * recog.c (get_referenced_operands): New function.
7038 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7039 operands have been referenced when recording LO_SUM references.
7040
7041 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7042
7043 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7044
7045 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7046
7047 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7048 Enable for generic and recent AMD targets.
7049
7050 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7051
7052 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7053 ARG_SIZE note when adjustment was eliminated.
7054
7055 2014-01-22 Jeff Law <law@redhat.com>
7056
7057 PR tree-optimization/59597
7058 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7059 in file. Accept new argument REGISTERING and use it to modify
7060 dump output appropriately.
7061 (register_jump_thread): Corresponding changes.
7062 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7063 thread paths involving joiner blocks. Add code to dump cancelled
7064 jump threading paths.
7065
7066 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7067
7068 PR rtl-optimization/59477
7069 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7070 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7071
7072 2014-01-22 Tom Tromey <tromey@redhat.com>
7073
7074 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7075 PARAMS.
7076 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7077
7078 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7079
7080 PR rtl-optimization/59896
7081 * lra-constraints.c (process_alt_operands): Check unused note for
7082 matched operands of insn with no output reloads.
7083
7084 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7085
7086 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7087 (mips_move_from_gpr_cost): Likewise.
7088
7089 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7090
7091 PR rtl-optimization/59858
7092 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7093 ira_class_hard_regs_num.
7094 (process_alt_operands): Increase reject for dying matched operand.
7095
7096 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7097
7098 PR target/59003
7099 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7100 smaller than size, perform several stores or loads and stores
7101 at dst + count - size to store or copy all of size bytes, rather
7102 than just last modesize bytes.
7103
7104 2014-01-20 DJ Delorie <dj@redhat.com>
7105
7106 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7107 that CLOBBERs are REGs before propogating their values.
7108
7109 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7110
7111 PR middle-end/59789
7112 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7113 (cgraph_inline_failed_type): New function.
7114 * cgraph.h (DEFCIFCODE): Add type.
7115 (cgraph_inline_failed_type_t): New enum.
7116 (cgraph_inline_failed_type): New prototype.
7117 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7118 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7119 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7120 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7121 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7122 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7123 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7124 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7125 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7126 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7127 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7128 OPTIMIZATION_MISMATCH.
7129 * tree-inline.c (expand_call_inline): Emit errors during
7130 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7131
7132 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7133
7134 PR target/59685
7135 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7136 mode attribute in insn output.
7137
7138 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7139
7140 * output.h (output_constant): Delete.
7141 * varasm.c (output_constant): Make private.
7142
7143 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7144
7145 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7146
7147 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7148
7149 PR middle-end/59860
7150 * tree.h (fold_builtin_strcat): New prototype.
7151 * builtins.c (fold_builtin_strcat): No longer static. Add len
7152 argument, if non-NULL, don't call c_strlen. Optimize
7153 directly into __builtin_memcpy instead of __builtin_strcpy.
7154 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7155 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7156
7157 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7158
7159 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7160 for SImode_address_operand operands, having only a REG argument.
7161
7162 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7163
7164 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7165 loader name using mbig-endian.
7166 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7167
7168 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7169
7170 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7171 (-mtune): Likewise.
7172 (-mcpu): Likewise.
7173
7174 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7175
7176 * config/aarch64/aarch64-protos.h
7177 (aarch64_cannot_change_mode_class_ptr): Declare.
7178 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7179 aarch64_cannot_change_mode_class_ptr): New.
7180 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7181 backend hook aarch64_cannot_change_mode_class.
7182
7183 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7184
7185 * common/config/aarch64/aarch64-common.c
7186 (aarch64_handle_option): Don't handle any option order logic here.
7187 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7188 selected_cpu, warn on architecture version mismatch.
7189 (aarch64_override_options): Fix parsing order for option strings.
7190
7191 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7192 Iain Sandoe <iain@codesourcery.com>
7193
7194 PR bootstrap/59496
7195 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7196 warning. Amend comment to reflect current functionality.
7197
7198 2014-01-20 Richard Biener <rguenther@suse.de>
7199
7200 PR middle-end/59860
7201 * builtins.c (fold_builtin_strcat): Remove case better handled
7202 by tree-ssa-strlen.c.
7203
7204 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7205
7206 * config/aarch64/aarch64.opt
7207 (mcpu, march, mtune): Make case-insensitive.
7208
7209 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7210
7211 PR target/59880
7212 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7213 if operands[1] is a REG or ZERO_EXTEND of a REG.
7214
7215 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7216
7217 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7218
7219 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7220
7221 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7222 long non-pic millicode calls.
7223
7224 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7225
7226 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7227
7228 2014-01-19 Kito Cheng <kito@0xlab.org>
7229
7230 * builtins.c (expand_movstr): Check movstr expand done or fail.
7231
7232 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7233 H.J. Lu <hongjiu.lu@intel.com>
7234
7235 PR target/59379
7236 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7237 to DImode for zero-extended addresses.
7238
7239 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7240
7241 PR rtl-optimization/57763
7242 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7243 on the new indirect jump_insn and increment LABEL_NUSES (label).
7244
7245 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7246
7247 PR bootstrap/59580
7248 PR bootstrap/59583
7249 * config.gcc (x86_archs): New variable.
7250 (x86_64_archs): Likewise.
7251 (x86_cpus): Likewise.
7252 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7253 --with-arch/--with-cpu= options.
