| blob | 5437c256fe2f953f4c2dbf6ed65553e4ebae3cfa |
1 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3 * configure.ac (tga_func): Remove.
4 (LIB_TLS_SPEC): Remove.
5 * configure: Regenerate.
6 * config.in: Regenerate.
7 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
9 2014-04-25 Richard Biener <rguenther@suse.de>
11 PR ipa/60912
12 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
13 call stmt use/clobber sets during stmt walk instead of
14 walking the possibly incomplete set of caller edges.
16 2014-04-25 Richard Biener <rguenther@suse.de>
18 PR ipa/60911
19 * passes.c (apply_ipa_transforms): Inline into only caller ...
20 (execute_one_pass): ... here. Properly bring in function
21 bodies for nodes we want to apply IPA transforms to.
23 2014-04-24 Cong Hou <congh@google.com>
25 PR tree-optimization/60896
26 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
27 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
28 (vect_mark_pattern_stmts): Set the def type of all statements in
29 PATTERN_DEF_SEQ as vect_internal_def.
31 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
33 * doc/extend.texi (PowerPC Built-in Functions): Document new
34 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
35 functions.
36 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
38 * config/rs6000/predicates.md (const_0_to_3_operand): New
39 predicate to match 0..3 integer constants.
41 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
42 to support adding miscellaneous builtin functions.
43 (BU_DFP_MISC_2): Likewise.
44 (BU_P7_MISC_1): Likewise.
45 (BU_P7_MISC_2): Likewise.
46 (BU_P8V_MISC_3): Likewise.
47 (BU_MISC_1): Likewise.
48 (BU_MISC_2): Likewise.
49 (DIVWE): Add extended divide builtin functions.
50 (DIVWEO): Likewise.
51 (DIVWEU): Likewise.
52 (DIVWEUO): Likewise.
53 (DIVDE): Likewise.
54 (DIVDEO): Likewise.
55 (DIVDEU): Likewise.
56 (DIVDEUO): Likewise.
57 (DXEX): Add decimal floating-point builtin functions.
58 (DXEXQ): Likewise.
59 (DDEDPD): Likewise.
60 (DDEDPDQ): Likewise.
61 (DENBCD): Likewise.
62 (DENBCDQ): Likewise.
63 (DIEX): Likewise.
64 (DIEXQ): Likewise.
65 (DSCLI): Likewise.
66 (DSCLIQ): Likewise.
67 (DSCRI): Likewise.
68 (DSCRIQ): Likewise.
69 (CDTBCD): Add new BCD builtin functions.
70 (CBCDTD): Likewise.
71 (ADDG6S): Likewise.
72 (BCDADD): Likewise.
73 (BCDADD_LT): Likewise.
74 (BCDADD_EQ): Likewise.
75 (BCDADD_GT): Likewise.
76 (BCDADD_OV): Likewise.
77 (BCDSUB): Likewise.
78 (BCDSUB_LT): Likewise.
79 (BCDSUB_EQ): Likewise.
80 (BCDSUB_GT): Likewise.
81 (BCDSUB_OV): Likewise.
82 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
83 (UNPACK_TD): Likewise.
84 (PACK_TF): Likewise.
85 (UNPACK_TF): Likewise.
86 (UNPACK_TF_0): Likewise.
87 (UNPACK_TF_1): Likewise.
88 (PACK_V1TI): Likewise.
89 (UNPACK_V1TI): Likewise.
91 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
92 support for decimal floating point builtin functions.
93 (rs6000_expand_ternop_builtin): Add checks for the new builtin
94 functions that take constant arguments.
95 (rs6000_invalid_builtin): Add decimal floating point builtin
96 support.
97 (rs6000_init_builtins): Setup long double, _Decimal64, and
98 _Decimal128 types for new builtin functions.
99 (builtin_function_type): Set the unsigned flags appropriately for
100 the new builtin functions.
101 (rs6000_opt_masks): Add support for decimal floating point builtin
102 functions.
104 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
105 floating point builtin functions.
106 (RS6000_BTM_COMMON): Likewise.
107 (RS6000_BTI_long_double): Likewise.
108 (RS6000_BTI_dfloat64): Likewise.
109 (RS6000_BTI_dfloat128): Likewise.
110 (long_double_type_internal_node): Likewise.
111 (dfloat64_type_internal_node): Likewise.
112 (dfloat128_type_internal_node): Likewise.
114 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
115 2.07 bcd arithmetic instructions.
116 (UNSPEC_BCDSUB): Likewise.
117 (UNSPEC_BCD_OVERFLOW): Likewise.
118 (UNSPEC_BCD_ADD_SUB): Likewise.
119 (bcd_add_sub): Likewise.
120 (BCD_TEST): Likewise.
121 (bcd<bcd_add_sub>): Likewise.
122 (bcd<bcd_add_sub>_test): Likewise.
123 (bcd<bcd_add_sub>_test2): Likewise.
124 (bcd<bcd_add_sub>_<code>): Likewise.
125 (peephole2 for combined bcd ops): Likewise.
127 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
128 decimal floating point builtin functions.
129 (UNSPEC_DENBCD): Likewise.
130 (UNSPEC_DXEX): Likewise.
131 (UNSPEC_DIEX): Likewise.
132 (UNSPEC_DSCLI): Likewise.
133 (UNSPEC_DSCRI): Likewise.
134 (D64_D128): Likewise.
135 (dfp_suffix): Likewise.
136 (dfp_ddedpd_<mode>): Likewise.
137 (dfp_denbcd_<mode>): Likewise.
138 (dfp_dxex_<mode>): Likewise.
139 (dfp_diex_<mode>): Likewise.
140 (dfp_dscli_<mode>): Likewise.
141 (dfp_dscri_<mode>): Likewise.
143 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
144 builtin functions.
145 (UNSPEC_CDTBCD): Likewise.
146 (UNSPEC_CBCDTD): Likewise.
147 (UNSPEC_DIVE): Add support for new extended divide builtin
148 functions.
149 (UNSPEC_DIVEO): Likewise.
150 (UNSPEC_DIVEU): Likewise.
151 (UNSPEC_DIVEUO): Likewise.
152 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
153 pack/unpack 128-bit types.
154 (UNSPEC_PACK_128BIT): Likewise.
155 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
156 (udiv<mode>3): Use idiv_ldiv mode attribute.
157 (div<mode>3): Likewise.
158 (addg6s): Add new BCD builtin functions.
159 (cdtbcd): Likewise.
160 (cbcdtd): Likewise.
161 (UNSPEC_DIV_EXTEND): Add support for new extended divide
162 instructions.
163 (div_extend): Likewise.
164 (div<div_extend>_<mode>"): Likewise.
165 (FP128_64): Add support for new builtin functions to pack/unpack
166 128-bit types.
167 (unpack<mode>): Likewise.
168 (unpacktf_0): Likewise.
169 (unpacktf_1): Likewise.
170 (unpack<mode>_dm): Likewise.
171 (unpack<mode>_nodm): Likewise.
172 (pack<mode>): Likewise.
173 (unpackv1ti): Likewise.
174 (packv1ti): Likewise.
176 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
178 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
179 is disabled.
181 2014-04-24 Jakub Jelinek <jakub@redhat.com>
183 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
184 * gimplify.c (omp_is_private): Change last argument's type to int.
185 Only diagnose lastprivate if the simd argument is 1, only diagnose
186 linear if the simd argument is 2.
187 (gimplify_omp_for): Adjust omp_is_private callers. When adding
188 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
189 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
190 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
191 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
192 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
193 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
194 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
195 * tree-nested.c (convert_nonlocal_omp_clauses,
196 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
198 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
200 PR target/60822
201 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
202 operand 1.
204 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
206 * flag-types.h (enum ivar_visibility): Add.
208 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
210 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
211 function * argument.
213 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
215 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
217 2014-04-24 Radovan Obradovic <robradovic@mips.com>
218 Tom de Vries <tom@codesourcery.com>
220 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
221 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
222 reg-note.
223 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
224 * emit-rtl.c (try_split): Same.
226 2014-04-24 Radovan Obradovic <robradovic@mips.com>
227 Tom de Vries <tom@codesourcery.com>
229 * common.opt (fuse-caller-save): New option.
231 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
233 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
234 for big-endian.
236 2014-04-24 Richard Biener <rguenther@suse.de>
238 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
239 during TER and instead use the sepops interface for expanding
240 non-GIMPLE_SINGLE_RHS.
242 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
244 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
245 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
247 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
249 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
250 assembler 64-bit option.
251 * configure: Regenerate.
253 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
255 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
256 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
257 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
258 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
259 (TARGET_CRYPTO): Take TARGET_SIMD into account.
261 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
263 * config/aarch64/aarch64-builtins.c
264 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
265 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
266 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
267 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
268 builtins.
269 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
270 (Vrevsuff): New mode attribute.
272 2014-04-24 Terry Guo <terry.guo@arm.com>
274 * config/arm/arm.h (machine_function): Define variable
275 after_arm_reorg here.
276 * config/arm/arm.c (after_arm_reorg): Remove the definition.
277 (arm_split_constant): Update the way to access variable
278 after_arm_reorg.
279 (arm_reorg): Ditto.
280 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
282 2014-04-23 Tom de Vries <tom@codesourcery.com>
284 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
286 2014-04-23 David Malcolm <dmalcolm@redhat.com>
288 * is-a.h: Update comments to reflect the following changes to the
289 "pointerness" of the API, making the template parameter match the
290 return type, allowing use of is-a.h with typedefs of pointers.
291 (is_a_helper::cast): Return a T rather then a pointer to a T, so
292 that the return type matches the parameter to the is_a_helper.
293 (as_a): Likewise.
294 (dyn_cast): Likewise.
296 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
297 pointer from the is-a.h API.
299 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
300 (is_a_helper <cgraph_node *>::test): ...this, matching change to
301 is-a.h API.
302 (is_a_helper <varpool_node>::test): Likewise, convert to...
303 (is_a_helper <varpool_node *>::test): ...this.
305 (varpool_first_variable): Update for removal of implicit pointer
306 from the is-a.h API.
307 (varpool_next_variable): Likewise.
308 (varpool_first_static_initializer): Likewise.
309 (varpool_next_static_initializer): Likewise.
310 (varpool_first_defined_variable): Likewise.
311 (varpool_next_defined_variable): Likewise.
312 (cgraph_first_defined_function): Likewise.
313 (cgraph_next_defined_function): Likewise.
314 (cgraph_first_function): Likewise.
315 (cgraph_next_function): Likewise.
316 (cgraph_first_function_with_gimple_body): Likewise.
317 (cgraph_next_function_with_gimple_body): Likewise.
318 (cgraph_alias_target): Likewise.
319 (varpool_alias_target): Likewise.
320 (cgraph_function_or_thunk_node): Likewise.
321 (varpool_variable_node): Likewise.
322 (symtab_real_symbol_p): Likewise.
323 * cgraphunit.c (referred_to_p): Likewise.
324 (analyze_functions): Likewise.
325 (handle_alias_pairs): Likewise.
326 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
327 * gimple-ssa.h (gimple_vuse_op): Likewise.
328 (gimple_vdef_op): Likewise.
329 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
330 * gimple.c (gimple_build_asm_1): Likewise.
331 (gimple_build_try): Likewise.
332 (gimple_build_resx): Likewise.
333 (gimple_build_eh_dispatch): Likewise.
334 (gimple_build_omp_for): Likewise.
335 (gimple_omp_for_set_clauses): Likewise.
337 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
338 (is_a_helper <gimple_statement_asm *>::test): ...this.
339 (is_a_helper <gimple_statement_bind>::test): Convert to...
340 (is_a_helper <gimple_statement_bind *>::test): ...this.
341 (is_a_helper <gimple_statement_call>::test): Convert to...
342 (is_a_helper <gimple_statement_call *>::test): ...this.
343 (is_a_helper <gimple_statement_catch>::test): Convert to...
344 (is_a_helper <gimple_statement_catch *>::test): ...this.
345 (is_a_helper <gimple_statement_resx>::test): Convert to...
346 (is_a_helper <gimple_statement_resx *>::test): ...this.
347 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
348 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
349 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
350 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
351 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
352 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
353 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
354 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
355 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
356 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
357 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
358 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
359 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
360 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
361 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
362 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
363 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
364 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
365 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
366 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
367 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
368 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
369 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
370 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
371 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
372 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
373 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
374 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
375 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
376 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
377 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
378 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
379 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
380 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
381 (is_a_helper <gimple_statement_phi>::test): Convert to...
382 (is_a_helper <gimple_statement_phi *>::test): ...this.
383 (is_a_helper <gimple_statement_transaction>::test): Convert to...
384 (is_a_helper <gimple_statement_transaction *>::test): ...this.
385 (is_a_helper <gimple_statement_try>::test): Convert to...
386 (is_a_helper <gimple_statement_try *>::test): ...this.
387 (is_a_helper <gimple_statement_wce>::test): Convert to...
388 (is_a_helper <gimple_statement_wce *>::test): ...this.
389 (is_a_helper <const gimple_statement_asm>::test): Convert to...
390 (is_a_helper <const gimple_statement_asm *>::test): ...this.
391 (is_a_helper <const gimple_statement_bind>::test): Convert to...
392 (is_a_helper <const gimple_statement_bind *>::test): ...this.
393 (is_a_helper <const gimple_statement_call>::test): Convert to...
394 (is_a_helper <const gimple_statement_call *>::test): ...this.
395 (is_a_helper <const gimple_statement_catch>::test): Convert to...
396 (is_a_helper <const gimple_statement_catch *>::test): ...this.
397 (is_a_helper <const gimple_statement_resx>::test): Convert to...
398 (is_a_helper <const gimple_statement_resx *>::test): ...this.
399 (is_a_helper <const gimple_statement_eh_dispatch>::test):
400 Convert to...
401 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
402 ...this.
403 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
404 to...
405 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
406 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
407 Convert to...
408 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
409 ...this.
410 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
411 Convert to...
412 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
413 ...this.
414 (is_a_helper <const gimple_statement_omp_return>::test): Convert
415 to...
416 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
417 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
418 to...
419 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
420 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
421 to...
422 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
423 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
424 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
425 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
426 to...
427 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
428 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
429 to...
430 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
431 (is_a_helper <const gimple_statement_omp_target>::test): Convert
432 to...
433 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
434 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
435 to...
436 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
437 (is_a_helper <const gimple_statement_omp_single>::test): Convert
438 to...
439 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
440 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
441 to...
442 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
443 (is_a_helper <const gimple_statement_omp_task>::test): Convert
444 to...
445 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
446 (is_a_helper <const gimple_statement_phi>::test): Convert to...
447 (is_a_helper <const gimple_statement_phi *>::test): ...this.
448 (is_a_helper <const gimple_statement_transaction>::test): Convert
449 to...
450 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
451 (is_a_helper <const gimple_statement_with_ops>::test): Convert
452 to...
453 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
454 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
455 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
456 (is_a_helper <const gimple_statement_with_memory_ops>::test):
457 Convert to...
458 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
459 ...this.
460 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
461 to...
462 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
464 (gimple_use_ops): Update for removal of implicit pointer from the
465 is-a.h API.
466 (gimple_set_use_ops): Likewise.
467 (gimple_vuse): Likewise.
468 (gimple_vdef): Likewise.
469 (gimple_vuse_ptr): Likewise.
470 (gimple_vdef_ptr): Likewise.
471 (gimple_set_vuse): Likewise.
472 (gimple_set_vdef): Likewise.
473 (gimple_omp_return_set_lhs): Likewise.
474 (gimple_omp_return_lhs): Likewise.
475 (gimple_omp_return_lhs_ptr): Likewise.
476 (gimple_call_fntype): Likewise.
477 (gimple_call_set_fntype): Likewise.
478 (gimple_call_set_internal_fn): Likewise.
479 (gimple_call_use_set): Likewise.
480 (gimple_call_clobber_set): Likewise.
481 (gimple_bind_vars): Likewise.
482 (gimple_bind_set_vars): Likewise.
483 (gimple_bind_body_ptr): Likewise.
484 (gimple_bind_set_body): Likewise.
485 (gimple_bind_add_stmt): Likewise.
486 (gimple_bind_block): Likewise.
487 (gimple_bind_set_block): Likewise.
488 (gimple_asm_ninputs): Likewise.
489 (gimple_asm_noutputs): Likewise.
490 (gimple_asm_nclobbers): Likewise.
491 (gimple_asm_nlabels): Likewise.
492 (gimple_asm_input_op): Likewise.
493 (gimple_asm_input_op_ptr): Likewise.
494 (gimple_asm_output_op): Likewise.
495 (gimple_asm_output_op_ptr): Likewise.
496 (gimple_asm_set_output_op): Likewise.
497 (gimple_asm_clobber_op): Likewise.
498 (gimple_asm_set_clobber_op): Likewise.
499 (gimple_asm_label_op): Likewise.
500 (gimple_asm_set_label_op): Likewise.
501 (gimple_asm_string): Likewise.
502 (gimple_catch_types): Likewise.
503 (gimple_catch_types_ptr): Likewise.
504 (gimple_catch_handler_ptr): Likewise.
505 (gimple_catch_set_types): Likewise.
506 (gimple_catch_set_handler): Likewise.
507 (gimple_eh_filter_types): Likewise.
508 (gimple_eh_filter_types_ptr): Likewise.
509 (gimple_eh_filter_failure_ptr): Likewise.
510 (gimple_eh_filter_set_types): Likewise.
511 (gimple_eh_filter_set_failure): Likewise.
512 (gimple_eh_must_not_throw_fndecl): Likewise.
513 (gimple_eh_must_not_throw_set_fndecl): Likewise.
514 (gimple_eh_else_n_body_ptr): Likewise.
515 (gimple_eh_else_e_body_ptr): Likewise.
516 (gimple_eh_else_set_n_body): Likewise.
517 (gimple_eh_else_set_e_body): Likewise.
518 (gimple_try_eval_ptr): Likewise.
519 (gimple_try_cleanup_ptr): Likewise.
520 (gimple_try_set_eval): Likewise.
521 (gimple_try_set_cleanup): Likewise.
522 (gimple_wce_cleanup_ptr): Likewise.
523 (gimple_wce_set_cleanup): Likewise.
524 (gimple_phi_capacity): Likewise.
525 (gimple_phi_num_args): Likewise.
526 (gimple_phi_result): Likewise.
527 (gimple_phi_result_ptr): Likewise.
528 (gimple_phi_set_result): Likewise.
529 (gimple_phi_arg): Likewise.
530 (gimple_phi_set_arg): Likewise.
531 (gimple_resx_region): Likewise.
532 (gimple_resx_set_region): Likewise.
533 (gimple_eh_dispatch_region): Likewise.
534 (gimple_eh_dispatch_set_region): Likewise.
535 (gimple_omp_critical_name): Likewise.
536 (gimple_omp_critical_name_ptr): Likewise.
537 (gimple_omp_critical_set_name): Likewise.
538 (gimple_omp_for_clauses): Likewise.
539 (gimple_omp_for_clauses_ptr): Likewise.
540 (gimple_omp_for_set_clauses): Likewise.
541 (gimple_omp_for_collapse): Likewise.
542 (gimple_omp_for_index): Likewise.
543 (gimple_omp_for_index_ptr): Likewise.
544 (gimple_omp_for_set_index): Likewise.
545 (gimple_omp_for_initial): Likewise.
546 (gimple_omp_for_initial_ptr): Likewise.
547 (gimple_omp_for_set_initial): Likewise.
548 (gimple_omp_for_final): Likewise.
549 (gimple_omp_for_final_ptr): Likewise.
550 (gimple_omp_for_set_final): Likewise.
551 (gimple_omp_for_incr): Likewise.
552 (gimple_omp_for_incr_ptr): Likewise.
553 (gimple_omp_for_set_incr): Likewise.
554 (gimple_omp_for_pre_body_ptr): Likewise.
555 (gimple_omp_for_set_pre_body): Likewise.
556 (gimple_omp_parallel_clauses): Likewise.
557 (gimple_omp_parallel_clauses_ptr): Likewise.
558 (gimple_omp_parallel_set_clauses): Likewise.
559 (gimple_omp_parallel_child_fn): Likewise.
560 (gimple_omp_parallel_child_fn_ptr): Likewise.
561 (gimple_omp_parallel_set_child_fn): Likewise.
562 (gimple_omp_parallel_data_arg): Likewise.
563 (gimple_omp_parallel_data_arg_ptr): Likewise.
564 (gimple_omp_parallel_set_data_arg): Likewise.
565 (gimple_omp_task_clauses): Likewise.
566 (gimple_omp_task_clauses_ptr): Likewise.
567 (gimple_omp_task_set_clauses): Likewise.
568 (gimple_omp_task_child_fn): Likewise.
569 (gimple_omp_task_child_fn_ptr): Likewise.
570 (gimple_omp_task_set_child_fn): Likewise.
571 (gimple_omp_task_data_arg): Likewise.
572 (gimple_omp_task_data_arg_ptr): Likewise.
573 (gimple_omp_task_set_data_arg): Likewise.
574 (gimple_omp_taskreg_clauses): Likewise.
575 (gimple_omp_taskreg_clauses_ptr): Likewise.
576 (gimple_omp_taskreg_set_clauses): Likewise.
577 (gimple_omp_taskreg_child_fn): Likewise.
578 (gimple_omp_taskreg_child_fn_ptr): Likewise.
579 (gimple_omp_taskreg_set_child_fn): Likewise.
580 (gimple_omp_taskreg_data_arg): Likewise.
581 (gimple_omp_taskreg_data_arg_ptr): Likewise.
582 (gimple_omp_taskreg_set_data_arg): Likewise.
583 (gimple_omp_task_copy_fn): Likewise.
584 (gimple_omp_task_copy_fn_ptr): Likewise.
585 (gimple_omp_task_set_copy_fn): Likewise.
586 (gimple_omp_task_arg_size): Likewise.
587 (gimple_omp_task_arg_size_ptr): Likewise.
588 (gimple_omp_task_set_arg_size): Likewise.
589 (gimple_omp_task_arg_align): Likewise.
590 (gimple_omp_task_arg_align_ptr): Likewise.
591 (gimple_omp_task_set_arg_align): Likewise.
592 (gimple_omp_single_clauses): Likewise.
593 (gimple_omp_single_clauses_ptr): Likewise.
594 (gimple_omp_single_set_clauses): Likewise.
595 (gimple_omp_target_clauses): Likewise.
596 (gimple_omp_target_clauses_ptr): Likewise.
597 (gimple_omp_target_set_clauses): Likewise.
598 (gimple_omp_target_child_fn): Likewise.
599 (gimple_omp_target_child_fn_ptr): Likewise.
600 (gimple_omp_target_set_child_fn): Likewise.
601 (gimple_omp_target_data_arg): Likewise.
602 (gimple_omp_target_data_arg_ptr): Likewise.
603 (gimple_omp_target_set_data_arg): Likewise.
604 (gimple_omp_teams_clauses): Likewise.
605 (gimple_omp_teams_clauses_ptr): Likewise.
606 (gimple_omp_teams_set_clauses): Likewise.
607 (gimple_omp_sections_clauses): Likewise.
608 (gimple_omp_sections_clauses_ptr): Likewise.
609 (gimple_omp_sections_set_clauses): Likewise.
610 (gimple_omp_sections_control): Likewise.
611 (gimple_omp_sections_control_ptr): Likewise.
612 (gimple_omp_sections_set_control): Likewise.
613 (gimple_omp_for_set_cond): Likewise.
614 (gimple_omp_for_cond): Likewise.
615 (gimple_omp_atomic_store_set_val): Likewise.
616 (gimple_omp_atomic_store_val): Likewise.
617 (gimple_omp_atomic_store_val_ptr): Likewise.
618 (gimple_omp_atomic_load_set_lhs): Likewise.
619 (gimple_omp_atomic_load_lhs): Likewise.
620 (gimple_omp_atomic_load_lhs_ptr): Likewise.
621 (gimple_omp_atomic_load_set_rhs): Likewise.
622 (gimple_omp_atomic_load_rhs): Likewise.
623 (gimple_omp_atomic_load_rhs_ptr): Likewise.
624 (gimple_omp_continue_control_def): Likewise.
625 (gimple_omp_continue_control_def_ptr): Likewise.
626 (gimple_omp_continue_set_control_def): Likewise.
627 (gimple_omp_continue_control_use): Likewise.
628 (gimple_omp_continue_control_use_ptr): Likewise.
629 (gimple_omp_continue_set_control_use): Likewise.
630 (gimple_transaction_body_ptr): Likewise.
631 (gimple_transaction_label): Likewise.
632 (gimple_transaction_label_ptr): Likewise.
633 (gimple_transaction_set_body): Likewise.
634 (gimple_transaction_set_label): Likewise.
636 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
637 * ipa-inline-analysis.c (inline_write_summary): Likewise.
638 * ipa-ref.c (ipa_record_reference): Likewise.
639 * ipa-reference.c (analyze_function): Likewise.
640 (ipa_reference_write_optimization_summary): Likewise.
641 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
642 (address_taken_from_non_vtable_p): Likewise.
643 (comdat_can_be_unshared_p_1): Likewise.
644 * lto-cgraph.c (lto_output_ref): Likewise.
645 (add_references): Likewise.
646 (compute_ltrans_boundary): Likewise.
647 (output_symtab): Likewise.
648 (input_ref): Likewise.
649 (input_cgraph_1): Likewise.
650 (output_cgraph_opt_summary): Likewise.
651 * lto-streamer-out.c (lto_output): Likewise.
652 (output_symbol_p): Likewise.
653 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
654 (lsei_start_function_in_partition): Likewise.
655 (lsei_next_variable_in_partition): Likewise.
656 (lsei_start_variable_in_partition): Likewise.
657 * symtab.c (insert_to_assembler_name_hash): Likewise.
658 (unlink_from_assembler_name_hash): Likewise.
659 (symtab_unregister_node): Likewise.
660 (symtab_remove_node): Likewise.
661 (dump_symtab_node): Likewise.
662 (verify_symtab_base): Likewise.
663 (verify_symtab_node): Likewise.
664 (symtab_make_decl_local): Likewise.
665 (symtab_alias_ultimate_target): Likewise.
666 (symtab_resolve_alias): Likewise.
667 (symtab_get_symbol_partitioning_class): Likewise.
668 * tree-phinodes.c (allocate_phi_node): Likewise.
669 (reserve_phi_args_for_new_edge): Likewise.
670 (remove_phi_args): Likewise.
671 * varpool.c (varpool_node_for_asm): Likewise.
672 (varpool_remove_unreferenced_decls): Likewise.
674 2014-04-23 Jeff Law <law@redhat.com>
676 PR tree-optimization/60902
677 * tree-ssa-threadedge.c
678 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
679 invalidate outputs from statements that do not produce useful
680 outputs for threading.
682 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
684 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
685 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
686 machine descriptions for Stack Smashing Protector.
688 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
690 * aarch64.md (<optab>_rol<mode>3): New pattern.
691 (<optab>_rolsi3_uxtw): Likewise.
692 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
694 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
696 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
697 (arm_cortex_a12_tune): Likewise.
699 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
701 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
703 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
705 * config/arm/arm.md (arm_rev16si2): New pattern.
706 (arm_rev16si2_alt): Likewise.
707 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
709 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
711 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
712 (rev16<mode>2_alt): Likewise.
713 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
714 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
715 (aarch_rev16_shleft_mask_imm_p): Likewise.
716 (aarch_rev16_p_1): Likewise.
717 (aarch_rev16_p): Likewise.
718 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
719 (aarch_rev16_shright_mask_imm_p): Likewise.
720 (aarch_rev16_shleft_mask_imm_p): Likewise.
722 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
724 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
725 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
726 rev cost.
727 (cortex_a53_extra_costs): Likewise.
728 (cortex_a57_extra_costs): Likewise.
729 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
730 (cortexa7_extra_costs): Likewise.
731 (cortexa8_extra_costs): Likewise.
732 (cortexa12_extra_costs): Likewise.
733 (cortexa15_extra_costs): Likewise.
734 (v7m_extra_costs): Likewise.
735 (arm_new_rtx_costs): Handle BSWAP.
737 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
739 * config/arm/arm.c (cortexa8_extra_costs): New table.
740 (arm_cortex_a8_tune): New tuning struct.
741 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
743 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
745 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
747 2014-04-23 Richard Biener <rguenther@suse.de>
749 * Makefile.in (OBJS): Remove loop-unswitch.o.
750 * loop-unswitch.c: Delete.
751 * tree-pass.h (make_pass_rtl_unswitch): Remove.
752 * passes.def (pass_rtl_unswitch): Likewise.
753 * loop-init.c (gate_rtl_unswitch): Likewise.
754 (rtl_unswitch): Likewise.
755 (pass_data_rtl_unswitch): Likewise.
756 (pass_rtl_unswitch): Likewise.
757 (make_pass_rtl_unswitch): Likewise.
758 * rtl.h (reversed_condition): Likewise.
759 (compare_and_jump_seq): Likewise.
760 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
761 and make static.
762 * loop-unroll.c (compare_and_jump_seq): Likewise.
764 2014-04-23 Richard Biener <rguenther@suse.de>
766 PR tree-optimization/60903
767 * tree-ssa-loop-im.c (analyze_memory_references): Remove
768 commented code block.
769 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
770 loop flags to newly created BBs and edges.
772 2014-04-23 Nick Clifton <nickc@redhat.com>
774 * config/msp430/msp430.c (msp430_handle_option): Move function
775 to msp430-common.c
776 (msp430_option_override): Simplify mcu and mcpu option handling.
777 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
778 support for -mhwmult command line option.
779 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
780 -mhwmult command line option.
781 (msp430_hwmult_enabled): Delete.
782 (msp43o_output_labelref): Add support for -mhwmult command line
783 option.
784 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
785 (umulsidi3): Likewise.
786 * config/msp430/msp430.opt (mmcu): Add Report attribute.
787 (mcpu, mlarge, msmall): Likewise.
788 (mhwmult): New option.
789 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
790 prototype.
791 (msp430_is_f5_mcu): Remove prototype.
792 (msp430_use_f5_series_hwmult): Add prototype.
793 * config/msp430/msp430-opts.h: New file.
794 * common/config/msp430: New directory.
795 * common/config/msp430/msp430-common.c: New file.
796 * config.gcc (msp430): Remove target_has_targetm_common.
797 * doc/invoke.texi: Document -mhwmult command line option.
799 2014-04-23 Nick Clifton <nickc@redhat.com>
801 * config/i386/cygwin.h (ENDFILE_SPEC): Include
802 default-manifest.o if it can be found in the search path.
803 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
805 2014-04-23 Terry Guo <terry.guo@arm.com>
807 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
809 2014-04-23 Richard Biener <rguenther@suse.de>
811 PR middle-end/60895
812 * tree-inline.c (declare_return_variable): Use mark_addressable.
814 2014-04-23 Richard Biener <rguenther@suse.de>
816 PR middle-end/60891
817 * loop-init.c (loop_optimizer_init): Make sure to apply
818 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
820 2014-04-22 Jakub Jelinek <jakub@redhat.com>
822 PR sanitizer/60275
823 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
824 New options.
825 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
826 if flag_sanitize_undefined_trap_on_error.
827 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
828 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
829 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
830 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
831 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
832 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
833 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
834 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
835 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
836 * ubsan.c (ubsan_instrument_unreachable): Return
837 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
838 (ubsan_expand_null_ifn): Emit __builtin_trap ()
839 if flag_sanitize_undefined_trap_on_error and
840 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
841 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
842 instrument_bool_enum_load): Emit __builtin_trap () if
843 flag_sanitize_undefined_trap_on_error and
844 __builtin_handle_*_abort () if !flag_sanitize_recover.
845 * doc/invoke.texi (-fsanitize-recover,
846 -fsanitize-undefined-trap-on-error): Document.
848 2014-04-22 Christian Bruel <christian.bruel@st.com>
850 * config/sh/sh.md (mov<mode>): Replace movQIHI.
851 Force immediates to SImode.
853 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
855 * config/nios2/nios2.md (UNSPEC_ROUND): New.
856 (lroundsfsi2): New.
857 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
858 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
859 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
860 (nios2_fpu_insn): Add entry for round.
861 (N2FPU_NO_ERRNO_P): Define.
862 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
863 flag_errno_math.
864 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
866 2014-04-22 Richard Henderson <rth@redhat.com>
868 * config/aarch64/aarch64 (addti3, subti3): New expanders.
869 (add<GPI>3_compare0): Remove leading * from name.
870 (add<GPI>3_carryin): Likewise.
871 (sub<GPI>3_compare0): Likewise.
872 (sub<GPI>3_carryin): Likewise.
873 (<su_optab>mulditi3): New expander.
874 (multi3): New expander.
875 (madd<GPI>): Remove leading * from name.
877 2014-04-22 Martin Jambor <mjambor@suse.cz>
879 * cgraphclones.c (cgraph_function_versioning): Copy
880 ipa_transforms_to_apply instead of asserting it is empty.
882 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
884 PR target/60868
885 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
886 on count_exp to get mode.
888 2014-04-22 Andrew Pinski <apinski@cavium.com>
890 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
891 Handle TLS for ILP32.
892 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
893 (tlsie_small_<mode>): this and handle PTR.
894 (tlsie_small_sidi): New pattern.
895 (tlsle_small): Change to an expand to handle ILP32.
896 (tlsle_small_<mode>): New pattern.
897 (tlsdesc_small): Rename to ...
898 (tlsdesc_small_<mode>): this and handle PTR.
900 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
902 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
904 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
906 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
907 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
908 (aarch64_types_signed_poly_qualifiers): Likewise.
909 (aarch64_types_unsigned_signed_qualifiers): Likewise.
910 (aarch64_types_poly_signed_qualifiers): Likewise.
911 (TYPES_REINTERP_SS): Type macro added.
912 (TYPES_REINTERP_SU): Likewise.
913 (TYPES_REINTERP_SP): Likewise.
914 (TYPES_REINTERP_US): Likewise.
915 (TYPES_REINTERP_PS): Likewise.
916 (aarch64_fold_builtin): New expression folding added.
917 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
918 Declarations removed.
919 (REINTERP_SS): Declarations added.
920 (REINTERP_US): Likewise.
921 (REINTERP_PS): Likewise.
922 (REINTERP_SU): Likewise.
923 (REINTERP_SP): Likewise.
924 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
925 (vreinterpretq_p8_f64): Likewise.
926 (vreinterpret_p16_f64): Likewise.
927 (vreinterpretq_p16_f64): Likewise.
928 (vreinterpret_f32_f64): Likewise.
929 (vreinterpretq_f32_f64): Likewise.
930 (vreinterpret_f64_f32): Likewise.
931 (vreinterpret_f64_p8): Likewise.
932 (vreinterpret_f64_p16): Likewise.
933 (vreinterpret_f64_s8): Likewise.
934 (vreinterpret_f64_s16): Likewise.
935 (vreinterpret_f64_s32): Likewise.
936 (vreinterpret_f64_s64): Likewise.
937 (vreinterpret_f64_u8): Likewise.
938 (vreinterpret_f64_u16): Likewise.
939 (vreinterpret_f64_u32): Likewise.
940 (vreinterpret_f64_u64): Likewise.
941 (vreinterpretq_f64_f32): Likewise.
942 (vreinterpretq_f64_p8): Likewise.
943 (vreinterpretq_f64_p16): Likewise.
944 (vreinterpretq_f64_s8): Likewise.
945 (vreinterpretq_f64_s16): Likewise.
946 (vreinterpretq_f64_s32): Likewise.
947 (vreinterpretq_f64_s64): Likewise.
948 (vreinterpretq_f64_u8): Likewise.
949 (vreinterpretq_f64_u16): Likewise.
950 (vreinterpretq_f64_u32): Likewise.
951 (vreinterpretq_f64_u64): Likewise.
952 (vreinterpret_s64_f64): Likewise.
953 (vreinterpretq_s64_f64): Likewise.
954 (vreinterpret_u64_f64): Likewise.
955 (vreinterpretq_u64_f64): Likewise.
956 (vreinterpret_s8_f64): Likewise.
957 (vreinterpretq_s8_f64): Likewise.
958 (vreinterpret_s16_f64): Likewise.
959 (vreinterpretq_s16_f64): Likewise.
960 (vreinterpret_s32_f64): Likewise.
961 (vreinterpretq_s32_f64): Likewise.
962 (vreinterpret_u8_f64): Likewise.
963 (vreinterpretq_u8_f64): Likewise.
964 (vreinterpret_u16_f64): Likewise.
965 (vreinterpretq_u16_f64): Likewise.
966 (vreinterpret_u32_f64): Likewise.
967 (vreinterpretq_u32_f64): Likewise.
969 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
971 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
972 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
973 (vreinterpret_p8_s8): Likewise.
974 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
975 (vreinterpret_p8_s16): Likewise.
976 (vreinterpret_p8_s32): Likewise.
977 (vreinterpret_p8_s64): Likewise.
978 (vreinterpret_p8_f32): Likewise.
979 (vreinterpret_p8_u8): Likewise.
980 (vreinterpret_p8_u16): Likewise.
981 (vreinterpret_p8_u32): Likewise.
982 (vreinterpret_p8_u64): Likewise.
983 (vreinterpret_p8_p16): Likewise.
984 (vreinterpretq_p8_s8): Likewise.
985 (vreinterpretq_p8_s16): Likewise.
986 (vreinterpretq_p8_s32): Likewise.
987 (vreinterpretq_p8_s64): Likewise.
988 (vreinterpretq_p8_f32): Likewise.
989 (vreinterpretq_p8_u8): Likewise.
990 (vreinterpretq_p8_u16): Likewise.
991 (vreinterpretq_p8_u32): Likewise.
992 (vreinterpretq_p8_u64): Likewise.
993 (vreinterpretq_p8_p16): Likewise.
994 (vreinterpret_p16_s8): Likewise.
995 (vreinterpret_p16_s16): Likewise.
996 (vreinterpret_p16_s32): Likewise.
997 (vreinterpret_p16_s64): Likewise.
998 (vreinterpret_p16_f32): Likewise.
999 (vreinterpret_p16_u8): Likewise.
1000 (vreinterpret_p16_u16): Likewise.
1001 (vreinterpret_p16_u32): Likewise.
1002 (vreinterpret_p16_u64): Likewise.
1003 (vreinterpret_p16_p8): Likewise.
1004 (vreinterpretq_p16_s8): Likewise.
1005 (vreinterpretq_p16_s16): Likewise.
1006 (vreinterpretq_p16_s32): Likewise.
1007 (vreinterpretq_p16_s64): Likewise.
1008 (vreinterpretq_p16_f32): Likewise.
1009 (vreinterpretq_p16_u8): Likewise.
1010 (vreinterpretq_p16_u16): Likewise.
1011 (vreinterpretq_p16_u32): Likewise.
1012 (vreinterpretq_p16_u64): Likewise.
1013 (vreinterpretq_p16_p8): Likewise.
1014 (vreinterpret_f32_s8): Likewise.
1015 (vreinterpret_f32_s16): Likewise.
1016 (vreinterpret_f32_s32): Likewise.
1017 (vreinterpret_f32_s64): Likewise.
1018 (vreinterpret_f32_u8): Likewise.
1019 (vreinterpret_f32_u16): Likewise.
1020 (vreinterpret_f32_u32): Likewise.
1021 (vreinterpret_f32_u64): Likewise.
1022 (vreinterpret_f32_p8): Likewise.
1023 (vreinterpret_f32_p16): Likewise.
1024 (vreinterpretq_f32_s8): Likewise.
1025 (vreinterpretq_f32_s16): Likewise.
1026 (vreinterpretq_f32_s32): Likewise.
1027 (vreinterpretq_f32_s64): Likewise.
1028 (vreinterpretq_f32_u8): Likewise.
1029 (vreinterpretq_f32_u16): Likewise.
1030 (vreinterpretq_f32_u32): Likewise.
1031 (vreinterpretq_f32_u64): Likewise.
1032 (vreinterpretq_f32_p8): Likewise.
1033 (vreinterpretq_f32_p16): Likewise.
1034 (vreinterpret_s64_s8): Likewise.
1035 (vreinterpret_s64_s16): Likewise.
1036 (vreinterpret_s64_s32): Likewise.
1037 (vreinterpret_s64_f32): Likewise.
1038 (vreinterpret_s64_u8): Likewise.
1039 (vreinterpret_s64_u16): Likewise.
1040 (vreinterpret_s64_u32): Likewise.
1041 (vreinterpret_s64_u64): Likewise.
1042 (vreinterpret_s64_p8): Likewise.
1043 (vreinterpret_s64_p16): Likewise.
1044 (vreinterpretq_s64_s8): Likewise.
1045 (vreinterpretq_s64_s16): Likewise.
1046 (vreinterpretq_s64_s32): Likewise.
1047 (vreinterpretq_s64_f32): Likewise.
1048 (vreinterpretq_s64_u8): Likewise.
1049 (vreinterpretq_s64_u16): Likewise.
1050 (vreinterpretq_s64_u32): Likewise.
1051 (vreinterpretq_s64_u64): Likewise.
1052 (vreinterpretq_s64_p8): Likewise.
1053 (vreinterpretq_s64_p16): Likewise.
1054 (vreinterpret_u64_s8): Likewise.
1055 (vreinterpret_u64_s16): Likewise.
1056 (vreinterpret_u64_s32): Likewise.
1057 (vreinterpret_u64_s64): Likewise.
1058 (vreinterpret_u64_f32): Likewise.
1059 (vreinterpret_u64_u8): Likewise.
1060 (vreinterpret_u64_u16): Likewise.
1061 (vreinterpret_u64_u32): Likewise.
1062 (vreinterpret_u64_p8): Likewise.
1063 (vreinterpret_u64_p16): Likewise.
1064 (vreinterpretq_u64_s8): Likewise.
1065 (vreinterpretq_u64_s16): Likewise.