7254 Support --with-arch=/--with-cpu={nehalem,westmere,
7255 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7256
7257 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7258
7259 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7260 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7261
7262 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7263
7264 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7265
7266 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7267
7268 PR target/58944
7269 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7270 clear cpp_get_options (parse_in)->warn_unused_macros for
7271 ix86_target_macros_internal with cpp_define.
7272
7273 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7274
7275 * jump.c (delete_related_insns): Keep (use (insn))s.
7276 * reorg.c (redundant_insn): Check for barriers too.
7277
7278 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7279
7280 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7281
7282 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7283
7284 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7285 call to $$dyncall when TARGET_LONG_CALLS is true.
7286
7287 2014-01-17 Jeff Law <law@redhat.com>
7288
7289 * ree.c (combine_set_extension): Temporarily disable test for
7290 changing number of hard registers.
7291
7292 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7293
7294 PR middle-end/58125
7295 * ipa-inline-analysis.c (inline_free_summary):
7296 Do not free summary of aliases.
7297
7298 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7299
7300 PR middle-end/59706
7301 * gimplify.c (gimplify_expr): Use create_tmp_var
7302 instead of create_tmp_var_raw. If cond doesn't have
7303 integral type, don't add the IFN_ANNOTATE builtin at all.
7304
7305 2014-01-17 Martin Jambor <mjambor@suse.cz>
7306
7307 PR ipa/59736
7308 * ipa-cp.c (prev_edge_clone): New variable.
7309 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7310 Also resize prev_edge_clone vector.
7311 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7312 (ipcp_edge_removal_hook): New function.
7313 (ipcp_driver): Register ipcp_edge_removal_hook.
7314
7315 2014-01-17 Andrew Pinski <apinski@cavium.com>
7316 Steve Ellcey <sellcey@mips.com>
7317
7318 PR target/59462
7319 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7320 of operator mode.
7321
7322 2014-01-17 Jeff Law <law@redhat.com>
7323
7324 PR middle-end/57904
7325 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7326 so that pass_ccp runs first.
7327
7328 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7329
7330 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7331 (ix86_adjust_cost): Use !TARGET_XXX.
7332 (do_reorder_for_imul): Likewise.
7333 (swap_top_of_ready_list): Likewise.
7334 (ix86_sched_reorder): Likewise.
7335
7336 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7337
7338 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7339 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7340 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7341 (intel_memset): New. Duplicate slm_memset.
7342 (intel_cost): New. Duplicate slm_cost.
7343 (m_INTEL): New macro.
7344 (processor_target_table): Add "intel".
7345 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7346 with PROCESSOR_INTEL for "intel".
7347 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7348 PROCESSOR_SILVERMONT.
7349 (ix86_issue_rate): Likewise.
7350 (ix86_adjust_cost): Likewise.
7351 (ia32_multipass_dfa_lookahead): Likewise.
7352 (swap_top_of_ready_list): Likewise.
7353 (ix86_sched_reorder): Likewise.
7354 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7355 instead of TARGET_OPT_AGU.
7356 * config/i386/i386.h (TARGET_INTEL): New.
7357 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7358 (processor_type): Add PROCESSOR_INTEL.
7359 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7360 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7361
7362 2014-01-17 Marek Polacek <polacek@redhat.com>
7363
7364 PR c/58346
7365 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7366 size is zero.
7367
7368 2014-01-17 Richard Biener <rguenther@suse.de>
7369
7370 PR tree-optimization/46590
7371 * opts.c (default_options_table): Add entries for
7372 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7373 all enabled at -O1 but not for -Og.
7374 * common.opt (fbranch-count-reg): Remove Init(1).
7375 (fmove-loop-invariants): Likewise.
7376 (ftree-pta): Likewise.
7377
7378 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7379
7380 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7381 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7382 decls to at least the GCC 4.8 used alignments.
7383
7384 PR fortran/59440
7385 * tree-nested.c (convert_nonlocal_reference_stmt,
7386 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7387 of GIMPLE_BIND stmts, adjust associated decls.
7388
7389 2014-01-17 Richard Biener <rguenther@suse.de>
7390
7391 PR tree-optimization/46590
7392 * vec.h (vec<>::bseach): New member function implementing
7393 binary search according to C89 bsearch.
7394 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7395 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7396 bitmap pointer again. Make accesses_in_loop a flat array.
7397 (mem_ref_obstack): New global.
7398 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7399 (mark_ref_stored): Likewise.
7400 (ref_indep_loop_p_2): Likewise.
7401 (set_ref_stored_in_loop): New helper function.
7402 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7403 (memref_free): Adjust.
7404 (record_mem_ref_loc): Simplify.
7405 (gather_mem_refs_stmt): Adjust.
7406 (sort_locs_in_loop_postorder_cmp): New function.
7407 (analyze_memory_references): Sort accesses_in_loop after
7408 loop postorder number.
7409 (find_ref_loc_in_loop_cmp): New function.
7410 (for_all_locs_in_loop): Find relevant cluster of locs in
7411 accesses_in_loop and iterate without recursion.
7412 (execute_sm): Avoid uninit warning.
7413 (struct ref_always_accessed): Simplify.
7414 (ref_always_accessed::operator ()): Likewise.
7415 (ref_always_accessed_p): Likewise.
7416 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7417 loop postorder numbers here.
7418 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7419 numbers.