1066 (vreinterpretq_u64_s32): Likewise.
1067 (vreinterpretq_u64_s64): Likewise.
1068 (vreinterpretq_u64_f32): Likewise.
1069 (vreinterpretq_u64_u8): Likewise.
1070 (vreinterpretq_u64_u16): Likewise.
1071 (vreinterpretq_u64_u32): Likewise.
1072 (vreinterpretq_u64_p8): Likewise.
1073 (vreinterpretq_u64_p16): Likewise.
1074 (vreinterpret_s8_s16): Likewise.
1075 (vreinterpret_s8_s32): Likewise.
1076 (vreinterpret_s8_s64): Likewise.
1077 (vreinterpret_s8_f32): Likewise.
1078 (vreinterpret_s8_u8): Likewise.
1079 (vreinterpret_s8_u16): Likewise.
1080 (vreinterpret_s8_u32): Likewise.
1081 (vreinterpret_s8_u64): Likewise.
1082 (vreinterpret_s8_p8): Likewise.
1083 (vreinterpret_s8_p16): Likewise.
1084 (vreinterpretq_s8_s16): Likewise.
1085 (vreinterpretq_s8_s32): Likewise.
1086 (vreinterpretq_s8_s64): Likewise.
1087 (vreinterpretq_s8_f32): Likewise.
1088 (vreinterpretq_s8_u8): Likewise.
1089 (vreinterpretq_s8_u16): Likewise.
1090 (vreinterpretq_s8_u32): Likewise.
1091 (vreinterpretq_s8_u64): Likewise.
1092 (vreinterpretq_s8_p8): Likewise.
1093 (vreinterpretq_s8_p16): Likewise.
1094 (vreinterpret_s16_s8): Likewise.
1095 (vreinterpret_s16_s32): Likewise.
1096 (vreinterpret_s16_s64): Likewise.
1097 (vreinterpret_s16_f32): Likewise.
1098 (vreinterpret_s16_u8): Likewise.
1099 (vreinterpret_s16_u16): Likewise.
1100 (vreinterpret_s16_u32): Likewise.
1101 (vreinterpret_s16_u64): Likewise.
1102 (vreinterpret_s16_p8): Likewise.
1103 (vreinterpret_s16_p16): Likewise.
1104 (vreinterpretq_s16_s8): Likewise.
1105 (vreinterpretq_s16_s32): Likewise.
1106 (vreinterpretq_s16_s64): Likewise.
1107 (vreinterpretq_s16_f32): Likewise.
1108 (vreinterpretq_s16_u8): Likewise.
1109 (vreinterpretq_s16_u16): Likewise.
1110 (vreinterpretq_s16_u32): Likewise.
1111 (vreinterpretq_s16_u64): Likewise.
1112 (vreinterpretq_s16_p8): Likewise.
1113 (vreinterpretq_s16_p16): Likewise.
1114 (vreinterpret_s32_s8): Likewise.
1115 (vreinterpret_s32_s16): Likewise.
1116 (vreinterpret_s32_s64): Likewise.
1117 (vreinterpret_s32_f32): Likewise.
1118 (vreinterpret_s32_u8): Likewise.
1119 (vreinterpret_s32_u16): Likewise.
1120 (vreinterpret_s32_u32): Likewise.
1121 (vreinterpret_s32_u64): Likewise.
1122 (vreinterpret_s32_p8): Likewise.
1123 (vreinterpret_s32_p16): Likewise.
1124 (vreinterpretq_s32_s8): Likewise.
1125 (vreinterpretq_s32_s16): Likewise.
1126 (vreinterpretq_s32_s64): Likewise.
1127 (vreinterpretq_s32_f32): Likewise.
1128 (vreinterpretq_s32_u8): Likewise.
1129 (vreinterpretq_s32_u16): Likewise.
1130 (vreinterpretq_s32_u32): Likewise.
1131 (vreinterpretq_s32_u64): Likewise.
1132 (vreinterpretq_s32_p8): Likewise.
1133 (vreinterpretq_s32_p16): Likewise.
1134 (vreinterpret_u8_s8): Likewise.
1135 (vreinterpret_u8_s16): Likewise.
1136 (vreinterpret_u8_s32): Likewise.
1137 (vreinterpret_u8_s64): Likewise.
1138 (vreinterpret_u8_f32): Likewise.
1139 (vreinterpret_u8_u16): Likewise.
1140 (vreinterpret_u8_u32): Likewise.
1141 (vreinterpret_u8_u64): Likewise.
1142 (vreinterpret_u8_p8): Likewise.
1143 (vreinterpret_u8_p16): Likewise.
1144 (vreinterpretq_u8_s8): Likewise.
1145 (vreinterpretq_u8_s16): Likewise.
1146 (vreinterpretq_u8_s32): Likewise.
1147 (vreinterpretq_u8_s64): Likewise.
1148 (vreinterpretq_u8_f32): Likewise.
1149 (vreinterpretq_u8_u16): Likewise.
1150 (vreinterpretq_u8_u32): Likewise.
1151 (vreinterpretq_u8_u64): Likewise.
1152 (vreinterpretq_u8_p8): Likewise.
1153 (vreinterpretq_u8_p16): Likewise.
1154 (vreinterpret_u16_s8): Likewise.
1155 (vreinterpret_u16_s16): Likewise.
1156 (vreinterpret_u16_s32): Likewise.
1157 (vreinterpret_u16_s64): Likewise.
1158 (vreinterpret_u16_f32): Likewise.
1159 (vreinterpret_u16_u8): Likewise.
1160 (vreinterpret_u16_u32): Likewise.
1161 (vreinterpret_u16_u64): Likewise.
1162 (vreinterpret_u16_p8): Likewise.
1163 (vreinterpret_u16_p16): Likewise.
1164 (vreinterpretq_u16_s8): Likewise.
1165 (vreinterpretq_u16_s16): Likewise.
1166 (vreinterpretq_u16_s32): Likewise.
1167 (vreinterpretq_u16_s64): Likewise.
1168 (vreinterpretq_u16_f32): Likewise.
1169 (vreinterpretq_u16_u8): Likewise.
1170 (vreinterpretq_u16_u32): Likewise.
1171 (vreinterpretq_u16_u64): Likewise.
1172 (vreinterpretq_u16_p8): Likewise.
1173 (vreinterpretq_u16_p16): Likewise.
1174 (vreinterpret_u32_s8): Likewise.
1175 (vreinterpret_u32_s16): Likewise.
1176 (vreinterpret_u32_s32): Likewise.
1177 (vreinterpret_u32_s64): Likewise.
1178 (vreinterpret_u32_f32): Likewise.
1179 (vreinterpret_u32_u8): Likewise.
1180 (vreinterpret_u32_u16): Likewise.
1181 (vreinterpret_u32_u64): Likewise.
1182 (vreinterpret_u32_p8): Likewise.
1183 (vreinterpret_u32_p16): Likewise.
1184 (vreinterpretq_u32_s8): Likewise.
1185 (vreinterpretq_u32_s16): Likewise.
1186 (vreinterpretq_u32_s32): Likewise.
1187 (vreinterpretq_u32_s64): Likewise.
1188 (vreinterpretq_u32_f32): Likewise.
1189 (vreinterpretq_u32_u8): Likewise.
1190 (vreinterpretq_u32_u16): Likewise.
1191 (vreinterpretq_u32_u64): Likewise.
1192 (vreinterpretq_u32_p8): Likewise.
1193 (vreinterpretq_u32_p16): Likewise.
1195 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1197 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1198 Pattern extended.
1199 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1200 extended.
1201 (sqabs): Likewise.
1202 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1203 (vqnegd_s64): Likewise.
1204 (vqabs_s64): Likewise.
1205 (vqabsd_s64): Likewise.
1207 2014-04-22 Richard Henderson <rth@redhat.com>
1209 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1210 computation to the top of the loop.
1212 2014-04-22 Renlin <renlin.li@arm.com>
1213 Jiong Wang <jiong.wang@arm.com>
1215 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1216 * config/aarch64/aarch64.c (aarch64_layout_frame)
1217 (aarch64_initial_elimination_offset): Likewise.
1219 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1221 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1222 Fix indentation.
1224 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1226 * machmode.h (bitwise_mode_for_mode): Declare.
1227 * stor-layout.h (bitwise_type_for_mode): Likewise.
1228 * stor-layout.c (bitwise_mode_for_mode): New function.
1229 (bitwise_type_for_mode): Likewise.
1230 * builtins.c (fold_builtin_memory_op): Use it instead of
1231 int_mode_for_mode and build_nonstandard_integer_type.
1233 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1235 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1236 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1237 (*-*-solaris2*): Simplify.
1238 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1239 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1240 *-*-solaris2.9* handling.
1242 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1243 as bug.
1244 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1245 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1246 handling, simplify.
1247 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1248 * configure: Regenerate.
1250 * config/i386/sol2-9.h: Remove.
1252 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1253 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1254 Remove Solaris 9 references.
1256 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1258 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1259 (floatuns<GPI:mode><GPF:mode>2): Remove.
1260 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1261 and floatuns conversions.
1262 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1263 and floatuns conversions.
1264 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1265 (w1,w2): New mode attributes for inequal width conversions.
1267 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1269 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1270 the output asm format.
1272 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1274 * config/aarch64/aarch64-simd.md
1275 (aarch64_cm<optab>di): Always split.
1276 (*aarch64_cm<optab>di): New.
1277 (aarch64_cmtstdi): Always split.
1278 (*aarch64_cmtstdi): New.
1280 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1282 PR tree-optimization/60823
1283 * omp-low.c (ipa_simd_modify_function_body): Go through
1284 all SSA_NAMEs and for those refering to vector arguments
1285 which are going to be replaced adjust SSA_NAME_VAR and,
1286 if it is a default definition, change it into a non-default
1287 definition assigned at the beginning of function from new_decl.
1288 (ipa_simd_modify_stmt_ops): Rewritten.
1289 * tree-dfa.c (set_ssa_default_def): When removing default def,
1290 check for NULL loc instead of NULL *loc.
1292 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1294 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1295 restrictions on core registers for DImode values in Thumb2.
1297 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1299 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1300 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1302 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1304 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1305 (*iordi_notzesidi_di): Likewise.
1306 (*iordi_notsesidi_di): Likewise.
1308 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1310 * config/arm/arm-protos.h (tune_params): New struct members.
1311 * config/arm/arm.c: Initialise tune_params per processor.
1312 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1313 for speed, based on new tune_params.
1315 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1317 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1318 added.
1319 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1320 macro.
1321 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1322 corrected.
1323 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1324 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1325 (vrnda_f64): Likewise.
1326 (vrndi_f64): Likewise.
1327 (vrndm_f64): Likewise.
1328 (vrndn_f64): Likewise.
1329 (vrndp_f64): Likewise.
1330 (vrndx_f64): Likewise.
1332 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1334 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1335 GET_MODE_SIZE argument is enum machine_mode.
1337 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1339 PR target/60910
1340 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1341 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1343 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1345 PR middle-end/60281
1346 * asan.c (asan_emit_stack_protection): Force the base to align to
1347 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1348 appropriate bits if STRICT_ALIGNMENT.
1349 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1350 when asan is on.
1351 (expand_used_vars): Leave a space in the stack frame for alignment
1352 if STRICT_ALIGNMENT.
1354 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1356 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1357 than a gimple.
1358 (gimple_store_p): Likewise.
1359 (gimple_assign_load_p): Likewise.
1360 (gimple_assign_cast_p): Likewise.
1361 (gimple_clobber_p): Likewise.
1363 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1364 rather than a gimple.
1365 (gimple_assign_cast_p): Likewise.
1367 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1369 PR target/60735
1370 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1371 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1373 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1374 more debug information for E500 if -mdebug=reg.
1376 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1378 PR target/60909
1379 * config/i386/i386.c (ix86_expand_builtin)
1380 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1381 register for target RTX.
1382 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1384 2014-04-18 Cong Hou <congh@google.com>
1386 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1387 the widen-mult pattern by handling two operands with different sizes,
1388 and operands whose size is smaller than half of the result type.
1390 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1392 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1393 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1394 (do_estimate_edge_time): Compute it.
1395 * ipa-inline.c (want_inline_small_function_p): Bypass
1396 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1398 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1400 * ipa-inline.c (spec_rem): New static variable.
1401 (dump_overall_stats): New function.
1402 (dump_inline_stats): New function.
1404 2014-04-18 Richard Henderson <rth@redhat.com>
1406 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1407 to GET_MODE_SIZE, not a reg_class_t.
1409 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1411 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1412 (vsx_xxmrglw_<mode>): Likewise.
1414 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1416 PR target/60876
1417 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1418 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1419 (rs6000_init_hard_regno_mode_ok): Likewise.
1421 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1423 * ipa-inline.c (inline_small_functions): Account only non-cold
1424 functions.
1425 * doc/invoke.texi (inline-unit-growth): Update documentation.
1427 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1429 * config/rs6000/rs6000.md (addti3, subti3): New.
1431 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1433 PR target/60863
1434 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1435 comment. Check optimize_insn_for_size_p instead of
1436 optimize_insn_for_speed_p.
1438 2014-04-17 Martin Jambor <mjambor@suse.cz>
1440 * gimple-iterator.c (gsi_start_edge): New function.
1441 * gimple-iterator.h (gsi_start_edge): Declare.
1442 * tree-sra.c (single_non_eh_succ): New function.
1443 (disqualify_ops_if_throwing_stmt): Renamed to
1444 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1445 having one non-EH successor BB.
1446 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1447 generate loads into replacements.
1448 (sra_modify_assign): Likewise and and also use the simple path for
1449 such statements.
1450 (sra_modify_function_body): Commit statements on edges.
1452 2014-04-17 Richard Biener <rguenther@suse.de>
1454 PR middle-end/60849
1455 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1456 comparison results and add clarifying comment.
1458 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1460 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1461 (blank_mode): Initialize it.
1462 (emit_mode_size_inline, emit_mode_nunits_inline,
1463 emit_mode_inner_inline): New functions.
1464 (emit_insn_modes_h): Call them and surround their output with
1465 #if GCC_VERSION >= 4001 ... #endif.
1466 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1467 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1468 mode_* arrays if the argument is __builtin_constant_p.
1469 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1470 is enum machine_mode.
1472 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1474 * passes.c (opt_pass::execute): Adjust.
1475 (pass_manager::execute_pass_mode_switching): Likewise.
1476 (early_local_passes::execute): Likewise.
1477 (execute_one_pass): Pass cfun to the pass's execute method.
1478 * tree-pass.h (opt_pass::execute): Add function * argument.
1479 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1480 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1481 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1482 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1483 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1484 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1485 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1486 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1487 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1488 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1489 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1490 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1491 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1492 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1493 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1494 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1495 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1496 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1497 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1498 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1499 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1500 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1501 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1502 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1503 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1504 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1505 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1506 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1507 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1508 Adjust.
1510 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1512 * passes.c (opt_pass::gate): Take function * argument.
1513 (gate_all_early_local_passes): Merge into
1514 (early_local_passes::gate): this.
1515 (gate_all_early_optimizations): Merge into
1516 (all_early_optimizations::gate): this.
1517 (gate_all_optimizations): Mege into
1518 (all_optimizations::gate): this.
1519 (gate_all_optimizations_g): Merge into
1520 (all_optimizations_g::gate): this.
1521 (gate_rest_of_compilation): Mege into
1522 (rest_of_compilation::gate): this.
1523 (gate_postreload): Merge into
1524 (postreload::gate): this.
1525 (dump_one_pass): Pass cfun to the pass's gate method.
1526 (execute_ipa_summary_passes): Likewise.
1527 (execute_one_pass): Likewise.
1528 (ipa_write_summaries_2): Likewise.
1529 (ipa_write_optimization_summaries_1): Likewise.
1530 (ipa_read_summaries_1): Likewise.
1531 (ipa_read_optimization_summaries_1): Likewise.
1532 (execute_ipa_stmt_fixups): Likewise.
1533 * tree-pass.h (opt_pass::gate): Add function * argument.
1534 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1535 combine-stack-adj.c, combine.c, compare-elim.c,
1536 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1537 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1538 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1539 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1540 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1541 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1542 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1543 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1544 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1545 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1546 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1547 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1548 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1549 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1550 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1551 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1552 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1553 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1554 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1555 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1556 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1557 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1558 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1559 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1560 var-tracking.c, vtable-verify.c, web.c: Adjust.
1562 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1564 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1565 * configure: Regenerate.
1567 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1569 * passes.c (dump_one_pass): don't check pass->has_gate.
1570 (execute_ipa_summary_passes): Likewise.
1571 (execute_one_pass): Likewise.
1572 (ipa_write_summaries_2): Likewise.
1573 (ipa_write_optimization_summaries_1): Likewise.
1574 (ipa_read_optimization_summaries_1): Likewise.
1575 (execute_ipa_stmt_fixups): Likewise.
1576 * tree-pass.h (pass_data::has_gate): Remove.
1577 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1578 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1579 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1580 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1581 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1582 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1583 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1584 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1585 gimple-low.c, gimple-ssa-isolate-paths.c,
1586 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1587 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1588 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1589 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1590 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1591 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1592 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1593 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1594 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1595 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1596 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1597 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1598 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1599 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1600 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1601 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1602 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1603 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1604 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1605 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1606 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1607 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1608 Adjust.
1610 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1612 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1613 declaration.
1614 * passes.c (pass_manager::register_dump_files_1): Merge into
1615 (pass_manager::register_dump_files): this, and remove its handling of
1616 properties since the pass always has the properties anyway.
1617 (pass_manager::pass_manager): Adjust.
1619 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1621 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1622 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1623 dealing with properties.
1624 (pass_manager::register_dump_files): Adjust.
1626 2014-03-20 Mark Wielaard <mjw@redhat.com>
1628 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1629 then represent the bound as normal constant value.
1631 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1633 PR target/60847
1634 Forward port from 4.8 branch
1635 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1637 * config/i386/bmiintrin.h (_blsi_u32): New.
1638 (_blsi_u64): Ditto.
1639 (_blsr_u32): Ditto.
1640 (_blsr_u64): Ditto.
1641 (_blsmsk_u32): Ditto.
1642 (_blsmsk_u64): Ditto.
1643 (_tzcnt_u32): Ditto.
1644 (_tzcnt_u64): Ditto.
1646 2014-04-17 Kito Cheng <kito@0xlab.org>
1648 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1650 2014-04-17 Richard Biener <rguenther@suse.de>
1652 PR middle-end/60849
1653 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1654 boolean results for comparisons.
1656 2014-04-17 Richard Biener <rguenther@suse.de>
1658 PR tree-optimization/60836
1659 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1660 initial PHI args to be gimple values.
1662 2014-04-17 Richard Biener <rguenther@suse.de>
1664 PR tree-optimization/60841
1665 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1666 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1667 of stmts to SLP build.
1668 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1669 (vect_analyze_slp): Likewise.
1670 (vect_analyze_slp_instance): Likewise.
1671 (vect_build_slp_tree): Limit overall SLP tree growth.
1672 * tree-vectorizer.h (vect_analyze_data_refs,
1673 vect_analyze_slp): Adjust prototypes.
1675 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1677 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1678 Silvermont.
1680 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1682 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1683 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1684 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1685 for TARGET_SLOW_PSHUFB
1687 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1689 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1690 * config/i386/i386.c (intel_cost): Ditto.
1692 2014-04-17 Joey Ye <joey.ye@arm.com>
1694 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1696 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1698 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1699 with -fuse-profile.
1701 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1703 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1704 (type_all_derivations_known_p): New predicate.
1705 (type_all_ctors_visible_p): New predicate.
1706 (type_possibly_instantiated_p): New predicate.
1707 (get_odr_type): Compute all_derivations_known.
1708 (dump_odr_type): Dump the flag.
1709 (maybe_record_type): Cleanup.
1710 (record_target_from_binfo): Add bases_to_consider array;
1711 record bases for types w/o instances and skip CXX destructor.
1712 (possible_polymorphic_call_targets_1): Add bases_to_consider
1713 and consider_construction parameters; check if type may have instance.
1714 (get_polymorphic_call_info): Set maybe_in_construction to true
1715 when we know nothing.
1716 (record_targets_from_bases): Skip CXX destructors; they are
1717 never called for types in construction.
1718 (possible_polymorphic_call_targets): Do not record target when
1719 type may not have instance.
1721 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1723 PR ipa/60854
1724 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1725 external aliases alive, too.
1727 2014-04-16 Andrew Pinski <apinski@cavium.com>
1729 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1730 definition.
1732 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1734 * final.c (compute_alignments): Do not apply loop alignment to a block
1735 falling through to the exit.
1737 2014-04-16 Catherine Moore <clm@codesourcery.com>
1739 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1740 Adjust constraints for microMIPS store patterns.
1742 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1744 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1746 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1748 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1749 (append_use): Run at -O0.
1750 (append_vdef): Likewise.
1751 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1752 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1754 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1756 PR tree-optimization/60844
1757 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1758 (propagate_op_to_single_use, remove_visited_stmt_chain,
1759 linearize_expr, repropagate_negates, reassociate_bb): Use it
1760 instead of gsi_remove.
1762 2014-04-16 Martin Jambor <mjambor@suse.cz>
1764 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1765 ipa_transforms_to_apply.
1766 (cgraph_function_versioning): Assert that old_node has empty
1767 ipa_transforms_to_apply.
1768 * trans-mem.c (ipa_tm_create_version): Likewise.
1769 * tree-inline.c (tree_function_versioning): Do not duplicate
1770 ipa_transforms_to_apply.
1772 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1774 PR target/60817
1775 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
1776 x86_64-*-* cases.
1777 Pass necessary as flags on 64-bit Solaris/x86.
1778 Use lowercase relocs for x86_64-*-*.
1779 * configure: Regenerate.
1781 2014-04-15 Jan Hubicka <jh@suse.cz>
1783 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
1784 (maybe_record_node, likely_target_p): Use it.
1786 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1788 PR target/60839
1789 Revert following patch
1791 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1793 PR target/60735
1794 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
1795 software floating point or no floating point registers, do not
1796 allow any type in the FPRs. Eliminate a test for SPE SIMD types
1797 in GPRs that occurs after we tested for GPRs that would never be
1798 true.
1800 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
1801 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
1802 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
1803 specifically allow DDmode, since that does not use the SPE SIMD
1804 instructions.
1806 2014-03-21 Mark Wielaard <mjw@redhat.com>
1808 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
1809 as unsigned or int depending on type and value used.
1811 2014-04-15 Richard Biener <rguenther@suse.de>
1813 PR rtl-optimization/56965
1814 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
1815 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
1816 ... here.
1817 * alias.c (true_dependence_1): Do not call
1818 nonoverlapping_component_refs_p.
1819 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
1820 nonoverlapping_component_refs_p.
1821 (indirect_refs_may_alias_p): Likewise.
1823 2014-04-15 Teresa Johnson <tejohnson@google.com>
1825 * cfg.c (dump_bb_info): Fix flags check.
1826 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
1828 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1830 PR rtl-optimization/60663
1831 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
1832 avoid 0 cost.
1834 2014-04-15 Richard Biener <rguenther@suse.de>
1836 * lto-streamer.h (LTO_major_version): Bump to 4.
1838 2014-04-15 Richard Biener <rguenther@suse.de>
1840 * common.opt (lto_partition_model): New enum.
1841 (flto-partition=): Merge separate options with a single with argument,
1842 add -flto-partition=one support.
1843 * flag-types.h (enum lto_partition_model): Declare.
1844 * opts.c (finish_options): Remove duplicate -flto-partition=
1845 option check.
1846 * lto-wrapper.c (run_gcc): Adjust.
1848 2014-04-15 Richard Biener <rguenther@suse.de>
1850 * alias.c (ncr_compar): New function.
1851 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
1853 2014-04-15 Richard Biener <rguenther@suse.de>
1855 * alias.c (record_component_aliases): Do not walk BINFOs.
1857 2014-04-15 Richard Biener <rguenther@suse.de>
1859 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1860 Add struct function argument and adjust.
1861 (find_func_aliases_for_call): Likewise.
1862 (find_func_aliases): Likewise.
1863 (find_func_clobbers): Likewise.
1864 (intra_create_variable_infos): Likewise.
1865 (compute_points_to_sets): Likewise.
1866 (ipa_pta_execute): Adjust. Do not push/pop cfun.
1868 2014-04-15 Richard Biener <rguenther@suse.de>
1870 * tree.c (iterative_hash_expr): Use enum tree_code_class
1871 to store TREE_CODE_CLASS.
1872 (tree_block): Likewise.
1873 (tree_set_block): Likewise.
1874 * tree.h (fold_build_pointer_plus_loc): Use
1875 convert_to_ptrofftype_loc.
1877 2014-04-15 Jakub Jelinek <jakub@redhat.com>
1879 PR plugins/59335
1880 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
1881 added in 4.9.
1883 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
1885 * cfgloop.h (struct loop): Move force_vectorize down.
1886 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
1887 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
1888 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
1889 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
1890 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
1891 * tree-core.h (enum annot_expr_kind): Add new kind values.
1892 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
1893 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
1894 kinds.
1895 * tree.def (ANNOTATE_EXPR): Tweak comment.
1897 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1899 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
1900 cxa_pure_virtual).
1902 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
1904 * tree.h (TYPE_IDENTIFIER): Declare.
1905 * tree.c (subrange_type_for_debug_p): Use it.
1906 * godump.c (go_format_type): Likewise.
1907 * dwarf2out.c (is_cxx_auto, modified_type_die,
1908 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
1909 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
1911 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1913 PR lto/60820
1914 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
1916 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
1918 * config/i386/i386.c (examine_argument): Return bool. Return true if
1919 parameter should be passed in memory.
1920 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
1921 (construct_container): Update calls to examine_argument.
1922 (function_arg_advance_64): Ditto.
1923 (return_in_memory_32): Merge with ix86_return_in_memory.
1924 (return_in_memory_64): Ditto.
1925 (return_in_memory_ms_64): Ditto.
1927 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1929 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
1930 * coverage.c (coverage_compute_profile_id): Handle externally visible
1931 symbols.
1933 2014-04-14 Martin Jambor <mjambor@suse.cz>
1935 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
1936 DECL_DISREGARD_INLINE_LIMITS functions.
1938 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
1940 PR target/60827
1941 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
1943 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
1945 PR target/60827
1946 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
1947 optimize_insn_for_speed_p instead of
1948 optimize_function_for_speed_p.
1950 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
1952 * doc/invoke.texi (free): Document AArch64.
1954 2014-04-14 Richard Biener <rguenther@suse.de>
1956 PR tree-optimization/60042
1957 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
1958 (insert_into_preds_of_block): Do not prevent PHI insertion
1959 for REFERENCE exprs here ...
1960 (eliminate_dom_walker::before_dom_children): ... but prevent
1961 their use here under similar conditions when applied to the
1962 IL after PRE optimizations.
1964 2014-04-14 Richard Biener <rguenther@suse.de>
1966 * passes.def: Move early points-to after early SRA.
1968 2014-04-14 Richard Biener <rguenther@suse.de>
1970 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
1971 check for which sign-changes we allow when forwarding
1972 a converted value into a switch.
1974 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
1976 * stor-layout.c (place_field): Finalize non-constant offset for the
1977 field, if any.
1979 2014-04-14 Richard Biener <rguenther@suse.de>
1981 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
1982 as argument.
1983 (expand_switch_using_bit_tests_p): Likewise.
1984 (process_switch): Compute and pass on speed_p based on the
1985 switch stmt.
1986 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
1987 optimize_bb_for_speed_p.
1989 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
1991 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
1992 * function.h (struct function): Rename has_force_vect_loops into
1993 has_force_vectorize_loops.
1994 * lto-streamer-in.c (input_cfg): Adjust for renaming.
1995 (input_struct_function_base): Likewise.
1996 * lto-streamer-out.c (output_cfg): Likewise.
1997 (output_struct_function_base): Likewise.
1998 * omp-low.c (expand_omp_simd): Likewise.
1999 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2000 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2001 (version_loop_for_if_conversion): Likewise.
2002 (tree_if_conversion): Likewise.
2003 (main_tree_if_conversion): Likewise.
2004 (gate_tree_if_conversion): Likewise.
2005 * tree-inline.c (copy_loops): Likewise.
2006 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2007 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2008 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2009 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2010 * tree-vectorizer.c (vectorize_loops): Likewise.
2011 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2013 2014-04-14 Richard Biener <rguenther@suse.de>
2015 PR lto/60720
2016 * lto-streamer-out.c (wrap_refs): New function.
2017 (lto_output): Wrap symbol references in global initializes in
2018 type-preserving MEM_REFs.
2020 2014-04-14 Christian Bruel <christian.bruel@st.com>
2022 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2024 2014-04-14 Christian Bruel <christian.bruel@st.com>
2026 * config/sh/sh.md (setmemqi): New expand pattern.
2027 * config/sh/sh.h (CLEAR_RATIO): Define.
2028 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2029 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2031 2014-04-14 Richard Biener <rguenther@suse.de>
2033 PR middle-end/55022
2034 * fold-const.c (negate_expr_p): Don't negate directional rounding
2035 division.
2036 (fold_negate_expr): Likewise.
2038 2014-04-14 Richard Biener <rguenther@suse.de>
2040 PR tree-optimization/59817
2041 PR tree-optimization/60453
2042 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2043 recursion to catch all CHRECs in the scalar evolution and restrict
2044 the predicate for the remains appropriately.
2046 2014-04-12 Catherine Moore <clm@codesourcery.com>
2048 * config/mips/constraints.md: Add new register constraint "kb".
2049 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2050 (*movhi_internal): Likewise.
2051 (*movqi_internal): Likewise.
2052 * config/mips/mips.h (M16_STORE_REGS): New register class.
2053 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2054 (REG_CLASS_CONTENTS): Likewise.
2055 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2057 2014-04-11 Tobias Burnus <burnus@net-b.de>
2059 PR c/60194
2060 * doc/invoke.texi (-Wformat-signedness): Document it.
2061 (Wformat=2): Mention that this enables -Wformat-signedness.
2063 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2065 * common/config/epiphany/epiphany-common.c
2066 (epiphany_option_optimization_table): Enable section anchors by
2067 default at -O1 or higher.
2068 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2069 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2070 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2071 carries no extra cost.
2072 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2073 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2074 * config/epiphany/predicates.md (memclob_operand): New predicate.
2075 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2076 Use memclob_operand predicate and X constraint for operand 3.
2078 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2080 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2081 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2082 its operands.
2084 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2086 PR rtl-optimization/60651
2087 * mode-switching.c (optimize_mode_switching): Make sure to emit
2088 sets of a lower numbered entity before sets of a higher numbered
2089 entity to a mode of the same or lower priority.
2090 When creating a seginfo for a basic block that starts with a code
2091 label, move the insertion point past the code label.
2092 (new_seginfo): Document and enforce requirement that
2093 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2094 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2095 * doc/tm.texi: Regenerate.
2097 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2099 PR target/60811
2100 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2102 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2104 * BASE-VER: Set to 4.10.0.
2106 2014-04-11 Tobias Burnus <burnus@net-b.de>
2108 PR other/59055
2109 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2110 * doc/gcc.texi (Service): Update description in the @menu
2111 * doc/invoke.texi (Option Summary): Remove misplaced and
2112 duplicated @menu.
2114 2014-04-11 Steve Ellcey <sellcey@mips.com>
2115 Jakub Jelinek <jakub@redhat.com>
2117 PR middle-end/60556
2118 * expr.c (convert_move): Use emit_store_flag_force instead of
2119 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2120 argument to it.
2122 2014-04-11 Richard Biener <rguenther@suse.de>
2124 PR middle-end/60797
2125 * varasm.c (assemble_alias): Avoid endless error reporting
2126 recursion by setting TREE_ASM_WRITTEN.
2128 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2130 * config/s390/s390.md: Add a splitter for NOT rtx.
2132 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2134 PR rtl-optimization/60663
2135 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2137 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2138 Jakub Jelinek <jakub@redhat.com>
2140 PR lto/60567
2141 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2142 flag from decl_node to node.
2144 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2146 PR debug/60655
2147 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2148 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2149 ameliorating the cases where it can be.
2151 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2153 Revert
2154 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2156 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2157 (loadsync_<mode>): Change mode.
2158 (load_quadpti, store_quadpti): New.
2159 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2160 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2161 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2163 2014-04-09 Cong Hou <congh@google.com>
2165 PR testsuite/60773
2166 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2167 documentation.
2169 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2171 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2172 instead of vnor to exploit possible fusion opportunity in the
2173 future.
2174 (altivec_expand_vec_perm_const_le): Likewise.
2176 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2178 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2179 (loadsync_<mode>): Change mode.
2180 (load_quadpti, store_quadpti): New.
2181 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2182 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2184 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2186 PR target/60763
2187 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2188 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2189 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2191 2014-04-08 Richard Biener <rguenther@suse.de>
2193 PR middle-end/60706
2194 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2195 a 64bit widest int print double-int similar to on HWI64 hosts.
2197 2014-04-08 Richard Biener <rguenther@suse.de>
2199 PR tree-optimization/60785
2200 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2201 default defs properly.
2203 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2205 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2206 (Weffc++): Likewise.
2208 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2210 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2211 set completep to false rather than true.
2213 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2215 PR target/60504
2216 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2217 ARM_TARGET2_DWARF_FORMAT.
2219 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2221 PR target/60609
2222 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2223 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2224 ADDR_DIFF_VEC.
2226 2014-04-07 Richard Biener <rguenther@suse.de>
2228 PR tree-optimization/60766
2229 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2230 (may_eliminate_iv): Convert cand_value_at result to desired type.
2232 2014-04-07 Jason Merrill <jason@redhat.com>
2234 PR c++/60731
2235 * common.opt (-fno-gnu-unique): Add.
2236 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2238 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2240 * haifa-sched.c: Fix outdated function reference and minor
2241 grammar errors in introductory comment.
2243 2014-04-07 Richard Biener <rguenther@suse.de>
2245 PR middle-end/60750
2246 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2247 for noreturn calls.
2248 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2250 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2252 PR debug/55794
2253 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2254 size accounting for thunks.
2255 (pa_asm_output_mi_thunk): Use final_start_function() and
2256 final_end_function() to output function start and end directives.
2258 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2260 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2261 device specific ISA/ feature information. Remove short_sp and
2262 errata_skip ds. Add avr_device_specific_features enum to have device
2263 specific info.
2264 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2265 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2266 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2267 updated device specific info.
2268 * config/avr/avr-mcus.def: Merge device specific details to
2269 dev_attribute field.
2270 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2271 errata_skip.
2272 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2273 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2274 assembler if RMW isa supported by current device.
2275 * config/avr/genmultilib.awk: Update as device info structure changed.
2276 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2278 2014-04-04 Cong Hou <congh@google.com>
2280 PR tree-optimization/60656
2281 * tree-vect-stmts.c (supportable_widening_operation):
2282 Fix a bug that elements in a vector with vect_used_by_reduction
2283 property are incorrectly reordered when the operation on it is not
2284 consistant with the one in reduction operation.
2286 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2288 PR rtl-optimization/60155
2289 * gcse.c (record_set_data): New function.
2290 (single_set_gcse): New function.
2291 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2292 (hoist_code): Likewise.
2293 (get_pressure_class_and_nregs): Likewise.
2295 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2297 * explow.c (probe_stack_range): Emit a final optimization blockage.
2299 2014-04-04 Anthony Green <green@moxielogic.com>
2301 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2302 typos.
2304 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2306 PR ipa/59626
2307 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2308 flags are set only during streaming.
2309 * ipa.c (process_references, walk_polymorphic_call_targets,
2310 symtab_remove_unreachable_nodes): Drop bodies of always inline
2311 after early inlining.
2312 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2314 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2315 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2317 PR debug/60655
2318 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2319 containing a NOT.
2321 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2323 PR bootstrap/60743
2324 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2325 duration.
2326 (cortex_a53_fdivd): Likewise.
2328 2014-04-04 Martin Jambor <mjambor@suse.cz>
2330 PR ipa/60640
2331 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2332 Adjust all callers.
2333 * cgraph.c (clone_of_p): Also return true if thunks match.
2334 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2335 cgraph_function_or_thunk_node and an obsolete comment.
2336 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2337 file.
2338 (build_function_decl_skip_args): Likewise.
2339 (set_new_clone_decl_and_node_flags): New function.
2340 (duplicate_thunk_for_node): Likewise.
2341 (redirect_edge_duplicating_thunks): Likewise.
2342 (cgraph_clone_node): New parameter args_to_skip, pass it to
2343 redirect_edge_duplicating_thunks which is called instead of
2344 cgraph_redirect_edge_callee.
2345 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2346 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2348 2014-04-04 Jeff Law <law@redhat.com>
2350 PR target/60657
2351 * config/arm/predicates.md (const_int_I_operand): New predicate.
2352 (const_int_M_operand): Similarly.
2353 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2354 const_int_operand.
2355 (insv_t2, extv_reg, extzv_t2): Likewise.
2356 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2357 (pop_multiple_with_writeback_and_return): Likewise.
2358 (vfp_pop_multiple_with_writeback): Likewise
2360 2014-04-04 Richard Biener <rguenther@suse.de>
2362 PR ipa/60746
2363 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2364 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2365 non-GIMPLE_LABELs.
2366 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2367 * gimplify.c (gimple_add_tmp_var_fn): New function.
2368 * gimple-expr.h (create_tmp_reg_fn): Declare.
2369 * gimple-expr.c (create_tmp_reg_fn): New function.
2370 * gimple-low.c (record_vars_into): Don't change cfun.
2371 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2372 code generation without cfun.
2374 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2376 PR bootstrap/60719
2377 * Makefile.in (install-driver): Fix shell scripting.
2379 2014-04-03 Cong Hou <congh@google.com>
2381 PR tree-optimization/60505
2382 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2383 threshold of number of iterations below which no vectorization
2384 will be done.
2385 * tree-vect-loop.c (new_loop_vec_info):
2386 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2387 * tree-vect-loop.c (vect_analyze_loop_operations):
2388 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2389 * tree-vect-loop.c (vect_transform_loop):
2390 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2391 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2392 of iterations of the loop and see if we should build the epilogue.
2394 2014-04-03 Richard Biener <rguenther@suse.de>
2396 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2397 (streamer_tree_cache_create): Adjust.
2398 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2399 to allow optional nodes array.
2400 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2401 (streamer_tree_cache_append): Likewise.
2402 (streamer_tree_cache_create): Create nodes array optionally
2403 as specified by parameter.
2404 * lto-streamer-out.c (create_output_block): Avoid maintaining
2405 the node array in the writer cache.
2406 (DFS_write_tree): Remove assertion.
2407 (produce_asm_for_decls): Free the out decl state hash table early.
2408 * lto-streamer-in.c (lto_data_in_create): Adjust for
2409 streamer_tree_cache_create prototype change.
2411 2014-04-03 Richard Biener <rguenther@suse.de>
2413 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2414 set TREE_CHAIN to NULL_TREE.
2416 2014-04-03 Richard Biener <rguenther@suse.de>
2418 PR tree-optimization/60740
2419 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2420 over all GIMPLE_COND operands.
2422 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2424 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2425 (Weffc++): Remove Scott's numbering, merge lists and reference
2426 Wnon-virtual-dtor.
2428 2014-04-03 Nick Clifton <nickc@redhat.com>
2430 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2431 properly.
2433 2014-04-03 Martin Jambor <mjambor@suse.cz>
2435 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2436 mention gcc_unreachable before failing.
2437 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2438 removed symbols.
2440 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2442 PR ipa/60659
2443 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2444 inconsistent code and instead mark the context inconsistent.
2445 (possible_polymorphic_call_targets): For inconsistent contexts
2446 return empty complete list.
2448 2014-04-02 Anthony Green <green@moxielogic.com>
2450 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2451 (extendqisi2, extendhisi2): Define.
2452 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2453 (WCHAR_TYPE): Change to unsigned int.
2455 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2457 PR tree-optimization/60733
2458 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2459 insertion point for PHI candidates to be the end of the feeding
2460 block for the PHI argument.
2462 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2464 PR rtl-optimization/60650
2465 * lra-constraints.c (process_alt_operands): Decrease reject for
2466 earlyclobber matching.
2468 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2470 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2472 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2474 * config/spu/spu.c (pad_bb): Do not crash when the last
2475 insn is CODE_FOR_blockage.
2477 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2479 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2480 lies outside the target mode.
2482 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2484 PR target/60735
2485 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2486 software floating point or no floating point registers, do not
2487 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2488 in GPRs that occurs after we tested for GPRs that would never be
2489 true.
2491 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2492 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2493 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2494 specifically allow DDmode, since that does not use the SPE SIMD
2495 instructions.
2497 2014-04-02 Richard Biener <rguenther@suse.de>
2499 PR middle-end/60729
2500 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2501 MODE_INTs. Properly use negv_optab.
2502 (expand_abs): Likewise.
2504 2014-04-02 Richard Biener <rguenther@suse.de>
2506 PR bootstrap/60719
2507 * Makefile.in (install-driver): Guard extra installs with special
2508 names properly.
2510 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2512 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2513 Document vec_vgbbd.
2515 2014-04-01 Richard Henderson <rth@redhat.com>
2517 PR target/60704
2518 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2519 alternative enabled before register allocation.
2521 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2523 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2524 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2525 typo.
2526 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2527 (nios2_got_address): Update nios2_large_got_address call site.
2528 (nios2_delegitimize_address): New function.
2529 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2530 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2531 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2533 2014-04-01 Martin Husemann <martin@duskware.de>
2535 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2536 for -mabi=32.
2538 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2540 PR rtl-optimization/60604
2541 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2542 check from register_operand.
2543 (register_operand): Redefine in terms of general_operand.
2544 (nonmemory_operand): Use register_operand for the non-constant cases.