7420
7421 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7422
7423 PR c++/57945
7424 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7425 on decls for which assemble_alias has been called.
7426
7427 2014-01-17 Nick Clifton <nickc@redhat.com>
7428
7429 * config/msp430/msp430.opt: (mcpu): New option.
7430 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7431 (msp430_option_override): Parse target_cpu. If the MCU name
7432 matches a generic string, clear target_mcu.
7433 (msp430_attr): Allow numeric interrupt values up to 63.
7434 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7435 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7436 option.
7437 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7438 Add mcpu matches.
7439 * config/msp430/msp430.md (popm): Use %J rather than %I.
7440 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7441 (addhi_cy_i): Use immediate_operand for operand 2.
7442 * doc/invoke.texi: Document -mcpu option.
7443
7444 2014-01-17 Richard Biener <rguenther@suse.de>
7445
7446 PR rtl-optimization/38518
7447 * df.h (df_analyze_loop): Declare.
7448 * df-core.c: Include cfgloop.h.
7449 (df_analyze_1): Split out main part of df_analyze.
7450 (df_analyze): Adjust.
7451 (loop_inverted_post_order_compute): New function.
7452 (loop_post_order_compute): Likewise.
7453 (df_analyze_loop): New function avoiding whole-function
7454 postorder computes.
7455 * loop-invariant.c (find_defs): Use df_analyze_loop.
7456 (find_invariants): Adjust.
7457 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7458
7459 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7460
7461 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7462 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7463
7464 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7465
7466 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7467 traversal when removing references.
7468
7469 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7470
7471 PR ipa/59775
7472 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7473
7474 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7475
7476 PR middle-end/56791
7477 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7478 pushing a reload for an autoinc when we had previously reloaded an
7479 inner part of the address.
7480
7481 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7482
7483 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7484 field.
7485 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7486 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7487 when not giving up or versioning for alias only because of
7488 loop->safelen.
7489 (vect_analyze_data_ref_dependences): Set to true.
7490 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7491 is a GIMPLE_PHI.
7492 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7493 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7494 to the condition.
7495
7496 PR middle-end/58344
7497 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7498
7499 PR target/59839
7500 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7501 operand 0 predicate for gathers, use a new pseudo as subtarget.
7502
7503 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7504
7505 PR middle-end/59609
7506 * lra-constraints.c (process_alt_operands): Add printing debug info.
7507 Check absence of input/output reloads for matched operands too.
7508
7509 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7510
7511 PR rtl-optimization/59835
7512 * ira.c (ira_init_register_move_cost): Increase cost for
7513 impossible modes.
7514
7515 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7516
7517 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7518
7519 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7520
7521 PR target/59780
7522 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7523 non-register objects. Use gen_(high/low)part more consistently.
7524 Fix assertions.
7525
7526 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7527
7528 PR target/59844
7529 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7530 endian support, remove tests for WORDS_BIG_ENDIAN.
7531 (p8_mfvsrd_3_<mode>): Likewise.
7532 (reload_gpr_from_vsx<mode>): Likewise.
7533 (reload_gpr_from_vsxsf): Likewise.
7534 (p8_mfvsrd_4_disf): Likewise.
7535
7536 2014-01-16 Richard Biener <rguenther@suse.de>
7537
7538 PR rtl-optimization/46590
7539 * lcm.c (compute_antinout_edge): Use postorder iteration.
7540 (compute_laterin): Use inverted postorder iteration.
7541
7542 2014-01-16 Nick Clifton <nickc@redhat.com>
7543
7544 PR middle-end/28865
7545 * varasm.c (output_constant): Return the number of bytes actually
7546 emitted.
7547 (output_constructor_array_range): Update the field size with the
7548 number of bytes emitted by output_constant.
7549 (output_constructor_regular_field): Likewise. Also do not
7550 complain if the total number of bytes emitted is now greater
7551 than the expected fieldpos.
7552 * output.h (output_constant): Update prototype and descriptive comment.
7553
7554 2014-01-16 Marek Polacek <polacek@redhat.com>
7555
7556 PR middle-end/59827
7557 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7558 it is error_mark_node.
7559
7560 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7561
7562 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7563 VALID_AVX256_REG_OR_OI_MODE.
7564
7565 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7566
7567 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7568 current procedure should be profiled.
7569
7570 2014-01-15 Andrew Pinski <apinski@cavium.com>
7571
7572 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7573 of moving from/to the STACK_REG register class.
7574
7575 2014-01-15 Richard Henderson <rth@redhat.com>
7576
7577 PR debug/54694
7578 * reginfo.c (global_regs_decl): Globalize.
7579 * rtl.h (global_regs_decl): Declare.
7580 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7581 reserved via global_regs.
7582
7583 2014-01-15 Teresa Johnson <tejohnson@google.com>
7584
7585 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7586
7587 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7588
7589 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7590 and vmulosh rather than call gen_vec_widen_smult_*.
7591 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7592 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7593 (vec_widen_smult_even_v16qi): Likewise.
7594 (vec_widen_umult_even_v8hi): Likewise.
7595 (vec_widen_smult_even_v8hi): Likewise.
7596 (vec_widen_umult_odd_v16qi): Likewise.
7597 (vec_widen_smult_odd_v16qi): Likewise.
7598 (vec_widen_umult_odd_v8hi): Likewise.
7599 (vec_widen_smult_odd_v8hi): Likewise.
7600 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7601 vmuloub rather than call gen_vec_widen_umult_*.
7602 (vec_widen_umult_lo_v16qi): Likewise.
7603 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7604 vmulosb rather than call gen_vec_widen_smult_*.
7605 (vec_widen_smult_lo_v16qi): Likewise.
7606 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7607 rather than call gen_vec_widen_umult_*.
7608 (vec_widen_umult_lo_v8hi): Likewise.
7609 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7610 rather than call gen_vec_widen_smult_*.
7611 (vec_widen_smult_lo_v8hi): Likewise.
7612
7613 2014-01-15 Jeff Law <law@redhat.com>
7614
7615 PR tree-optimization/59747
7616 * ree.c (find_and_remove_re): Properly handle case where a second
7617 eliminated extension requires widening a copy created for elimination
7618 of a prior extension.
7619 (combine_set_extension): Ensure that the number of hard regs needed
7620 for a destination register does not change when we widen it.
7621
7622 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7623
7624 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7625 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7626 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7627 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7628 (avr-*-rtems*): Likewise.
7629 (bfin*-rtems*): Likewise.
7630 (moxie-*-rtems*): Likewise.
7631 (h8300-*-rtems*): Likewise.
7632 (i[34567]86-*-rtems*): Likewise.
7633 (lm32-*-rtems*): Likewise.
7634 (m32r-*-rtems*): Likewise.
7635 (m68k-*-rtems*): Likewise.
7636 (microblaze*-*-rtems*): Likewise.
7637 (mips*-*-rtems*): Likewise.
7638 (powerpc-*-rtems*): Likewise.
7639 (sh-*-rtems*): Likewise.
7640 (sparc-*-rtems*): Likewise.
7641 (sparc64-*-rtems*): Likewise.
7642 (v850-*-rtems*): Likewise.
7643 (m32c-*-rtems*): Likewise.
7644
7645 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7646
7647 PR rtl-optimization/59511
7648 * ira.c (ira_init_register_move_cost): Use memory costs for some
7649 cases of register move cost calculations.
7650 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7651 instead of BB frequency.
7652 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7653 * lra-assigns.c (find_hard_regno_for): Ditto.
7654
7655 2014-01-15 Richard Biener <rguenther@suse.de>
7656
7657 PR tree-optimization/59822
7658 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7659 (vectorizable_load): Use it to hoist defs of uses of invariant
7660 loads out of the loop.
7661
7662 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7663 Kugan Vivekanandarajah <kuganv@linaro.org>
7664
7665 PR target/59695
7666 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7667 truncation.
7668
7669 2014-01-15 Richard Biener <rguenther@suse.de>
7670
7671 PR rtl-optimization/59802
7672 * lcm.c (compute_available): Use inverted postorder to seed
7673 the initial worklist.
7674
7675 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7676
7677 PR target/59803
7678 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7679 ADDR_REGS for invalid symrefs in non-PIC code.
7680
7681 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7682
7683 PR other/58712
7684 * builtins.c (determine_block_size): Initialize *probable_max_size
7685 even if len_rtx is CONST_INT.
7686
7687 2014-01-14 Andrew Pinski <apinski@cavium.com>
7688
7689 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7690 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7691 (cortexa53_tunings): Likewise.
7692 (aarch64_sched_issue_rate): New function.
7693 (TARGET_SCHED_ISSUE_RATE): Define.
7694
7695 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7696
7697 * ira-costs.c (find_costs_and_classes): Add missed
7698 ira_init_register_move_cost_if_necessary.
7699
7700 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7701
7702 PR target/59787
7703 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7704
7705 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7706
7707 PR target/59794
7708 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7709 to indicate if type is used for function return value. Warn ABI
7710 change if the vector mode isn't available for function return value.
7711 (ix86_function_arg_advance): Pass false to type_natural_mode.
7712 (ix86_function_arg): Likewise.
7713 (ix86_gimplify_va_arg): Likewise.
7714 (function_arg_32): Don't warn ABI change.
7715 (ix86_function_value): Pass true to type_natural_mode.
7716 (ix86_return_in_memory): Likewise.
7717 (ix86_struct_value_rtx): Removed.
7718 (TARGET_STRUCT_VALUE_RTX): Likewise.
7719
7720 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7721
7722 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7723 converting a conditional jump into a conditional return.
7724
7725 2014-01-14 Richard Biener <rguenther@suse.de>
7726
7727 PR tree-optimization/58921
7728 PR tree-optimization/59006
7729 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7730 hoisting invariant stmts.
7731 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7732 invariant loads on the preheader edge if possible.
7733
7734 2014-01-14 Joey Ye <joey.ye@arm.com>
7735
7736 * doc/plugin.texi (Building GCC plugins): Update to C++.
7737
7738 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7739
7740 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7741 (_mm_rcp28_round_ss): Ditto.
7742 (_mm_rsqrt28_round_sd): Ditto.
7743 (_mm_rsqrt28_round_ss): Ditto.
7744 (_mm_rcp28_sd): Ditto.
7745 (_mm_rcp28_ss): Ditto.
7746 (_mm_rsqrt28_sd): Ditto.
7747 (_mm_rsqrt28_ss): Ditto.
7748 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7749 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7750 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7751 (IX86_BUILTIN_RCP28SD): Ditto.
7752 (IX86_BUILTIN_RCP28SS): Ditto.