2546 2014-04-01 Richard Biener <rguenther@suse.de>
2548 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2550 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2552 * doc/invoke.texi (mapp-regs): Clarify.
2554 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2556 * config/i386/avx512fintrin.h (__v32hi): Define type.
2557 (__v64qi): Likewise.
2558 (_mm512_set1_epi8): Define.
2559 (_mm512_set1_epi16): Define.
2560 (_mm512_set4_epi32): Define.
2561 (_mm512_set4_epi64): Define.
2562 (_mm512_set4_pd): Define.
2563 (_mm512_set4_ps): Define.
2564 (_mm512_setr4_epi64): Define.
2565 (_mm512_setr4_epi32): Define.
2566 (_mm512_setr4_pd): Define.
2567 (_mm512_setr4_ps): Define.
2568 (_mm512_setzero_epi32): Define.
2570 2014-03-31 Martin Jambor <mjambor@suse.cz>
2572 PR middle-end/60647
2573 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2574 callsite_arguments_match_p. Updated all callers. Also check types of
2575 corresponding formal parameters and actual arguments.
2576 (not_all_callers_have_enough_arguments_p) Renamed to
2577 some_callers_have_mismatched_arguments_p.
2579 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2581 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2583 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2585 PR target/60034
2586 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2587 section anchor.
2589 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2591 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2592 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2593 Split out
2594 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2595 Use FMAMODE_NOVF512 mode iterator.
2596 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2597 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2598 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2599 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2600 Split out
2601 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2602 Use VF_128_256 mode iterator.
2603 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2604 Ditto.
2606 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2608 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2609 static chain if needed.
2611 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2613 PR target/60697
2614 * lra-constraints.c (index_part_to_reg): New.
2615 (process_address): Use it.
2617 2014-03-27 Jeff Law <law@redhat.com>
2618 Jakub Jelinek <jakub@redhat.com>
2620 PR target/60648
2621 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2622 gen_rtx_{PLUS,MULT} to build up the address expression.
2624 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2625 creating non-canonical RTL.
2627 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2629 PR ipa/60243
2630 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2631 functions; reorganize to make cheap checks first.
2632 (inline_small_functions): Do not estimate growth when dumping;
2633 it is expensive.
2634 * ipa-inline.h (inline_summary): Add min_size.
2635 (growth_likely_positive): New function.
2636 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2637 (set_cond_stmt_execution_predicate): Cleanup.
2638 (estimate_edge_size_and_time): Compute min_size.
2639 (estimate_calls_size_and_time): Likewise.
2640 (estimate_node_size_and_time): Likewise.
2641 (inline_update_overall_summary): Update min_size.
2642 (do_estimate_edge_time): Likewise.
2643 (do_estimate_edge_size): Update.
2644 (do_estimate_edge_hints): Update.
2645 (growth_likely_positive): New function.
2647 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2649 PR target/60693
2650 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2651 also if addr has VOIDmode.
2653 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2655 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2656 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2657 Declare extern.
2658 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2659 instructions as well as AdvancedSIMD loads.
2661 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2663 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2664 Use crypto_aese type.
2665 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2666 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2667 crypto_aese, crypto_aesmc. Move to types.md.
2668 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2669 crypto_aesmc.
2670 * config/arm/iterators.md (crypto_type): Likewise.
2672 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2674 * cgraph.c: Include expr.h and tree-dfa.h.
2675 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2676 remove LHS.
2678 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2680 PR target/60675
2681 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2682 regs from checking multi-reg pseudos.
2684 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2686 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2688 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2690 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2691 if it would clobber the stack pointer, even temporarily.
2693 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2695 * mode-switching.c: Make small adjustments to the top comment.
2697 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2699 * config/rs6000/constraints.md (wD constraint): New constraint to
2700 match the constant integer to get the top DImode/DFmode out of a
2701 vector in a VSX register.
2703 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2704 match the constant integer to get the top DImode/DFmode out of a
2705 vector in a VSX register.
2707 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2708 for ISA 2.07.
2710 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2711 vbpermq builtins.
2713 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2714 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2716 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2717 Optimize vec_extract of 64-bit values, where the value being
2718 extracted is in the top word, where we can use scalar
2719 instructions. Add direct move and store support. Combine the big
2720 endian/little endian vector select load support into a single insn.
2721 (vsx_extract_<mode>_internal1): Likewise.
2722 (vsx_extract_<mode>_internal2): Likewise.
2723 (vsx_extract_<mode>_load): Likewise.
2724 (vsx_extract_<mode>_store): Likewise.
2725 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2726 combined into vsx_extract_<mode>_load.
2727 (vsx_extract_<mode>_one_le): Likewise.
2729 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2730 define the top 64-bit vector element.
2732 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2733 constraint.
2735 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2736 Document vec_vbpermq builtin.
2738 PR target/60672
2739 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2740 enable use of xxsldwi and xxpermdi builtin functions.
2741 (vec_xxpermdi): Likewise.
2743 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2744 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2746 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2748 PR rtl-optimization/60650
2749 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2750 first_p. Use it.
2751 (find_spills_for): New.
2752 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2753 Spill all pseudos on the second iteration.
2755 2014-03-27 Marek Polacek <polacek@redhat.com>
2757 PR c/50347
2758 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2759 types.
2761 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2763 * config/s390/s390.c (s390_can_use_return_insn): Check for
2764 call-saved FPRs on 31 bit.
2766 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2768 PR middle-end/60682
2769 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2770 if they need regimplification, just drop them instead of
2771 calling gimple_regimplify_operands on them.
2773 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2775 PR target/60580
2776 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
2777 (aarch64_frame_pointer_required): Adjust logic.
2778 (aarch64_can_eliminate): Adjust logic.
2779 (aarch64_override_options_after_change): Adjust logic.
2781 2014-03-27 Dehao Chen <dehao@google.com>
2783 * ipa-inline.c (early_inliner): Update node's inline info.
2785 2014-03-26 Dehao Chen <dehao@google.com>
2787 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
2788 compiler inserted conditional jumps for NAN float check.
2790 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2792 * ubsan.h (ubsan_create_data): Change second argument's type
2793 to const location_t *.
2794 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
2795 _("<unknown>").
2796 (ubsan_create_data): Change second argument to const location_t *PLOC.
2797 Create Loc field whenever PLOC is non-NULL.
2798 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
2799 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
2800 callers.
2802 PR other/59545
2803 * real.c (real_to_integer2): Change type of low to UHWI.
2805 2014-03-26 Tobias Burnus <burnus@net-b.de>
2807 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
2808 (CILK_SELF_SPECS): New define.
2809 (driver_self_specs): Use it.
2811 2014-03-26 Richard Biener <rguenther@suse.de>
2813 * tree-pretty-print.c (percent_K_format): Implement special
2814 case for LTO and its stripped down BLOCK tree.
2816 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2818 PR sanitizer/60636
2819 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
2821 * tree-vrp.c (simplify_internal_call_using_ranges): If only
2822 one range is range_int_cst_p, but not both, at least optimize
2823 addition/subtraction of 0 and multiplication by 0 or 1.
2824 * gimple-fold.c (gimple_fold_call): Fold
2825 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
2826 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
2827 INTEGER_CSTs, try to fold at least x * 0 and y - y.
2829 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
2831 PR rtl-optimization/60452
2832 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
2833 <case REG>: Return 1 for invalid offsets from the frame pointer.
2835 2014-03-26 Marek Polacek <polacek@redhat.com>
2837 PR c/37428
2838 * doc/extend.texi (C Extensions): Mention variable-length arrays in
2839 a structure/union.
2841 2014-03-26 Marek Polacek <polacek@redhat.com>
2843 PR c/39525
2844 * doc/extend.texi (Designated Inits): Describe what happens to omitted
2845 field members.
2847 2014-03-26 Marek Polacek <polacek@redhat.com>
2849 PR other/59545
2850 * ira-color.c (update_conflict_hard_regno_costs): Perform the
2851 multiplication in unsigned type.
2853 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2855 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
2857 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2859 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
2861 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2863 PR ipa/60315
2864 * cif-code.def (UNREACHABLE) New code.
2865 * ipa-inline.c (inline_small_functions): Skip edges to
2866 __builtlin_unreachable.
2867 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
2868 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
2869 predicate to __bulitin_unreachable.
2870 (set_cond_stmt_execution_predicate): Fix issue when
2871 invert_tree_comparison returns ERROR_MARK.
2872 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
2873 propagate to inline clones.
2874 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
2875 to unreachable.
2876 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
2877 * cgraphclones.c (cgraph_clone_node): If call destination is already
2878 ureachable, do not redirect it back.
2879 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
2880 unreachable.
2882 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2884 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
2885 Do not modify inline clones.
2887 2014-03-25 Jakub Jelinek <jakub@redhat.com>
2889 * config/i386/i386.md (general_sext_operand): New mode attr.
2890 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
2891 don't generate (sign_extend (const_int)).
2892 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
2893 operands[2]. Use We constraint instead of <i> and
2894 <general_sext_operand> predicate instead of <general_operand>.
2895 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
2896 * config/i386/constraints.md (We): New constraint.
2897 * config/i386/predicates.md (x86_64_sext_operand,
2898 sext_operand): New predicates.
2900 2014-03-25 Martin Jambor <mjambor@suse.cz>
2902 PR ipa/60600
2903 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
2904 inconsistent devirtualizations to __builtin_unreachable.
2906 2014-03-25 Marek Polacek <polacek@redhat.com>
2908 PR c/35449
2909 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
2911 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
2913 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
2914 order of elements for big-endian.
2916 2014-03-25 Richard Biener <rguenther@suse.de>
2918 PR middle-end/60635
2919 * gimplify-me.c (gimple_regimplify_operands): Update the
2920 re-gimplifed stmt.
2922 2014-03-25 Martin Jambor <mjambor@suse.cz>
2924 PR ipa/59176
2925 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
2926 (lto_output_varpool_node): Likewise.
2927 (input_overwrite_node): Likewise.
2928 (input_varpool_node): Likewise.
2930 2014-03-25 Richard Biener <rguenther@suse.de>
2932 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
2933 (run_gcc): Likewise.
2935 2014-03-25 Jakub Jelinek <jakub@redhat.com>
2937 * combine.c (simplify_compare_const): Add MODE argument.
2938 Handle mode_width 0 as very large mode_width.
2939 (try_combine, simplify_comparison): Adjust callers.
2941 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
2942 type to avoid signed integer overflow.
2943 * explow.c (plus_constant): Likewise.
2945 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
2947 * doc/generic.texi: Correct typos.
2949 2014-03-24 Tobias Burnus <burnus@net-b.de>
2951 * doc/invoke.texi (-flto): Expand section about
2952 using static libraries with LTO.
2954 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2956 PR rtl-optimization/60501
2957 * optabs.def (addptr3_optab): New optab.
2958 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
2959 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
2960 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
2962 * lra.c (emit_add3_insn): Use the addptr pattern if available.
2964 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
2966 2014-03-24 Ulrich Drepper <drepper@gmail.com>
2968 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
2969 _mm512_set1_pd.
2971 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
2972 (_mm256_undefined_ps): Define.
2973 (_mm256_undefined_pd): Define.
2974 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
2975 (_mm_undefined_pd): Define.
2976 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
2977 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
2978 (_mm512_undefined_ps): Define.
2979 (_mm512_undefined_pd): Define.
2980 Use _mm*_undefined_*.
2981 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
2983 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
2985 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
2986 (lshr_simd): DI mode added.
2987 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
2988 (aarch64_ushr_simddi): Likewise.
2989 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
2990 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
2991 (vshrd_n_u64): Likewise.
2993 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2995 * Makefile.in (s-macro_list): Depend on cc1.
2997 2014-03-23 Teresa Johnson <tejohnson@google.com>
2999 * ipa-utils.c (ipa_print_order): Use specified dump file.
3001 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3003 PR rtl-optimization/60601
3004 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3006 * gcc.c (eval_spec_function): Initialize save_growing_value.
3008 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3010 PR sanitizer/60613
3011 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3012 code == MINUS_EXPR, never swap op0 with op1.
3014 * toplev.c (init_local_tick): Avoid signed integer multiplication
3015 overflow.
3016 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3017 shift by first operand's bitsize.
3019 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3021 PR target/60610
3022 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3023 redefine to 1 or 0.
3024 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3025 TARGET_ISA_64BIT_P(x).
3027 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3029 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3030 pattern for vector nor instead of subtract from splat(-1).
3031 (altivec_expand_vec_perm_const_le): Likewise.
3033 2014-03-21 Richard Henderson <rth@twiddle.net>
3035 PR target/60598
3036 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3037 related insns after epilogue_completed.
3039 2014-03-21 Martin Jambor <mjambor@suse.cz>
3041 PR ipa/59176
3042 * cgraph.h (symtab_node): New flag body_removed.
3043 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3044 when removing bodies.
3045 * symtab.c (dump_symtab_base): Dump body_removed flag.
3046 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3047 had their bodies removed.
3049 2014-03-21 Martin Jambor <mjambor@suse.cz>
3051 PR ipa/60419
3052 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3053 in the border.
3055 2014-03-21 Richard Biener <rguenther@suse.de>
3057 PR tree-optimization/60577
3058 * tree-core.h (struct tree_base): Document nothrow_flag use
3059 in VAR_DECL_NONALIASED.
3060 * tree.h (VAR_DECL_NONALIASED): New.
3061 (may_be_aliased): Adjust.
3062 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3064 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3066 * expr.c (expand_expr_real_1): Remove outdated comment.
3068 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3070 PR middle-end/60597
3071 * ira.c (adjust_cleared_regs): Call copy_rtx on
3072 *reg_equiv[REGNO (loc)].src_p before passing it to
3073 simplify_replace_fn_rtx.
3075 PR target/60568
3076 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3077 into CONST, put pic register as first operand of PLUS. Use
3078 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3080 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3082 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3084 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3086 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3087 around for store forwarding issue in the FPU on the UT699.
3088 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3089 loads and operations if -mfix-ut699 is specified.
3090 (divtf3_hq): Tweak attribute.
3091 (sqrttf2_hq): Likewise.
3093 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3095 * calls.c (store_one_arg): Remove incorrect const qualification on the
3096 type of the temporary.
3097 * cfgexpand.c (expand_return): Likewise.
3098 * expr.c (expand_constructor): Likewise.
3099 (expand_expr_real_1): Likewise.
3101 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3103 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3104 of parts.
3106 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3108 PR target/60039
3109 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3111 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3113 * config/arm/aarch-common-protos.h
3114 (alu_cost_table): Fix spelling of "extend".
3115 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3117 2014-03-19 Richard Biener <rguenther@suse.de>
3119 PR middle-end/60553
3120 * tree-core.h (tree_type_common): Re-order pointer members
3121 to reduce recursion depth during GC walks.
3123 2014-03-19 Marek Polacek <polacek@redhat.com>
3125 PR sanitizer/60569
3126 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3127 before accessing it.
3129 2014-03-19 Richard Biener <rguenther@suse.de>
3131 PR lto/59543
3132 * lto-streamer-in.c (input_function): In WPA stage do not drop
3133 debug stmts.
3135 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3137 PR tree-optimization/60559
3138 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3139 with build_zero_cst assignment.
3141 2014-03-18 Kai Tietz <ktietz@redhat.com>
3143 PR rtl-optimization/56356
3144 * sdbout.c (sdbout_parms): Verify that parms'
3145 incoming argument is valid.
3146 (sdbout_reg_parms): Likewise.
3148 2014-03-18 Richard Henderson <rth@redhat.com>
3150 PR target/60562
3151 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3152 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3153 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3155 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3157 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3158 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3159 Italicize plugin event names in description. Explain that
3160 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3161 Remind that no GCC functions should be called after PLUGIN_FINISH.
3162 Explain what pragmas with expansion are.
3164 2014-03-18 Martin Liska <mliska@suse.cz>
3166 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3167 gimple call statement is update.
3168 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3169 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3171 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3173 PR sanitizer/60557
3174 * ubsan.c (ubsan_instrument_unreachable): Call
3175 initialize_sanitizer_builtins.
3176 (ubsan_pass): Likewise.
3178 PR sanitizer/60535
3179 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3180 varpool_finalize_decl instead of rest_of_decl_compilation.
3182 2014-03-18 Richard Biener <rguenther@suse.de>
3184 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3185 by using bitmap_and_compl instead of bitmap_and_compl_into.
3186 (df_rd_transfer_function): Likewise.
3188 2014-03-18 Richard Biener <rguenther@suse.de>
3190 * doc/lto.texi (fresolution): Fix typo.
3192 2014-03-18 Richard Biener <rguenther@suse.de>
3194 * doc/invoke.texi (flto): Update for changes in 4.9.
3196 2014-03-18 Richard Biener <rguenther@suse.de>
3198 * doc/loop.texi: Remove section on the removed lambda framework.
3199 Update loop docs with recent changes in preserving loop structure.
3201 2014-03-18 Richard Biener <rguenther@suse.de>
3203 * doc/lto.texi (-fresolution): Document.
3205 2014-03-18 Richard Biener <rguenther@suse.de>
3207 * doc/contrib.texi: Adjust my name.
3209 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3211 PR ipa/58721
3212 * internal-fn.c: Include diagnostic-core.h.
3213 (expand_BUILTIN_EXPECT): New function.
3214 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3215 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3216 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3217 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3218 IFN_BUILTIN_EXPECT.
3219 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3220 Revert 3 argument __builtin_expect code.
3221 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3222 * gimple-fold.c (gimple_fold_call): Likewise.
3223 * tree.h (fold_builtin_expect): New prototype.
3224 * builtins.c (build_builtin_expect_predicate): Add predictor
3225 argument, if non-NULL, create 3 argument __builtin_expect.
3226 (fold_builtin_expect): No longer static. Add ARG2 argument,
3227 pass it through to build_builtin_expect_predicate.
3228 (fold_builtin_2): Adjust caller.
3229 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3230 * internal-fn.def (BUILTIN_EXPECT): New.
3232 2014-03-18 Tobias Burnus <burnus@net-b.de>
3234 PR ipa/58721
3235 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3236 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3237 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3239 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3241 PR ipa/58721
3242 * predict.c (combine_predictions_for_bb): Fix up formatting.
3243 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3244 fill what it points to if non-NULL.
3245 (tree_predict_by_opcode): Adjust caller, use the predictor.
3246 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3248 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3250 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3251 proper constant for the store mode.
3253 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3255 * symtab.c (change_decl_assembler_name): Fix transparent alias
3256 chain construction.
3258 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3260 * config/aarch64/aarch64.c: Correct the comments about the
3261 aarch64 stack layout.
3263 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3265 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3266 check for GF_OMP_FOR_KIND_FOR.
3268 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3270 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3271 ymm and zmm register names.
3273 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3275 PR target/60516
3276 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3277 note creation for the 2010-08-31 changes.
3279 2014-03-17 Marek Polacek <polacek@redhat.com>
3281 PR middle-end/60534
3282 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3283 as -fno-tree-loop-vectorize.
3284 (expand_omp_simd): Likewise.
3286 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3288 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3289 (eligible_for_call_delay): New prototype.
3290 * config/sparc/sparc.c (tls_call_delay): Rename into...
3291 (eligible_for_call_delay): ...this. Return false if the instruction
3292 cannot be put in the delay slot of a branch.
3293 (eligible_for_restore_insn): Simplify.
3294 (eligible_for_return_delay): Return false if the instruction cannot be
3295 put in the delay slot of a branch and simplify.
3296 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3297 put in the delay slot of a branch.
3298 * config/sparc/sparc.md (fix_ut699): New attribute.
3299 (tls_call_delay): Delete.
3300 (in_call_delay): Reimplement.
3301 (eligible_for_sibcall_delay): Rename into...
3302 (in_sibcall_delay): ...this.
3303 (eligible_for_return_delay): Rename into...
3304 (in_return_delay): ...this.
3305 (in_branch_delay): Reimplement.
3306 (in_uncond_branch_delay): Delete.
3307 (in_annul_branch_delay): Delete.
3309 2014-03-14 Richard Henderson <rth@redhat.com>
3311 PR target/60525
3312 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3313 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3314 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3315 (floathi splitters): Remove.
3316 (float<SWI48x>xf2): New pattern.
3317 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3318 code that tried to handle DImode for 32-bit, but which was excluded
3319 by the pattern's condition. Drop allocation of stack temporary.
3320 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3321 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3322 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3323 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3324 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3325 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3326 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3327 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3328 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3329 (*float<SWI48x><X87MODEF>2_i387): Remove.
3330 (all float _with_temp splitters): Remove.
3331 (*float<SWI48x><MODEF>2_i387): New pattern.
3332 (*float<SWI48><MODEF>2_sse): New pattern.
3333 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3334 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3336 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3337 Marek Polacek <polacek@redhat.com>
3339 PR middle-end/60484
3340 * common.opt (dump_base_name_prefixed): New Variable.
3341 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3342 if x_dump_base_name_prefixed is already set, set it at the end.
3344 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3346 PR rtl-optimization/60508
3347 * lra-constraints.c (get_reload_reg): Add new parameter
3348 in_subreg_p.