7753 (IX86_BUILTIN_RSQRT28SD): Ditto.
7754 (IX86_BUILTIN_RSQRT28SS): Ditto.
7755 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7756 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7757 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7758 (ix86_expand_special_args_builtin): Expand new FTYPE.
7759 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7760 (srcp14<mode>): Make insn unary.
7761 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7762 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7763 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7764 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7765 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7766 Fix rounding: make it SAE only.
7767 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7768 Ditto.
7769 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7770 Ditto.
7771 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7772 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7773 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7774 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7775 (round_saeonly_mask_scalar_operand4): Ditto.
7776 (round_saeonly_mask_scalar_op3): Ditto.
7777 (round_saeonly_mask_scalar_op4): Ditto.
7778
7779 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7780
7781 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7782 Implement -maltivec=be for vec_insert and vec_extract.
7783
7784 2014-01-10 DJ Delorie <dj@redhat.com>
7785
7786 * config/msp430/msp430.md (call_internal): Don't allow memory
7787 references with SP as the base register.
7788 (call_value_internal): Likewise.
7789 * config/msp430/constraints.md (Yc): New. For memory references
7790 that don't use SP as a base register.
7791
7792 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7793 "an integer without a # prefix"
7794 * config/msp430/msp430.md (epilogue_helper): Use it.
7795
7796 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7797
7798 PR target/59617
7799 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7800 AVX512F gather builtins.
7801 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7802 on gather decls with INTEGER_TYPE masktype.
7803 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7804 directly into the builtin rather than hoisting it before loop.
7805
7806 PR tree-optimization/59387
7807 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7808 (scev_const_prop): If folded_casts and type has undefined overflow,
7809 use force_gimple_operand instead of force_gimple_operand_gsi and
7810 for each added stmt if it is assign with
7811 arith_code_with_undefined_signed_overflow, call
7812 rewrite_to_defined_overflow.
7813 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7814 gimple-fold.h instead.
7815 (arith_code_with_undefined_signed_overflow,
7816 rewrite_to_defined_overflow): Moved to ...
7817 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7818 rewrite_to_defined_overflow): ... here. No longer static.
7819 Include gimplify-me.h.
7820 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7821 rewrite_to_defined_overflow): New prototypes.
7822
7823 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7824
7825 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7826
7827 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7828
7829 * builtins.c (get_object_alignment_2): Minor tweak.
7830 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7831
7832 2014-01-13 Christian Bruel <christian.bruel@st.com>
7833
7834 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7835 optimized non constant lengths.
7836
7837 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7838
7839 PR libgomp/59194
7840 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7841 load as __atomic_load_N if possible.
7842
7843 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7844
7845 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7846 target parameter.
7847 (rs6000_expand_builtin): Adjust call.
7848
7849 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7850
7851 PR target/58115
7852 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7853 * config/rs6000/rs6000.c: Include target-globals.h.
7854 (rs6000_set_current_function): Instead of doing target_reinit
7855 unconditionally, use save_target_globals_default_opts and
7856 restore_target_globals.
7857
7858 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7859 FPSCR.
7860 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7861 (rs6000_expand_builtin): Handle mffs and mtfsf.
7862 (rs6000_init_builtins): Define mffs and mtfsf.
7863 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7864 (rs6000_mffs): New pattern.
7865 (rs6000_mtfsf): New pattern.
7866
7867 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7868
7869 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7870 Start narrowing with START. Apply candidate-use pair
7871 and check overall cost in narrowing.
7872 (iv_ca_prune): Pass new argument.
7873
7874 2014-01-10 Jeff Law <law@redhat.com>
7875
7876 PR middle-end/59743
7877 * ree.c (combine_reaching_defs): Ensure the defining statement
7878 occurs before the extension when optimizing extensions with
7879 different source and destination hard registers.
7880
7881 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7882
7883 PR ipa/58585
7884 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7885 vtables into the type inheritance graph.
7886
7887 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7888
7889 PR rtl-optimization/59754
7890 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7891 modes in the REGNO != REGNO case.
7892
7893 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7894
7895 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7896
7897 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7898
7899 PR tree-optimization/59745
7900 * tree-predcom.c (tree_predictive_commoning_loop): Call
7901 free_affine_expand_cache if giving up because components is NULL.
7902
7903 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7904 GC in payload of target_globals struct instead of allocating them on
7905 the heap and the larger structs separately using GC.
7906 * target-globals.h (struct target_globals): Make regs, hard_regs,
7907 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7908 of GTY((skip)) and change type to void *.
7909 (reset_target_globals): Cast loads from those fields to corresponding
7910 types.
7911
7912 2014-01-10 Steve Ellcey <sellcey@mips.com>
7913
7914 PR plugins/59335
7915 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7916 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7917 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7918
7919 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7920
7921 PR target/59744
7922 * aarch64-modes.def (CC_Zmode): New flags mode.
7923 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7924 represents an equality.
7925 (aarch64_get_condition_code): Handle CC_Zmode.
7926 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
7927
7928 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7929
7930 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
7931 extraction in good case.
7932
7933 2014-01-10 Richard Biener <rguenther@suse.de>
7934
7935 PR tree-optimization/59374
7936 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
7937 checking after SLP discovery. Mark stmts not participating
7938 in any SLP instance properly.
7939
7940 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7941
7942 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
7943 when handling a SET rtx.
7944
7945 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7946
7947 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
7948 (cortex-a57): Likewise.
7949 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
7950
7951 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7952
7953 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
7954 non-iwmmxt builtins.
7955
7956 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7957
7958 PR ipa/58252
7959 PR ipa/59226
7960 * ipa-devirt.c record_target_from_binfo): Take as argument
7961 stack of binfos and lookup matching one for virtual inheritance.
7962 (possible_polymorphic_call_targets_1): Update.
7963
7964 2014-01-10 Huacai Chen <chenhc@lemote.com>
7965
7966 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
7967 kernel strings for Loongson-2E/2F/3A.
7968
7969 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7970
7971 PR middle-end/59670
7972 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
7973 is_gimple_call before calling gimple_call_internal_p.
7974
7975 2014-01-09 Steve Ellcey <sellcey@mips.com>
7976
7977 * Makefile.in (TREE_FLOW_H): Remove.
7978 (TREE_SSA_H): Add file names from tree-flow.h.
7979 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
7980 * tree.h: Remove tree-flow.h reference.
7981 * hash-table.h: Remove tree-flow.h reference.
7982 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
7983 reference with tree-ssa-loop.h.
7984
7985 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7986
7987 * doc/invoke.texi: Add -maltivec={be,le} options, and document
7988 default element-order behavior for -maltivec.
7989 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
7990 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
7991 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
7992 when targeting big endian, at least for now.
7993 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
7994
7995 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7996
7997 PR middle-end/47735
7998 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
7999 var satisfies use_register_for_decl, just take into account type
8000 alignment, rather than decl alignment.
8001
8002 PR tree-optimization/59622
8003 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8004 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8005 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8006 Don't devirtualize for inplace at all. For targets.length () == 1,
8007 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8008
8009 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8010
8011 * config/i386/i386.md (cpu): Remove the unused btver1.
8012
8013 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8014
8015 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8016
8017 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8018
8019 PR target/58115
8020 * tree-core.h (struct target_globals): New forward declaration.
8021 (struct tree_target_option): Add globals field.
8022 * tree.h (TREE_TARGET_GLOBALS): Define.
8023 (prepare_target_option_nodes_for_pch): New prototype.
8024 * target-globals.h (struct target_globals): Define even if
8025 !SWITCHABLE_TARGET.
8026 * tree.c (prepare_target_option_node_for_pch,
8027 prepare_target_option_nodes_for_pch): New functions.
8028 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8029 * config/i386/i386.c: Include target-globals.h.
8030 (ix86_set_current_function): Instead of doing target_reinit
8031 unconditionally, use save_target_globals_default_opts and
8032 restore_target_globals.
8033
8034 2014-01-09 Richard Biener <rguenther@suse.de>
8035
8036 PR tree-optimization/59715
8037 * tree-cfg.h (split_critical_edges): Declare.
8038 * tree-cfg.c (split_critical_edges): Export.
8039 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8040
8041 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8042
8043 * cfgexpand.c (expand_stack_vars): Optionally disable
8044 asan stack protection.
8045 (expand_used_vars): Likewise.
8046 (partition_stack_vars): Likewise.
8047 * asan.c (asan_emit_stack_protection): Optionally disable
8048 after return stack usage.
8049 (instrument_derefs): Optionally disable memory access instrumentation.
8050 (instrument_builtin_call): Likewise.
8051 (instrument_strlen_call): Likewise.
8052 (asan_protect_global): Optionally disable global variables protection.
8053 * doc/invoke.texi: Added doc for new options.
8054 * params.def: Added new options.
8055 * params.h: Likewise.
8056
8057 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8058
8059 PR rtl-optimization/59724
8060 * ifcvt.c (cond_exec_process_if_block): Don't call
8061 flow_find_head_matching_sequence with 0 longest_match.
8062 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8063 non-active insns if !stop_after.
8064 (try_head_merge_bb): Revert 2014-01-07 changes.
8065
8066 2014-01-08 Jeff Law <law@redhat.com>
8067
8068 * ree.c (get_sub_rtx): New function, extracted from...
8069 (merge_def_and_ext): Here.
8070 (combine_reaching_defs): Use get_sub_rtx.
8071
8072 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8073
8074 * cgraph.h (varpool_variable_node): Do not choke on null node.
8075
8076 2014-01-08 Catherine Moore <clm@codesourcery.com>
8077
8078 * config/mips/mips.md (simple_return): Attempt to use JRC
8079 for microMIPS.
8080 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8081
8082 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8083
8084 PR rtl-optimization/59137
8085 * reorg.c (steal_delay_list_from_target): Call update_block for
8086 elided insns.
8087 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8088
8089 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8090
8091 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8092 two duplicate entries.
8093
8094 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8095
8096 Revert:
8097 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8098
8099 * config/mips/mips.c (mips_truncated_op_cost): New function.
8100 (mips_rtx_costs): Adjust test for BADDU.
8101 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8102
8103 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8104
8105 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8106 (*baddu_si): ...this new pattern.
8107
8108 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8109
8110 PR ipa/59722
8111 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8112
8113 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8114
8115 PR middle-end/57748
8116 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8117 inner_reference_p.
8118 (expand_expr, expand_normal): Adjust.
8119 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8120 inner_reference_p. Use inner_reference_p to expand inner references.
8121 (store_expr): Adjust.
8122 * cfgexpand.c (expand_call_stmt): Adjust.