3349 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3350 Pass the new parameter values.
3352 2014-03-14 Richard Biener <rguenther@suse.de>
3354 * common.opt: Revert unintented changes from r205065.
3355 * opts.c: Likewise.
3357 2014-03-14 Richard Biener <rguenther@suse.de>
3359 PR middle-end/60518
3360 * cfghooks.c (split_block): Properly adjust all loops the
3361 block was a latch of.
3363 2014-03-14 Martin Jambor <mjambor@suse.cz>
3365 PR lto/60461
3366 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3367 and simplify it.
3369 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3371 PR target/59396
3372 * config/avr/avr.c (avr_set_current_function): Pass function name
3373 through default_strip_name_encoding before sanity checking instead
3374 of skipping the first char of the assembler name.
3376 2014-03-13 Richard Henderson <rth@redhat.com>
3378 PR debug/60438
3379 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3380 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3381 * config/i386/i386-protos.h: Likewise.
3382 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3383 in the expander instead of a splitter.
3384 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3385 any possibility of requiring a memory.
3386 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3387 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3388 (fp branch splitters): Update for ix86_split_fp_branch.
3389 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3390 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3391 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3392 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3393 (*fop_<MODEF>_3_i387): Likewise.
3394 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3395 (splitters for the fop_* register patterns): Remove.
3396 (fscalexf4_i387): Rename from *fscalexf4_i387.
3397 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3399 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3401 PR tree-optimization/59779
3402 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3403 type for bitsize and maxsize instead of HOST_WIDE_INT.
3405 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3407 PR rtl-optimization/57320
3408 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3409 the CFG after thread_prologue_and_epilogue_insns.
3411 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3413 PR rtl-optimization/57189
3414 * lra-constraints.c (process_alt_operands): Disfavor spilling
3415 vector pseudos.
3417 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3419 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3421 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3423 PR tree-optimization/59025
3424 PR middle-end/60418
3425 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3426 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3428 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3430 PR target/60486
3431 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3432 calls of avr_out_plus_1.
3434 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3436 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3437 BB's single pred and update the father loop's latch info later.
3439 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3441 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3442 (VEC_M): Likewise.
3443 (VEC_N): Likewise.
3444 (VEC_R): Likewise.
3445 (VEC_base): Likewise.
3446 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3447 registers, we need to swap double words in little endian mode.
3449 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3450 to be a container mode for 128-bit integer operations added in ISA
3451 2.07. Unlike TImode and PTImode, the preferred register set is
3452 the Altivec/VMX registers for the 128-bit operations.
3454 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3455 declarations.
3456 (rs6000_split_128bit_ok_p): Likewise.
3458 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3459 macros for creating ISA 2.07 normal and overloaded builtin
3460 functions with 3 arguments.
3461 (BU_P8V_OVERLOAD_3): Likewise.
3462 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3463 for use as overloaded functions.
3464 (VPERM_1TI_UNS): Likewise.
3465 (VSEL_1TI): Likewise.
3466 (VSEL_1TI_UNS): Likewise.
3467 (ST_INTERNAL_1ti): Likewise.
3468 (LD_INTERNAL_1ti): Likewise.
3469 (XXSEL_1TI): Likewise.
3470 (XXSEL_1TI_UNS): Likewise.
3471 (VPERM_1TI): Likewise.
3472 (VPERM_1TI_UNS): Likewise.
3473 (XXPERMDI_1TI): Likewise.
3474 (SET_1TI): Likewise.
3475 (LXVD2X_V1TI): Likewise.
3476 (STXVD2X_V1TI): Likewise.
3477 (VEC_INIT_V1TI): Likewise.
3478 (VEC_SET_V1TI): Likewise.
3479 (VEC_EXT_V1TI): Likewise.
3480 (EQV_V1TI): Likewise.
3481 (NAND_V1TI): Likewise.
3482 (ORC_V1TI): Likewise.
3483 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3484 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3485 overloaded builtin.
3486 (VADDUQM): Likewise.
3487 (VSUBCUQ): Likewise.
3488 (VADDEUQM): Likewise.
3489 (VADDECUQ): Likewise.
3490 (VSUBEUQM): Likewise.
3491 (VSUBECUQ): Likewise.
3493 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3494 __int128_t and __uint128_t types.
3495 (__uint128_type): Likewise.
3496 (altivec_categorize_keyword): Add support for vector __int128_t,
3497 vector __uint128_t, vector __int128, and vector unsigned __int128
3498 as a container type for TImode operations that need to be done in
3499 VSX/Altivec registers.
3500 (rs6000_macro_to_expand): Likewise.
3501 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3502 to support 128-bit integer instructions vaddcuq, vadduqm,
3503 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3504 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3506 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3507 for V1TImode, and set up preferences to use VSX/Altivec registers.
3508 Setup VSX reload handlers.
3509 (rs6000_debug_reg_global): Likewise.
3510 (rs6000_init_hard_regno_mode_ok): Likewise.
3511 (rs6000_preferred_simd_mode): Likewise.
3512 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3513 (easy_altivec_constant): Likewise.
3514 (output_vec_const_move): Likewise.
3515 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3516 simple move.
3517 (rs6000_expand_vector_extract): Likewise.
3518 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3519 addressing.
3520 (rs6000_const_vec): Add support for V1TImode.
3521 (rs6000_emit_le_vsx_load): Swap double words when loading or
3522 storing TImode/V1TImode.
3523 (rs6000_emit_le_vsx_store): Likewise.
3524 (rs6000_emit_le_vsx_move): Likewise.
3525 (rs6000_emit_move): Add support for V1TImode.
3526 (altivec_expand_ld_builtin): Likewise.
3527 (altivec_expand_st_builtin): Likewise.
3528 (altivec_expand_vec_init_builtin): Likewise.
3529 (altivec_expand_builtin): Likewise.
3530 (rs6000_init_builtins): Add support for V1TImode type. Add
3531 support for ISA 2.07 128-bit integer builtins. Define type names
3532 for the VSX/Altivec vector types.
3533 (altivec_init_builtins): Add support for overloaded vector
3534 functions with V1TImode type.
3535 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3536 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3537 external function.
3538 (rs6000_split_128bit_ok_p): Likewise.
3539 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3540 __int128_t and vector __uint128_t.
3542 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3543 and mode attributes.
3544 (VSX_M): Likewise.
3545 (VSX_M2): Likewise.
3546 (VSm): Likewise.
3547 (VSs): Likewise.
3548 (VSr): Likewise.
3549 (VSv): Likewise.
3550 (VS_scalar): Likewise.
3551 (VS_double): Likewise.
3552 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3554 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3555 we support the ISA 2.07 128-bit integer arithmetic instructions.
3556 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3557 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3558 and TImode types for use with the builtin functions.
3559 (V1TI_type_node): Likewise.
3560 (unsigned_V1TI_type_node): Likewise.
3561 (intTI_type_internal_node): Likewise.
3562 (uintTI_type_internal_node): Likewise.
3564 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3565 128-bit builtin functions.
3566 (UNSPEC_VADDEUQM): Likewise.
3567 (UNSPEC_VADDECUQ): Likewise.
3568 (UNSPEC_VSUBCUQ): Likewise.
3569 (UNSPEC_VSUBEUQM): Likewise.
3570 (UNSPEC_VSUBECUQ): Likewise.
3571 (VM): Add V1TImode to vector mode iterators.
3572 (VM2): Likewise.
3573 (VI_unit): Likewise.
3574 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3575 (altivec_vaddcuq): Likewise.
3576 (altivec_vsubuqm): Likewise.
3577 (altivec_vsubcuq): Likewise.
3578 (altivec_vaddeuqm): Likewise.
3579 (altivec_vaddecuq): Likewise.
3580 (altivec_vsubeuqm): Likewise.
3581 (altivec_vsubecuq): Likewise.
3583 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3584 mode iterators.
3585 (BOOL_128): Likewise.
3586 (BOOL_REGS_OUTPUT): Likewise.
3587 (BOOL_REGS_OP1): Likewise.
3588 (BOOL_REGS_OP2): Likewise.
3589 (BOOL_REGS_UNARY): Likewise.
3590 (BOOL_REGS_AND_CR0): Likewise.
3592 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3593 128-bit integer builtin support.
3594 (vec_vadduqm): Likewise.
3595 (vec_vaddecuq): Likewise.
3596 (vec_vaddeuqm): Likewise.
3597 (vec_vsubecuq): Likewise.
3598 (vec_vsubeuqm): Likewise.
3599 (vec_vsubcuq): Likewise.
3600 (vec_vsubuqm): Likewise.
3602 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3603 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3604 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3605 128-bit integer add/subtract to ISA 2.07.
3607 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3609 * config/arc/arc.c (arc_predicate_delay_insns):
3610 Fix third argument passed to conditionalize_nonjump.
3612 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3614 * config/aarch64/aarch64-builtins.c
3615 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3616 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3617 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3618 instead of __builtin_lfloor.
3619 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3621 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3623 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3624 (tree_ssa_ifcombine_bb_1): New function.
3625 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3626 is an empty forwarder block to then_bb or vice versa and then_bb
3627 and else_bb are effectively swapped.
3629 2014-03-12 Christian Bruel <christian.bruel@st.com>
3631 PR target/60264
3632 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3633 REG_CFA_DEF_CFA note.
3634 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3635 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3637 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3639 PR tree-optimization/60454
3640 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3642 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3644 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3645 Do not define target_cpu_default2 to generic.
3646 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3647 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3648 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3650 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3651 Marc Glisse <marc.glisse@inria.fr>
3653 PR tree-optimization/60502
3654 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3655 instead of build_low_bits_mask.
3657 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3659 PR middle-end/60482
3660 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3661 if there are multiple uses, but op doesn't live on E edge.
3662 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3663 clobber stmts before __builtin_unreachable.
3665 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3667 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3668 hard_frame_pointer_rtx.
3669 * cse.c (cse_insn): Remove volatile check.
3670 * cselib.c (cselib_process_insn): Likewise.
3671 * dse.c (scan_insn): Likewise.
3673 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3675 * config/arc/arc.c (conditionalize_nonjump): New function,
3676 broken out of ...
3677 (arc_ifcvt): ... this.
3678 (arc_predicate_delay_insns): Use it.
3680 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3682 * config/arc/predicates.md (extend_operand): During/after reload,
3683 allow const_int_operand.
3684 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3685 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3686 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3687 to "i".
3688 (umulsi3_highpart_i): Likewise.
3690 2014-03-11 Richard Biener <rguenther@suse.de>
3692 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3693 Add asserts to guard possible wrong-code bugs.
3695 2014-03-11 Richard Biener <rguenther@suse.de>
3697 PR tree-optimization/60429
3698 PR tree-optimization/60485
3699 * tree-ssa-structalias.c (set_union_with_increment): Properly
3700 take into account all fields that overlap the shifted vars.
3701 (do_sd_constraint): Likewise.
3702 (do_ds_constraint): Likewise.
3703 (get_constraint_for_ptr_offset): Likewise.
3705 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3707 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3708 (nios2_compute_frame_layout):
3709 Add calculation of cfun->machine->fp_save_offset.
3710 (nios2_expand_prologue): Correct setting of frame pointer register
3711 in prologue.
3712 (nios2_expand_epilogue): Update recovery of stack pointer from
3713 frame pointer accordingly.
3714 (nios2_initial_elimination_offset): Update calculation of offset
3715 for eliminating to HARD_FRAME_POINTER_REGNUM.
3717 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3719 PR ipa/60457
3720 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3721 cgraph_get_create_node on VAR_DECLs.
3723 2014-03-10 Richard Biener <rguenther@suse.de>
3725 PR middle-end/60474
3726 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3728 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3730 * config/vms/vms.opt (vms_float_format): New variable.
3732 2014-03-08 Tobias Burnus <burnus@net-b.de>
3734 * doc/invoke.texi (-fcilkplus): Update implementation status.
3736 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3737 Richard Biener <rguenther@suse.de>
3739 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3740 consistently accross all TUs.
3741 (run_gcc): Enable -fshort-double automatically at link at link-time
3742 and disallow override.
3744 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3746 PR target/58271
3747 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3748 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3749 if they can't be used.
3751 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3753 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3754 for Solaris 11/x86 ld.
3755 * configure: Regenerate.
3757 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3759 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3760 (LIB_TLS_SPEC): Save as ld_tls_libs.
3761 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3762 (HAVE_AS_IX86_TLSLDM): New test.
3763 * configure, config.in: Regenerate.
3764 * config/i386/i386.c (legitimize_tls_address): Fall back to
3765 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3766 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3767 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3768 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3770 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3772 * common.opt (fira-loop-pressure): Mark as optimization.
3774 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3776 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
3777 an OpenMP mappable type.
3779 2014-03-06 Matthias Klose <doko@ubuntu.com>
3781 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
3782 MULTILIB_OSDIRNAMES is not defined.
3784 2014-03-06 Jakub Jelinek <jakub@redhat.com>
3785 Meador Inge <meadori@codesourcery.com>
3787 PR target/58595
3788 * config/arm/arm.c (arm_tls_symbol_p): Remove.
3789 (arm_legitimize_address): Call legitimize_tls_address for any
3790 arm_tls_referenced_p expression, handle constant addend. Call it
3791 before testing for !TARGET_ARM.
3792 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
3794 2014-03-06 Richard Biener <rguenther@suse.de>
3796 PR middle-end/60445
3797 PR lto/60424
3798 PR lto/60427
3799 Revert
3800 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3802 * tree-streamer.c (record_common_node): Assert we don't record
3803 nodes with type double.
3804 (preload_common_node): Skip type double, complex double and double
3805 pointer since it is now frontend dependent due to fshort-double option.
3807 2014-03-06 Richard Biener <rguenther@suse.de>
3809 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
3810 or -fno-lto is specified and the linker has full plugin support.
3811 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
3812 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
3813 * lto-wrapper.c (merge_and_complain): Merge compile-time
3814 optimization levels.
3815 (run_gcc): And pass it through to the link options.
3817 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
3819 PR debug/60381
3820 Revert:
3821 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3822 PR debug/59992
3823 * cselib.c (remove_useless_values): Skip to avoid quadratic
3824 behavior if the condition moved from...
3825 (cselib_process_insn): ... here holds.
3827 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3829 PR plugins/59335
3830 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
3831 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
3833 PR plugins/59335
3834 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
3835 (TM_H): Add x86-tune.def.
3837 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3839 * config/aarch64/aarch64.c (generic_tunings):
3840 Use cortexa57_extra_costs.
3842 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3844 PR lto/60404
3845 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
3846 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
3847 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
3848 cost for in_lto_p.
3850 2014-03-04 Heiher <r@hev.cc>
3852 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
3853 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
3855 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
3857 * config/i386/predicates.md (const2356_operand): Change to ...
3858 (const2367_operand): ... this.
3859 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
3860 const2367_operand.
3861 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
3862 (*avx512pf_scatterpf<mode>sf): Ditto.
3863 (avx512pf_scatterpf<mode>df): Ditto.
3864 (*avx512pf_scatterpf<mode>df_mask): Ditto.
3865 (*avx512pf_scatterpf<mode>df): Ditto.
3866 * config/i386/i386.c (ix86_expand_builtin): Update
3867 incorrect hint operand error message.
3869 2014-03-04 Richard Biener <rguenther@suse.de>
3871 * lto-section-in.c (lto_get_section_data): Fix const cast.
3873 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3875 * tree-streamer.c (record_common_node): Assert we don't record
3876 nodes with type double.
3877 (preload_common_node): Skip type double, complex double and double
3878 pointer since it is now frontend dependent due to fshort-double option.
3880 2014-03-04 Richard Biener <rguenther@suse.de>
3882 PR lto/60405
3883 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
3884 (lto_input_toplevel_asms): Likewise.
3885 * lto-section-in.c (lto_get_section_data): Instead do it here
3886 for every section.
3888 2014-03-04 Richard Biener <rguenther@suse.de>
3890 PR tree-optimization/60382
3891 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
3892 dead PHIs a reduction.
3894 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
3896 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
3897 hint value.
3898 (_mm_prefetch): Move out of GCC target("sse") pragma.
3899 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
3900 GCC target("prfchw") pragma.
3901 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
3902 for locality <= 2.
3903 * config/i386/i386.c (ix86_option_override_internal): Enable
3904 -mprfchw with -mprefetchwt1.
3906 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
3908 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
3909 Mark as varying.
3911 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
3913 * opts.h (CL_PCH_IGNORE): Define.
3914 * targhooks.c (option_affects_pch_p):
3915 Return false for options that have CL_PCH_IGNORE set.
3916 * opt-functions.awk: Process PchIgnore.
3917 * doc/options.texi: Document PchIgnore.
3919 * config/arc/arc.opt (misize): Add PchIgnore property.
3921 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3923 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
3924 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
3925 constraint on constants to permit them being loaded into
3926 GENERAL_REGS or BASE_REGS.
3928 2014-03-03 Nick Clifton <nickc@redhat.com>
3930 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
3931 anti-cacnonical alternatives.
3932 (negandhi3_real): New pattern.
3933 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
3935 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
3937 * config/avr/avr-mcus.def: Remove atxmega16x1.
3938 * config/avr/avr-tables.opt: Regenerate.
3939 * config/avr/t-multilib: Regenerate.
3940 * doc/avr-mmcu.texi: Regenerate.
3942 2014-03-03 Tobias Grosser <tobias@grosser.es>
3943 Mircea Namolaru <mircea.namolaru@inria.fr>
3945 PR tree-optimization/58028
3946 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
3947 scalar dimensions.
3949 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3951 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
3952 not handled by recognizers.
3954 2014-03-03 Jakub Jelinek <jakub@redhat.com>
3956 PR middle-end/60175
3957 * function.c (expand_function_end): Don't emit
3958 clobber_return_register sequence if clobber_after is a BARRIER.
3959 * cfgexpand.c (construct_exit_block): Append instructions before
3960 return_label to prev_bb.
3962 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3964 * config/rs6000/constraints.md: Document reserved use of "wc".
3966 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
3968 PR ipa/60150
3969 * ipa.c (function_and_variable_visibility): When dissolving comdat
3970 group, also set all symbols to local.
3972 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
3974 PR ipa/60306
3976 Revert:
3977 2013-12-14 Jan Hubicka <jh@suse.cz>
3978 PR middle-end/58477
3979 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
3981 2014-03-02 Jon Beniston <jon@beniston.com>
3983 PR bootstrap/48230
3984 PR bootstrap/50927
3985 PR bootstrap/52466
3986 PR target/46898
3987 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
3988 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
3989 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
3990 (simple_return, *simple_return): New patterns
3991 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
3992 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
3994 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
3996 * dwarf2out.c (gen_subprogram_die): Tidy.
3998 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4000 PR target/60071
4001 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4002 (*mov_t_msb_neg_negc): ... this new insn.
4004 2014-02-28 Jason Merrill <jason@redhat.com>
4006 PR c++/58678
4007 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4008 function.
4010 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4012 PR c++/60314
4013 * dwarf2out.c (decltype_auto_die): New static.
4014 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4015 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4016 (is_cxx_auto): Likewise.
4018 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4020 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4021 we are not using general regs only.
4023 2014-02-28 Richard Biener <rguenther@suse.de>
4025 PR target/60280
4026 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4027 previous fix and only allow to remove trivial pre-headers
4028 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4029 (remove_forwarder_block): Properly update the latch of a loop.
4031 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4033 PR debug/59992
4034 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4035 (cselib_preserved_hash_table): New.
4036 (preserve_constants_and_equivs): Move preserved vals to it.
4037 (cselib_find_slot): Look it up first.
4038 (cselib_init): Initialize it.
4039 (cselib_finish): Release it.
4040 (dump_cselib_table): Dump it.
4042 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4044 PR debug/59992
4045 * cselib.c (remove_useless_values): Skip to avoid quadratic
4046 behavior if the condition moved from...
4047 (cselib_process_insn): ... here holds.
4049 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4051 PR debug/57232
4052 * var-tracking.c (vt_initialize): Apply the same condition to
4053 preserve the CFA base value.
4055 2014-02-28 Joey Ye <joey.ye@arm.com>
4057 PR target/PR60169
4058 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4059 if reload in progress or completed.
4061 2014-02-28 Tobias Burnus <burnus@net-b.de>
4063 PR middle-end/60147
4064 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4065 NAMELIST_DECL.
4067 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4069 * doc/tm.texi.in (Condition Code Status): Update documention for
4070 relative locations of cc0-setter and cc0-user.
4072 2014-02-27 Jeff Law <law@redhat.com>
4074 PR rtl-optimization/52714
4075 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4076 into two independent simple sets, if I3 is a jump, ensure the
4077 pattern we place into I3 is a (set (pc) ...).
4079 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4080 Jeff Law <law@redhat.com>
4082 PR rtl-optimization/49847
4083 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4084 are in different blocks.
4085 * doc/tm.texi (Condition Code Status): Update documention for
4086 relative locations of cc0-setter and cc0-user.
4088 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4090 PR target/59222
4091 * lra.c (lra_emit_add): Check SUBREG too.