8123
8124 2014-01-08 Rong Xu <xur@google.com>
8125
8126 * gcov-io.c (gcov_var): Move from gcov-io.h.
8127 (gcov_position): Ditto.
8128 (gcov_is_error): Ditto.
8129 (gcov_rewrite): Ditto.
8130 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8131 only part to libgcc/libgcov.h.
8132
8133 2014-01-08 Marek Polacek <polacek@redhat.com>
8134
8135 PR middle-end/59669
8136 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8137
8138 2014-01-08 Marek Polacek <polacek@redhat.com>
8139
8140 PR sanitizer/59667
8141 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8142
8143 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8144
8145 PR rtl-optimization/59649
8146 * stor-layout.c (get_mode_bounds): For BImode return
8147 0 and STORE_FLAG_VALUE.
8148
8149 2014-01-08 Richard Biener <rguenther@suse.de>
8150
8151 PR middle-end/59630
8152 * gimple.h (is_gimple_builtin_call): Remove.
8153 (gimple_builtin_call_types_compatible_p): New.
8154 (gimple_call_builtin_p): New overload.
8155 * gimple.c (is_gimple_builtin_call): Remove.
8156 (validate_call): Rename to ...
8157 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8158 check return types.
8159 (validate_type): New static function.
8160 (gimple_call_builtin_p): New overload and adjust.
8161 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8162 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8163 (gimple_fold_stmt_to_constant_1): Likewise.
8164 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8165
8166 2014-01-08 Richard Biener <rguenther@suse.de>
8167
8168 PR middle-end/59471
8169 * gimplify.c (gimplify_expr): Gimplify register-register type
8170 VIEW_CONVERT_EXPRs to separate stmts.
8171
8172 2014-01-07 Jeff Law <law@redhat.com>
8173
8174 PR middle-end/53623
8175 * ree.c (combine_set_extension): Handle case where source
8176 and destination registers in an extension insn are different.
8177 (combine_reaching_defs): Allow source and destination registers
8178 in extension to be different under limited circumstances.
8179 (add_removable_extension): Remove restriction that the
8180 source and destination registers in the extension are the same.
8181 (find_and_remove_re): Emit a copy from the extension's
8182 destination to its source after the defining insn if
8183 the source and destination registers are different.
8184
8185 PR middle-end/59285
8186 * ifcvt.c (merge_if_block): If we are merging a block with more than
8187 one successor with a block with no successors, remove any BARRIER
8188 after the second block.
8189
8190 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8191
8192 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8193
8194 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8195
8196 PR target/59652
8197 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8198 for 14-bit register offsets when INT14_OK_STRICT is false.
8199
8200 2014-01-07 Roland Stigge <stigge@antcom.de>
8201 Michael Meissner <meissner@linux.vnet.ibm.com>
8202
8203 PR 57386/target
8204 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8205 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8206
8207 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8208
8209 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8210 -mcpu.
8211
8212 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8213
8214 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8215 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8216 rtx is const0_rtx or not.
8217
8218 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8219
8220 PR target/58115
8221 * target-globals.c (save_target_globals): Remove this_fn_optab
8222 handling.
8223 * toplev.c: Include optabs.h.
8224 (target_reinit): Temporarily restore the global options if another
8225 set of options are in force.
8226
8227 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8228
8229 PR rtl-optimization/58668
8230 * cfgcleanup.c (flow_find_cross_jump): Don't count
8231 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8232 to determine what is counted.
8233 (flow_find_head_matching_sequence): Use active_insn_p to determine
8234 what is counted.
8235 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8236 counting change.
8237 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8238 determine what is counted.
8239
8240 PR tree-optimization/59643
8241 * tree-predcom.c (split_data_refs_to_components): If one dr is
8242 read and one write, determine_offset fails and the write isn't
8243 in the bad component, just put the read into the bad component.
8244
8245 2014-01-07 Mike Stump <mikestump@comcast.net>
8246 Jakub Jelinek <jakub@redhat.com>
8247
8248 PR pch/59436
8249 * tree-core.h (struct tree_optimization_option): Change optabs
8250 type from unsigned char * to void *.
8251 * optabs.c (init_tree_optimization_optabs): Adjust
8252 TREE_OPTIMIZATION_OPTABS initialization.
8253
8254 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8255
8256 PR target/59644
8257 * config/i386/i386.h (struct machine_function): Add
8258 no_drap_save_restore field.
8259 * config/i386/i386.c (ix86_save_reg): Use
8260 !cfun->machine->no_drap_save_restore instead of
8261 crtl->stack_realign_needed.
8262 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8263 this function clears frame_pointer_needed. Set
8264 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8265 and DRAP reg is needed.
8266
8267 2014-01-06 Marek Polacek <polacek@redhat.com>
8268
8269 PR c/57773
8270 * doc/implement-c.texi: Mention that other integer types are
8271 permitted as bit-field types in strictly conforming mode.
8272
8273 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8274
8275 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8276 is newly allocated.
8277
8278 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8279
8280 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8281
8282 2014-01-06 Martin Jambor <mjambor@suse.cz>
8283
8284 PR ipa/59008
8285 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8286 to int.
8287 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8288
8289 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8290
8291 PR debug/59350
8292 PR debug/59510
8293 * var-tracking.c (add_stores): Preserve the value of the source even if
8294 we don't record the store.