4093 2014-02-27 Andreas Schwab <schwab@suse.de>
4095 * config/m68k/m68k.c (m68k_option_override): Disable
4096 -flive-range-shrinkage for classic m68k.
4097 (m68k_override_options_after_change): Likewise.
4099 2014-02-27 Marek Polacek <polacek@redhat.com>
4101 PR middle-end/59223
4102 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4103 -Wmaybe-uninitialized.
4105 2014-02-27 Alan Modra <amodra@gmail.com>
4107 PR target/57936
4108 * reload1.c (emit_input_reload_insns): When reload_override_in,
4109 set old to rl->in_reg when rl->in_reg is a subreg.
4111 2014-02-26 Richard Biener <rguenther@suse.de>
4113 PR bootstrap/60343
4114 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4116 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4118 * common/config/i386/predicates.md (const1256_operand): Remove.
4119 (const2356_operand): New.
4120 (const_1_to_2_operand): Remove.
4121 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4122 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4123 (*avx512pf_gatherpf<mode>sf): Ditto.
4124 (avx512pf_gatherpf<mode>df): Ditto.
4125 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4126 (*avx512pf_gatherpf<mode>df): Ditto.
4127 (avx512pf_scatterpf<mode>sf): Ditto.
4128 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4129 (*avx512pf_scatterpf<mode>sf): Ditto.
4130 (avx512pf_scatterpf<mode>df): Ditto.
4131 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4132 (*avx512pf_scatterpf<mode>df): Ditto.
4133 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4135 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4137 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4138 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4139 (_mm512_mask_testn_epi64_mask): Move to ...
4140 * config/i386/avx512cdintrin.h: Here.
4141 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4142 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4143 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4144 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4145 TARGET_AVX512F from TARGET_AVX512CD.
4147 2014-02-26 Richard Biener <rguenther@suse.de>
4149 PR ipa/60327
4150 * ipa.c (walk_polymorphic_call_targets): Properly guard
4151 call to inline_update_overall_summary.
4153 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4155 PR target/60280
4156 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4157 and latches only if requested. Fix latch if it is removed.
4158 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4159 LOOPS_HAVE_PREHEADERS.
4161 2014-02-25 Andrew Pinski <apinski@cavium.com>
4163 * builtins.c (expand_builtin_thread_pointer): Create a new target
4164 when the target is NULL.
4166 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4168 PR rtl-optimization/60317
4169 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4170 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4171 * lra-assigns.c: Include params.h.
4172 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4173 other reload pseudos considerations.
4175 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4177 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4178 to use canonical form for nor<mode>3.
4180 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4182 PR target/55426
4183 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4184 conversions.
4186 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4188 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4189 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4190 (ix86_handle_option): Handle OPT_mprefetchwt1.
4191 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4192 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4193 PREFETCHWT1 CPUID.
4194 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4195 OPTION_MASK_ISA_PREFETCHWT1.
4196 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4197 (PTA_PREFETCHWT1): New.
4198 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4199 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4200 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4201 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4202 (*prefetch_avx512pf_<mode>_: Change into ...
4203 (*prefetch_prefetchwt1_<mode>: This.
4204 * config/i386/i386.opt (mprefetchwt1): New.
4205 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4206 (_mm_prefetch): Handle intent to write.
4207 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4209 2014-02-25 Richard Biener <rguenther@suse.de>
4211 PR middle-end/60291
4212 * emit-rtl.c (mem_attrs_htab): Remove.
4213 (mem_attrs_htab_hash): Likewise.
4214 (mem_attrs_htab_eq): Likewise.
4215 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4216 (init_emit_once): Do not allocate mem_attrs_htab.
4218 2014-02-25 Richard Biener <rguenther@suse.de>
4220 PR lto/60319
4221 * lto-opts.c (lto_write_options): Output non-explicit conservative
4222 -fwrapv, -fno-trapv and -fno-strict-overflow.
4223 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4224 (run_gcc): And pass them through.
4226 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4228 * sel-sched.c (calculate_new_fences): New parameter ptime.
4229 Calculate it as a maximum over all fence cycles.
4230 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4231 Print the final schedule timing when sched_verbose.
4233 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4235 PR rtl-optimization/60292
4236 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4237 bit fot the fence instruction.
4239 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4241 * calls.h: Fix typo in comment.
4243 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4245 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4246 adjusting offsetable addresses.
4248 2014-02-24 Guozhi Wei <carrot@google.com>
4250 * sparseset.h (sparseset_pop): Fix the wrong index.
4252 2014-02-24 Walter Lee <walt@tilera.com>
4254 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4255 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4256 triplet.
4257 * common/config/tilegx/tilegx-common.c
4258 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4259 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4260 (LINK_SPEC): Ditto.
4261 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4262 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4263 (tilegx_gimplify_va_arg_expr): Handle big endian.
4264 (tilegx_expand_unaligned_load): Ditto.
4265 (tilegx_expand_unaligned_store): Ditto.
4266 (TARGET_RETURN_IN_MSB): New.
4267 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4268 (TARGET_ENDIAN_DEFAULT): New.
4269 (TARGET_BIG_ENDIAN): Handle big endian.
4270 (BYTES_BIG_ENDIAN): Ditto.
4271 (WORDS_BIG_ENDIAN): Ditto.
4272 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4273 (ENDIAN_SPEC): New.
4274 (EXTRA_SPECS): New.
4275 * config/tilegx/tilegx.md (extv): Handle big endian.
4276 (extzv): Ditto.
4277 (insn_st<n>): Ditto.
4278 (insn_st<n>_add<bitsuffix>): Ditto.
4279 (insn_stnt<n>): Ditto.
4280 (insn_stnt<n>_add<bitsuffix>):Ditto.
4281 (vec_interleave_highv8qi): Handle big endian.
4282 (vec_interleave_highv8qi_be): New.
4283 (vec_interleave_highv8qi_le): New.
4284 (insn_v1int_h): Handle big endian.
4285 (vec_interleave_lowv8qi): Handle big endian.
4286 (vec_interleave_lowv8qi_be): New.
4287 (vec_interleave_lowv8qi_le): New.
4288 (insn_v1int_l): Handle big endian.
4289 (vec_interleave_highv4hi): Handle big endian.
4290 (vec_interleave_highv4hi_be): New.
4291 (vec_interleave_highv4hi_le): New.
4292 (insn_v2int_h): Handle big endian.
4293 (vec_interleave_lowv4hi): Handle big endian.
4294 (vec_interleave_lowv4hi_be): New.
4295 (vec_interleave_lowv4hi_le): New.
4296 (insn_v2int_l): Handle big endian.
4297 (vec_interleave_highv2si): Handle big endian.
4298 (vec_interleave_highv2si_be): New.
4299 (vec_interleave_highv2si_le): New.
4300 (insn_v4int_h): Handle big endian.
4301 (vec_interleave_lowv2si): Handle big endian.
4302 (vec_interleave_lowv2si_be): New.
4303 (vec_interleave_lowv2si_le): New.
4304 (insn_v4int_l): Handle big endian.
4305 * config/tilegx/tilegx.opt (mbig-endian): New option.
4306 (mlittle-endian): New option.
4307 * doc/install.texi: Document tilegxbe-linux.
4308 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4310 2014-02-24 Martin Jambor <mjambor@suse.cz>
4312 PR ipa/60266
4313 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4314 there are no parameter descriptors.
4316 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4318 PR rtl-optimization/60268
4319 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4320 initialization to ...
4321 (sched_rgn_init): ... here.
4322 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4324 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4326 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4327 names.
4329 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4331 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4332 definition.
4334 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4336 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4337 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4339 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4341 * config/microblaze/predicates.md: Add cmp_op predicate.
4342 * config/microblaze/microblaze.md: Add branch_compare instruction
4343 which uses cmp_op predicate and emits cmp insn before branch.
4344 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4345 to microblaze_expand_conditional_branch and consolidate logic.
4346 (microblaze_expand_conditional_branch): emit branch_compare
4347 insn instead of handling cmp op separate from branch insn.
4349 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4351 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4352 to permit subregs.
4354 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4356 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4357 define_insn with define_expand and new define_insn
4358 *altivec_lve<VI_char>x_internal.
4359 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4360 and new define_insn *altivec_stve<VI_char>x_internal.
4361 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4362 prototype.
4363 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4364 lve*x built-ins.
4365 (altivec_expand_stvex_be): New function.
4367 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4369 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4370 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4371 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4372 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4374 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4376 PR target/60298
4377 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4378 instead of emit_move_insn.
4380 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4382 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4383 vspltw with vsldoi.
4384 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4385 gen_altivec_vsumsws.
4387 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4389 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4390 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4391 (altivec_lvxl_<mode>): New define_expand incorporating
4392 -maltivec=be semantics where needed.
4393 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4394 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4395 semantics where needed.
4396 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4397 (altivec_stvx_<mode>): New define_expand incorporating
4398 -maltivec=be semantics where needed.
4399 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4400 VM2 iterator instead of V4SI.
4401 (altivec_stvxl_<mode>): New define_expand incorporating
4402 -maltivec=be semantics where needed.
4403 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4404 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4405 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4406 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4407 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4408 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4409 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4410 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4411 ALTIVEC_BUILTIN_STVXL.
4412 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4413 (altivec_expand_stvx_be): Likewise.
4414 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4415 (altivec_expand_lvx_be): Likewise.
4416 (altivec_expand_stvx_be): Likewise.
4417 (altivec_expand_builtin): Add cases for
4418 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4419 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4420 (altivec_init_builtins): Add definitions for
4421 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4422 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4424 2014-02-21 Catherine Moore <clm@codesourcery.com>
4426 * doc/invoke.texi (mvirt, mno-virt): Document.
4427 * config/mips/mips.opt (mvirt): New option.
4428 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4430 2014-02-21 Richard Biener <rguenther@suse.de>
4432 PR tree-optimization/60276
4433 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4434 (STMT_VINFO_MIN_NEG_DIST): New macro.
4435 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4436 STMT_VINFO_MIN_NEG_DIST.
4437 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4438 made for negative dependence distances still hold.
4440 2014-02-21 Richard Biener <rguenther@suse.de>
4442 PR middle-end/60291
4443 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4444 DECL_INITIAL for globals not in the current function context.
4446 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4448 PR tree-optimization/56490
4449 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4450 * tree-ssa-uninit.c: Include params.h.
4451 (compute_control_dep_chain): Add num_calls argument, return false
4452 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4453 num_calls to recursive call.
4454 (find_predicates): Change dep_chain into normal array,
4455 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4456 variable and adjust compute_control_dep_chain caller.
4457 (find_def_preds): Likewise.
4459 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4461 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4462 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4464 2014-02-21 Nick Clifton <nickc@redhat.com>
4466 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4467 (pushhi1): Likewise.
4468 (popqi1): Add mode to pre_dec.
4469 (pophi1): Likewise.
4471 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4473 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4474 mode for mask of V8SFmode permutation.
4476 2014-02-20 Richard Henderson <rth@redhat.com>
4478 PR c++/60272
4479 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4480 a new pseudo for OLDVAL.
4482 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4484 PR target/57896
4485 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4486 gen_reg_rtx if d->testing_p.
4487 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4488 if d->testing_p and we will certainly return true.
4489 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4490 if d->testing_p.
4492 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4494 * emit-rtl.c (gen_reg_rtx): Assert that
4495 crtl->emit.regno_pointer_align_length is non-zero.
4497 2014-02-20 Richard Henderson <rth@redhat.com>
4499 PR c++/60272
4500 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4501 on failure the store back into EXPECT.
4503 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4504 Sandra Loosemore <sandra@codesourcery.com>
4506 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4507 * config/nios2/nios2.c (nios2_function_profiler): Add
4508 -fPIC (flag_pic == 2) support.
4509 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4510 (nios2_large_offset_p): New function.
4511 (nios2_unspec_reloc_p): Move up position, update to use
4512 nios2_large_offset_p.
4513 (nios2_unspec_address): Remove function.
4514 (nios2_unspec_offset): New function.
4515 (nios2_large_got_address): New function.
4516 (nios2_got_address): Add large offset support.
4517 (nios2_legitimize_tls_address): Update usage of removed and new
4518 functions.
4519 (nios2_symbol_binds_local_p): New function.
4520 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4521 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4522 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4523 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4524 processing for (const (unspec ...)).
4525 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4527 2014-02-20 Richard Biener <rguenther@suse.de>
4529 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4530 doing the substitution.
4531 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4533 2014-02-20 Martin Jambor <mjambor@suse.cz>
4535 PR ipa/55260
4536 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4537 info when checking whether lattices are bottom.
4539 2014-02-20 Richard Biener <rguenther@suse.de>
4541 PR middle-end/60221
4542 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4543 regions at -O0.
4545 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4547 PR ipa/58555
4548 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4549 parameter specifying the scaling.
4550 (inline_call): Update.
4551 (want_inline_recursively): Guard division by zero.
4552 (recursive_inlining): Update.
4553 * ipa-inline.h (clone_inlined_nodes): Update.
4555 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4557 PR target/60204
4558 * config/i386/i386.c (classify_argument): Pass structures of size
4559 64 bytes or less in register.
4561 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4562 Kirill Yukhin <kirill.yukhin@intel.com>
4564 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4565 (_mm_rcp28_round_ss): Ditto.
4566 (_mm_rsqrt28_round_sd): Ditto.
4567 (_mm_rsqrt28_round_ss): Ditto.
4568 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4569 (_mm_rcp14_round_ss): Ditto.
4570 (_mm_rsqrt14_round_sd): Ditto.
4571 (_mm_rsqrt14_round_ss): Ditto.
4572 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4573 the first input operand, get rid of match_dup.
4574 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4575 attribute to sse.
4576 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4577 Ditto.
4578 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4579 operand as the first input operand, set type attribute.
4580 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4581 Set type attribute.
4582 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4583 operand as the first input operand, set type attribute.
4585 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4587 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4588 bit of zero.
4590 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4592 PR target/60207
4593 * config/i386/i386.c (construct_container): Remove TFmode check
4594 for X86_64_INTEGER_CLASS.
4596 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4598 PR target/59794
4599 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4600 only when -Wpsabi is enabled.
4602 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4604 PR target/59799
4605 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4606 passing arrays in registers are the same as for structs, so remove the
4607 special case for them.
4609 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4611 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4612 destination type, extract only the valid bits if the source type is not
4613 integral and has a different mode.
4615 2014-02-19 Richard Biener <rguenther@suse.de>
4617 PR ipa/60243
4618 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4619 for all calls.
4621 2014-02-19 Richard Biener <rguenther@suse.de>
4623 PR ipa/60243
4624 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4625 (ipa_modify_call_arguments): Emit an argument load explicitely and
4626 preserve virtual SSA form there and for the replacement call.
4627 Do not update SSA form nor free dominance info.
4629 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4631 * ipa.c (function_and_variable_visibility): Also clear WEAK
4632 flag when disolving COMDAT_GROUP.
4634 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4636 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4637 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4638 not devirtualizing.
4639 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4640 do more sanity checks.
4641 (detect_type_change): Return true when giving up early.
4642 (compute_complex_assign_jump_func): Fix type parameter of
4643 ipa_set_ancestor_jf.
4644 (compute_complex_ancestor_jump_func): Likewise.
4645 (update_jump_functions_after_inlining): Fix updating of
4646 ancestor function.
4647 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4649 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4651 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4652 inline clones when edge disappears.
4654 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4656 PR target/60203
4657 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4658 Split 64-bit moves into 2 patterns. Do not allow the use of
4659 direct move for TDmode in little endian, since the decimal value
4660 has little endian bytes within a word, but the 64-bit pieces are
4661 ordered in a big endian fashion, and normal subreg's of TDmode are
4662 not allowed.
4663 (mov<mode>_64bit_dm): Likewise.
4664 (movtd_64bit_nodm): Likewise.
4666 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4668 PR tree-optimization/60174
4669 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4670 statement of an SSA_NAME that occurs in an abnormal PHI node.
4672 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4674 PR sanitizer/60142
4675 * final.c (SEEN_BB): Remove.
4676 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4677 (final_scan_insn): Don't force_source_line on second
4678 NOTE_INSN_BASIC_BLOCK.
4680 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4682 PR target/60205
4683 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4684 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4685 (type_natural_mode): Warn ABI change when %zmm register is not
4686 available for AVX512F vector value passing.
4688 2014-02-18 Kai Tietz <ktietz@redhat.com>
4690 PR target/60193
4691 * config/i386/i386.c (ix86_expand_prologue): Use value in
4692 rax register as displacement when restoring %r10 or %rax.
4693 Fix wrong offset when restoring both registers.
4695 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4697 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4698 assertion with conditional return.
4700 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4701 Uros Bizjak <ubizjak@gmail.com>
4703 PR driver/60233
4704 * config/i386/driver-i386.c (host_detect_local_cpu): If
4705 YMM state is not saved by the OS, also clear has_f16c. Move
4706 CPUID 0x80000001 handling before YMM state saving checking.
4708 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4710 PR rtl-optimization/58960
4711 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4712 factored out from ...
4713 (sched_init): ... here.
4714 (free_global_sched_pressure_data): New, factored out from ...
4715 (sched_finish): ... here.
4716 * sched-int.h (free_global_sched_pressure_data): Declare.
4717 * sched-rgn.c (nr_regions_initial): New static global.
4718 (haifa_find_rgns): Initialize it.
4719 (schedule_region): Disable sched-pressure for the newly
4720 generated regions.
4722 2014-02-17 Richard Biener <rguenther@suse.de>
4724 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4725 release SSA defs of pattern stmts.
4727 2014-02-17 Richard Biener <rguenther@suse.de>
4729 * tree-inline.c (expand_call_inline): Release the virtual
4730 operand defined by the call we are about to inline.
4732 2014-02-17 Richard Biener <rguenther@suse.de>
4734 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4736 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4737 Ilya Tocar <ilya.tocar@intel.com>
4739 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4740 arguments order in builtin.
4741 (_mm512_permutexvar_epi64): Ditto.
4742 (_mm512_mask_permutexvar_epi64): Ditto
4743 (_mm512_maskz_permutexvar_epi32): Ditto
4744 (_mm512_permutexvar_epi32): Ditto
4745 (_mm512_mask_permutexvar_epi32): Ditto
4747 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4749 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4750 (p8_vmrgow): Likewise.
4752 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4754 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4755 endian targets.
4757 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4759 PR target/60203
4760 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4761 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4762 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4763 using direct move instructions on ISA 2.07. Also adjust
4764 instruction length for 64-bit.
4765 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4766 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4768 2014-02-15 Alan Modra <amodra@gmail.com>
4770 PR target/58675
4771 PR target/57935
4772 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4773 find_replacement on parts of insn rtl that might be reloaded.
4775 2014-02-15 Richard Biener <rguenther@suse.de>
4777 PR tree-optimization/60183
4778 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
4779 (tree_ssa_phiprop): Calculate and free post-dominators.
4781 2014-02-14 Jeff Law <law@redhat.com>
4783 PR rtl-optimization/60131
4784 * ree.c (get_extended_src_reg): New function.
4785 (combine_reaching_defs): Use it rather than assuming location of REG.
4786 (find_and_remove_re): Verify first operand of extension is
4787 a REG before adding the insns to the copy list.
4789 2014-02-14 Roland McGrath <mcgrathr@google.com>
4791 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
4792 * configure: Regenerated.
4793 * config.in: Regenerated.
4794 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
4795 instead of ASM_SHORT.
4797 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
4798 Richard Earnshaw <rearnsha@arm.com>
4800 PR rtl-optimization/59535
4801 * lra-constraints.c (process_alt_operands): Encourage alternative
4802 when unassigned pseudo class is superset of the alternative class.
4803 (inherit_reload_reg): Don't inherit when optimizing for code size.
4804 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
4805 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
4806 modes not less than 4 for Thumb1.
4808 2014-02-14 Kyle McMartin <kyle@redhat.com>
4810 PR pch/60010
4811 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
4813 2014-02-14 Richard Biener <rguenther@suse.de>
4815 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
4816 (get_frame_arg): Drop the assert with langhook types_compatible_p.
4817 Do not strip INDIRECT_REFs.
4819 2014-02-14 Richard Biener <rguenther@suse.de>
4821 PR lto/60179
4822 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
4823 DECL_FUNCTION_SPECIFIC_TARGET.
4824 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
4825 * tree-streamer-out.c (pack_ts_target_option): Remove.
4826 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
4827 (write_ts_function_decl_tree_pointers): Do not stream
4828 DECL_FUNCTION_SPECIFIC_TARGET.
4829 * tree-streamer-in.c (unpack_ts_target_option): Remove.
4830 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
4831 (lto_input_ts_function_decl_tree_pointers): Do not stream
4832 DECL_FUNCTION_SPECIFIC_TARGET.
4834 2014-02-14 Jakub Jelinek <jakub@redhat.com>
4836 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
4837 (get_initial_def_for_induction, vectorizable_induction): Ignore
4838 debug stmts when looking for exit_phi.
4839 (vectorizable_live_operation): Fix up condition.