8295
8296 2014-01-06 Terry Guo <terry.guo@arm.com>
8297
8298 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8299
8300 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8301
8302 PR bootstrap/59541
8303 * config/darwin.c (darwin_function_section): Adjust return values to
8304 correspond to optimisation changes made in r206070.
8305
8306 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8307
8308 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8309 from prefetch_block tune setting.
8310 (nocona_cost): Correct size of prefetch block to 64.
8311
8312 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8313
8314 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8315 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8316 used to save the static chain register in the computation of the offset
8317 from which the FP registers need to be restored.
8318
8319 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8320
8321 PR tree-optimization/59519
8322 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8323 ICE if get_current_def (current_new_name) is already non-NULL, as long
8324 as it is a phi result of some other phi in *new_exit_bb that has
8325 the same argument.
8326
8327 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8328 or vmovdqu* for misaligned_operand.
8329 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8330 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8331 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8332 aligned_mem for AVX512F masked aligned load and store builtins and for
8333 non-temporal moves.
8334
8335 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8336
8337 PR tree-optimization/59651
8338 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8339 Address range for negative step should be added by TYPE_SIZE_UNIT.
8340
8341 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8342
8343 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8344 overlapping registers also for registers other than the stack pointer.
8345
8346 2014-01-03 Marek Polacek <polacek@redhat.com>
8347
8348 PR other/59661
8349 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8350 __builtin_FILE.
8351
8352 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8353
8354 PR target/59625
8355 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8356 asm goto as jump.
8357
8358 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8359 (push splitter): Use <P:MODE_SIZE> instead of
8360 GET_MODE_SIZE (<P:MODE>mode).
8361 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8362 (mov -1, reg peephole2): Likewise.
8363 * config/i386/sse.md (*mov<mode>_internal,
8364 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8365 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8366 *<code><mode>3, *andnot<mode>3<mask_name>,
8367 <mask_codefor><code><mode>3<mask_name>): Likewise.
8368 * config/i386/subst.md (mask_mode512bit_condition,
8369 sd_mask_mode512bit_condition): Likewise.
8370
8371 2014-01-02 Xinliang David Li <davidxl@google.com>
8372
8373 PR tree-optimization/59303
8374 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8375 (dump_predicates): Better output format.
8376 (pred_equal_p): New function.
8377 (is_neq_relop_p): Ditto.
8378 (is_neq_zero_form_p): Ditto.
8379 (pred_expr_equal_p): Ditto.
8380 (pred_neg_p): Ditto.
8381 (simplify_pred): Ditto.
8382 (simplify_preds_2): Ditto.
8383 (simplify_preds_3): Ditto.
8384 (simplify_preds_4): Ditto.
8385 (simplify_preds): Ditto.
8386 (push_pred): Ditto.
8387 (push_to_worklist): Ditto.
8388 (get_pred_info_from_cmp): Ditto.
8389 (is_degenerated_phi): Ditto.
8390 (normalize_one_pred_1): Ditto.
8391 (normalize_one_pred): Ditto.
8392 (normalize_one_pred_chain): Ditto.
8393 (normalize_preds): Ditto.
8394 (normalize_cond_1): Remove function.
8395 (normalize_cond): Ditto.
8396 (is_gcond_subset_of): Ditto.
8397 (is_subset_of_any): Ditto.
8398 (is_or_set_subset_of): Ditto.
8399 (is_and_set_subset_of): Ditto.
8400 (is_norm_cond_subset_of): Ditto.
8401 (pred_chain_length_cmp): Ditto.
8402 (convert_control_dep_chain_into_preds): Type change.
8403 (find_predicates): Ditto.
8404 (find_def_preds): Ditto.
8405 (destroy_predicates_vecs): Ditto.
8406 (find_matching_predicates_in_rest_chains): Ditto.
8407 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8408 (is_pred_expr_subset): Ditto.
8409 (is_pred_chain_subset_of): Ditto.
8410 (is_included_in): Ditto.
8411 (is_superset_of): Ditto.
8412
8413 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8414
8415 Update copyright years.
8416
8417 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8418
8419 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8420 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8421 config/arc/arc.md, config/arc/arc.opt,
8422 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8423 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8424 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8425 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8426 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8427 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8428 vtable-verify.c, vtable-verify.h: Use the standard form for the
8429 copyright notice.
8430
8431 2014-01-02 Tobias Burnus <burnus@net-b.de>
8432
8433 * gcc.c (process_command): Update copyright notice dates.
8434 * gcov-dump.c: Ditto.
8435 * gcov.c: Ditto.
8436 * doc/cpp.texi: Bump @copying's copyright year.
8437 * doc/cppinternals.texi: Ditto.
8438 * doc/gcc.texi: Ditto.
8439 * doc/gccint.texi: Ditto.
8440 * doc/gcov.texi: Ditto.
8441 * doc/install.texi: Ditto.
8442 * doc/invoke.texi: Ditto.
8443
8444 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8445
8446 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8447
8448 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8449
8450 * config/i386/sse.md (*mov<mode>_internal): Guard
8451 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8452
8453 PR rtl-optimization/59647
8454 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8455 new_rtx into UNSIGNED_FLOAT rtxes.
8456 \f
8457 Copyright (C) 2014 Free Software Foundation, Inc.
8458
8459 Copying and distribution of this file, with or without modification,
8460 are permitted in any medium without royalty provided the copyright
8461 notice and this notice are preserved.