4841 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4843 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
4844 nreverse() because it changes the content of original tree list.
4846 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4848 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
4849 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
4851 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4853 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
4854 GNU coding standards.
4856 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4858 PR debug/60152
4859 * dwarf2out.c (gen_subprogram_die): Don't call
4860 add_calling_convention_attribute if subr_die is old_die.
4862 2014-02-13 Sharad Singhai <singhai@google.com>
4864 * doc/optinfo.texi: Fix order of nodes.
4866 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
4868 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
4869 operands[2], not operands[3].
4871 2014-02-13 Richard Biener <rguenther@suse.de>
4873 PR bootstrap/59878
4874 * doc/install.texi (ISL): Update recommended version to 0.12.2,
4875 mention the possibility of an in-tree build.
4876 (CLooG): Update recommended version to 0.18.1, mention the
4877 possibility of an in-tree build and clarify that the ISL
4878 bundled with CLooG does not work.
4880 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4882 PR target/43546
4883 * expr.c (compress_float_constant): If x is a hard register,
4884 extend into a pseudo and then move to x.
4886 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
4888 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
4889 caused by bad second argument to warning_at() with -mhotpatch and
4890 nested functions (e.g. with gfortran).
4892 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
4894 * opts.c (option_name): Remove "enabled by default" rider.
4896 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
4898 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
4900 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
4901 Uros Bizjak <ubizjak@gmail.com>
4903 PR target/60151
4904 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
4905 * configure: Regenerated.
4907 2014-02-12 Richard Biener <rguenther@suse.de>
4909 * vec.c (vec_prefix::calculate_allocation): Move as
4910 inline variant to vec.h.
4911 (vec_prefix::calculate_allocation_1): New out-of-line version.
4912 * vec.h (vec_prefix::calculate_allocation_1): Declare.
4913 (vec_prefix::m_has_auto_buf): Rename to ...
4914 (vec_prefix::m_using_auto_storage): ... this.
4915 (vec_prefix::calculate_allocation): Inline the easy cases
4916 and dispatch to calculate_allocation_1 which doesn't need the
4917 prefix address.
4918 (va_heap::reserve): Use gcc_checking_assert.
4919 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
4920 m_using_auto_storage.
4921 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
4922 member and adjust.
4923 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
4924 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
4925 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
4927 2014-02-12 Richard Biener <rguenther@suse.de>
4929 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
4930 when we found a dependence.
4932 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
4934 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
4935 common code...
4936 (maybe_fold_stmt): ... into this new function.
4937 * omp-low.c (lower_omp): Update comment.
4939 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
4940 last use.
4942 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
4943 dereference.
4945 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
4947 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
4948 identifiers in comments.
4949 (cortexa53_extra_costs): Likewise.
4950 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
4951 (cortexa7_extra_costs): Likewise.
4952 (cortexa12_extra_costs): Likewise.
4953 (cortexa15_extra_costs): Likewise.
4954 (v7m_extra_costs): Likewise.
4956 2014-02-12 Richard Biener <rguenther@suse.de>
4958 PR middle-end/60092
4959 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
4960 of posix_memalign being successful.
4961 (lower_stmt): Restrict lowering of posix_memalign to when
4962 -ftree-bit-ccp is enabled.
4964 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4966 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
4967 arg_loc.
4968 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
4970 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
4972 PR rtl-optimization/60116
4973 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
4974 other_insn once the combination has been validated.
4976 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
4978 PR lto/59468
4979 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
4980 and wrapper.
4981 * ipa-devirt.c: Include demangle.h
4982 (odr_violation_reported): New static variable.
4983 (add_type_duplicate): Update odr_violations.
4984 (maybe_record_node): Add completep parameter; update it.
4985 (record_target_from_binfo): Add COMPLETEP parameter;
4986 update it as needed.
4987 (possible_polymorphic_call_targets_1): Likewise.
4988 (struct polymorphic_call_target_d): Add nonconstruction_targets;
4989 rename FINAL to COMPLETE.
4990 (record_targets_from_bases): Sanity check we found the binfo;
4991 fix COMPLETEP updating.
4992 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
4993 parameter, fix computing of COMPLETEP.
4994 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
4995 at LTO time do demangling.
4996 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
4997 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
4998 parameter.
4999 (gimple_get_virt_method_for_binfo): Likewise.
5000 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5001 gimple_get_virt_method_for_vtable): Update prototypes.
5003 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5005 PR target/49008
5006 * genautomata.c (add_presence_absence): Fix typo with
5007 {final_}presence_list.
5009 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5011 PR target/60137
5012 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5013 for VSX/Altivec vectors that land in GPR registers.
5015 2014-02-11 Richard Henderson <rth@redhat.com>
5016 Jakub Jelinek <jakub@redhat.com>
5018 PR debug/59776
5019 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5020 around drhs if type conversion to lacc->type is not useless.
5022 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5024 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5025 tuning struct.
5026 (cortex-a57.cortex-a53): Likewise.
5027 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5029 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5031 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5032 arm_restrict_it.
5034 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5036 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5037 add_options_for_arm_vfp3.
5039 2014-02-11 Jeff Law <law@redhat.com>
5041 PR middle-end/54041
5042 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5043 object with an undesirable mode.
5045 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5047 PR libgomp/60107
5048 * config/i386/sol2-9.h: New file.
5049 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5050 *-*-solaris2.9*): Use it.
5052 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5054 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5055 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5057 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5059 * config/microblaze/microblaze.c: Extend mcpu version format
5061 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5063 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5065 2014-02-10 Richard Henderson <rth@redhat.com>
5067 PR target/59927
5068 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5069 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5070 ms-abi vs -mno-accumulate-outgoing-args.
5071 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5072 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5073 respect to ms-abi.
5075 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5077 PR middle-end/60080
5078 * cfgexpand.c (expand_asm_operands): Attach source location to
5079 ASM_INPUT rtx objects.
5080 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5082 2014-02-10 Nick Clifton <nickc@redhat.com>
5084 * config/mn10300/mn10300.c (popcount): New function.
5085 (mn10300_expand_prologue): Include saved registers in stack usage
5086 count.
5088 2014-02-10 Jeff Law <law@redhat.com>
5090 PR middle-end/52306
5091 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5092 when changing the SET_DEST of a prior insn to avoid an input reload.
5094 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5096 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5097 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5098 -mcall-openbsd, or -mcall-linux.
5099 (CC1_ENDIAN_BIG_SPEC): Remove.
5100 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5101 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5102 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5103 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5104 and %cc1_endian_default.
5105 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5107 2014-02-10 Richard Biener <rguenther@suse.de>
5109 PR tree-optimization/60115
5110 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5111 MEM_REF handling. Properly verify that the accesses are not
5112 out of the objects bound.
5114 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5116 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5117 coretex to cortex.
5119 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5121 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5122 proper constants and fix formatting.
5123 (possible_polymorphic_call_targets): Fix formatting.
5125 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5126 Ilya Tocar <ilya.tocar@intel.com>
5128 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5129 (_mm512_loadu_epi32): Renamed into...
5130 (_mm512_loadu_si512): This.
5131 (_mm512_storeu_epi32): Renamed into...
5132 (_mm512_storeu_si512): This.
5133 (_mm512_maskz_ceil_ps): Removed.
5134 (_mm512_maskz_ceil_pd): Ditto.
5135 (_mm512_maskz_floor_ps): Ditto.
5136 (_mm512_maskz_floor_pd): Ditto.
5137 (_mm512_floor_round_ps): Ditto.
5138 (_mm512_floor_round_pd): Ditto.
5139 (_mm512_ceil_round_ps): Ditto.
5140 (_mm512_ceil_round_pd): Ditto.
5141 (_mm512_mask_floor_round_ps): Ditto.
5142 (_mm512_mask_floor_round_pd): Ditto.
5143 (_mm512_mask_ceil_round_ps): Ditto.
5144 (_mm512_mask_ceil_round_pd): Ditto.
5145 (_mm512_maskz_floor_round_ps): Ditto.
5146 (_mm512_maskz_floor_round_pd): Ditto.
5147 (_mm512_maskz_ceil_round_ps): Ditto.
5148 (_mm512_maskz_ceil_round_pd): Ditto.
5149 (_mm512_expand_pd): Ditto.
5150 (_mm512_expand_ps): Ditto.
5151 * config/i386/i386.c (ix86_builtins): Remove
5152 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5153 (bdesc_args): Ditto.
5154 * config/i386/predicates.md (const1256_operand): New.
5155 (const_1_to_2_operand): Ditto.
5156 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5157 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5158 (*avx512pf_gatherpf<mode>sf): Ditto.
5159 (avx512pf_gatherpf<mode>df): Ditto.
5160 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5161 (*avx512pf_gatherpf<mode>df): Ditto.
5162 (avx512pf_scatterpf<mode>sf): Ditto.
5163 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5164 (*avx512pf_scatterpf<mode>sf): Ditto.
5165 (avx512pf_scatterpf<mode>df): Ditto.
5166 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5167 (*avx512pf_scatterpf<mode>df): Ditto.
5168 (avx512f_expand<mode>): Removed.
5169 (<shift_insn><mode>3<mask_name>): Change predicate type.
5171 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5173 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5174 not at the end of datarefs vector use ordered_remove to avoid
5175 reordering datarefs vector.
5177 PR c/59984
5178 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5179 mark local addressable non-static vars as GOVD_PRIVATE
5180 instead of GOVD_LOCAL.
5181 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5182 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5183 than copying them.
5185 PR middle-end/60092
5186 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5187 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5188 assume_aligned or alloc_align attributes.
5189 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5190 arguments. Handle also assume_aligned and alloc_align attributes.
5191 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5192 calls to functions with assume_aligned or alloc_align attributes.
5193 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5195 2014-02-08 Terry Guo <terry.guo@arm.com>
5197 * doc/invoke.texi: Document ARM -march=armv7e-m.
5199 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5201 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5202 flag on __cilkrts_rethrow builtin.
5204 PR ipa/60026
5205 * ipa-cp.c (determine_versionability): Fail at -O0
5206 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5207 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5209 Revert:
5210 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5212 PR ipa/60026
5213 * tree-inline.c (copy_forbidden): Fail for
5214 __attribute__((optimize (0))) functions.
5216 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5218 * varpool.c: Include pointer-set.h.
5219 (varpool_remove_unreferenced_decls): Variables in other partitions
5220 will not be output; be however careful to not lose information
5221 about partitioning.
5223 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5225 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5226 lookup in the vtable constructor.
5228 2014-02-07 Jeff Law <law@redhat.com>
5230 PR target/40977
5231 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5232 define_insn_and_split.
5234 * ipa-inline.c (inline_small_functions): Fix typos.
5236 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5238 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5239 (s390_can_use_return_insn): Declare.
5240 * config/s390/s390.h (EPILOGUE_USES): Define.
5241 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5242 instructions.
5243 (s390_chunkify_start): Handle return JUMP_LABELs.
5244 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5245 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5246 (s390_can_use_return_insn): New functions.
5247 (s390_fix_long_loop_prediction): Handle conditional returns.
5248 (TARGET_SET_UP_BY_PROLOGUE): Define.
5249 * config/s390/s390.md (ANY_RETURN): New code iterator.
5250 (*creturn, *csimple_return, return, simple_return): New patterns.
5252 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5254 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5255 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5256 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5257 REG_CFA_RESTORE list when deciding not to restore a register.
5259 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5261 * config/s390/s390.c: Include tree-pass.h and context.h.
5262 (s390_early_mach): New function, split out from...
5263 (s390_emit_prologue): ...here.
5264 (pass_data_s390_early_mach): New pass structure.
5265 (pass_s390_early_mach): New class.
5266 (s390_option_override): Create and register early_mach pass.
5267 Move to end of file.
5269 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5271 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5272 to match for the exit block.
5274 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5276 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5277 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5278 Reject misaligned operands.
5280 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5282 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5284 2014-02-07 Richard Biener <rguenther@suse.de>
5286 PR middle-end/60092
5287 * gimple-low.c (lower_builtin_posix_memalign): New function.
5288 (lower_stmt): Call it to lower posix_memalign in a way
5289 to make alignment info accessible.
5291 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5293 PR c++/60082
5294 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5295 __builtin_setjmp_receiver.
5297 2014-02-07 Richard Biener <rguenther@suse.de>
5299 PR middle-end/60092
5300 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5301 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5302 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5303 Handle BUILT_IN_POSIX_MEMALIGN.
5304 (find_func_clobbers): Likewise.
5305 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5306 (call_may_clobber_ref_p_1): Likewise.
5308 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5310 PR ipa/59918
5311 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5312 sanity check.
5314 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5316 PR ipa/59469
5317 * lto-cgraph.c (lto_output_node): Use
5318 symtab_get_symbol_partitioning_class.
5319 (lto_output_varpool_node): likewise.
5320 (symtab_get_symbol_partitioning_class): Move here from
5321 lto/lto-partition.c
5322 * cgraph.h (symbol_partitioning_class): Likewise.
5323 (symtab_get_symbol_partitioning_class): Declare.
5325 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5327 * ggc.h (ggc_internal_cleared_alloc): New macro.
5328 * vec.h (vec_safe_copy): Handle memory stats.
5329 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5330 * target-globals.c (save_target_globals): Likewise.
5332 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5334 PR target/60077
5335 * expr.c (emit_move_resolve_push): Export; be bit more selective
5336 on when to clear alias set.
5337 * expr.h (emit_move_resolve_push): Declare.
5338 * function.h (struct function): Add tail_call_marked.
5339 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5340 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5341 * config/i386/i386.md (TImode move expander): De not call
5342 ix86_expand_push.
5343 (FP push expanders): Preserve memory attributes.
5344 * config/i386/sse.md (push<mode>1): Remove.
5345 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5346 (ix86_expand_push): Remove.
5347 * config/i386/mmx.md (push<mode>1): Remove.
5349 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5351 PR rtl-optimization/60030
5352 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5353 lopart with paradoxical subreg before shifting it up by hprec.
5355 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5357 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5358 Remove extra newline at end of file.
5359 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5360 (arm_issue_rate): Handle cortexa57.
5361 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5362 (cortex-a57.cortex-a53): Likewise.
5364 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5366 PR target/59575
5367 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5368 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5369 bitmask.
5370 (arm_expand_prologue): Adjust all callers.
5371 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5372 info, registers also at the lowest numbered registers side. Use
5373 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5374 XEXP.
5376 PR debug/59992
5377 * var-tracking.c (adjust_mems): Before adding a SET to
5378 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5380 2014-02-06 Alan Modra <amodra@gmail.com>
5382 PR target/60032
5383 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5384 change SDmode to DDmode when lra_in_progress.
5386 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5388 PR middle-end/59150
5389 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5390 free_data_ref on the dr first, and before goto again also set dr
5391 to the next dr. For simd_lane_access, free old datarefs[i] before
5392 overwriting it. For get_vectype_for_scalar_type failure, don't
5393 free_data_ref if simd_lane_access.
5395 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5397 PR target/60062
5398 * tree.h (opts_for_fn): New inline function.
5399 (opt_for_fn): Define.
5400 * config/i386/i386.c (ix86_function_regparm): Use
5401 opt_for_fn (decl, optimize) instead of optimize.
5403 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5405 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5406 for SYMBOL_REF in large memory model.
5408 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5410 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5411 and crypto support.
5412 (cortex-a57): Likewise.
5413 (cortex-a57.cortex-a53): Likewise.
5415 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5416 Kugan Vivekanandarajah <kuganv@linaro.org>
5418 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5419 unaligned_access.
5420 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5422 2014-02-06 Richard Biener <rguenther@suse.de>
5424 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5425 set_loop_copy and initialize_original_copy_tables.
5427 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5429 * config/aarch64/aarch64-simd.md
5430 (aarch64_ashr_simddi): Change QI to SI.
5432 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5433 Jakub Jelinek <jakub@redhat.com>
5435 PR middle-end/60013
5436 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5437 of the dataflow.
5439 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5441 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5442 CODE_FOR_altivec_vpku[hw]um to
5443 CODE_FOR_altivec_vpku[hw]um_direct.
5444 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5445 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5446 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5447 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5449 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5451 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5452 generation for -maltivec=be.
5453 (altivec_vsumsws): Simplify redundant test.
5455 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5457 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5458 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5459 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5460 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5461 gen_altivec_vpkuwum.
5462 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5463 BYTES_BIG_ENDIAN.
5464 (altivec_vpks<VI_char>ss): Likewise.
5465 (altivec_vpks<VI_char>us): Likewise.
5466 (altivec_vpku<VI_char>us): Likewise.
5467 (altivec_vpku<VI_char>um): Likewise.
5468 (altivec_vpku<VI_char>um_direct): New (copy of
5469 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5470 internal use).
5471 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5472 target is little endian and -maltivec=be is not specified.
5473 (*altivec_vupkhs<VU_char>_direct): New (copy of
5474 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5475 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5476 target is little endian and -maltivec=be is not specified.
5477 (*altivec_vupkls<VU_char>_direct): New (copy of
5478 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5479 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5480 little endian and -maltivec=be is not specified.
5481 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5482 little endian and -maltivec=be is not specified.
5484 2014-02-05 Richard Henderson <rth@redhat.com>
5486 PR debug/52727
5487 * combine-stack-adj.c: Revert r206943.
5488 * sched-int.h (struct deps_desc): Add last_args_size.
5489 * sched-deps.c (init_deps): Initialize it.
5490 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5491 contain REG_ARGS_SIZE notes.
5493 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5495 * lto-cgraph.c (asm_nodes_output): Make global.
5496 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5497 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5498 (driver_handle_option): Handle OPT_fwpa.
5500 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5502 PR ipa/59947
5503 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5504 a comment typo and formatting issue. If odr_hash hasn't been
5505 created, return vNULL and set *completep to false.
5507 PR middle-end/57499
5508 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5509 bb with no successors.
5511 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5513 PR target/59718
5514 * doc/invoke.texi (-march): Clarify documentation for ARM.
5515 (-mtune): Likewise.
5516 (-mcpu): Likewise.
5518 2014-02-05 Richard Biener <rguenther@suse.de>
5520 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5521 when not vectorizing because of too many alias checks.
5522 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5523 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5525 2014-02-05 Nick Clifton <nickc@redhat.com>
5527 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5528 accept extended registers in any mode when compiling for the MN10300.
5530 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5532 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5533 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5534 sanitization attributes.
5535 (can_inline_edge_p): Likewise.
5536 (sanitize_attrs_match_for_inline_p): New function.
5538 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5540 * ipa-prop.c (detect_type_change): Shor circuit testing of
5541 type changes on THIS pointer.
5543 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5545 PR target/59777
5546 * config/pa/pa.c (legitimize_tls_address): Return original address
5547 if not passed a SYMBOL_REF rtx.
5548 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5549 addresses.
5550 (pa_emit_move_sequence): Simplify TLS source operands.
5551 (pa_legitimate_constant_p): Reject all TLS constants.
5552 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5553 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5555 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5557 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5558 groups when we know they are controlled by LTO.
5559 * varasm.c (default_binds_local_p_1): If object is in other partition,
5560 it will be resolved locally.
5562 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5564 * config/host-linux.c (linux_gt_pch_use_address): Don't
5565 use SSIZE_MAX because it is not always defined.
5567 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5569 PR bootstrap/59913
5570 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5571 threshold for pseudo splitting.
5572 (update_ebb_live_info): Process call argument hard registers and
5573 hard registers from insn definition too.
5574 (max_small_class_regs_num): New constant.
5575 (inherit_in_ebb): Update live hard regs through EBBs. Update
5576 reloads_num only for small register classes. Don't split for
5577 outputs of jumps.
5579 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5581 PR ipa/60058
5582 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5583 is non-null.
5585 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5587 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5588 visibility is safe.
5590 2014-02-04 Marek Polacek <polacek@redhat.com>
5592 * gdbinit.in (pel): Define.
5594 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5596 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5597 behavior.
5599 2014-02-04 Richard Biener <rguenther@suse.de>
5601 PR lto/59723
5602 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5603 in function context local.
5604 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5605 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5606 similar to LTO_imported_decl_ref.
5608 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5610 PR tree-optimization/60002
5611 * cgraphclones.c (build_function_decl_skip_args): Clear
5612 DECL_LANG_SPECIFIC.
5614 PR tree-optimization/60023
5615 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5616 false to gsi_replace.
5617 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5618 has been in some EH region and vec_stmt could throw, add
5619 vec_stmt into the same EH region.
5620 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5621 has no lhs, ignore it.
5622 * internal-fn.c (expand_MASK_LOAD): Likewise.
5624 PR ipa/60026
5625 * tree-inline.c (copy_forbidden): Fail for
5626 __attribute__((optimize (0))) functions.
5628 PR other/58712
5629 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5630 is set, copy one less argument.
5631 (expand_simd_clones): Don't subtract clone_info->inbranch
5632 from simd_clone_struct_alloc argument.
5634 PR rtl-optimization/57915
5635 * recog.c (simplify_while_replacing): If all unary/binary/relational
5636 operation arguments are constant, attempt to simplify those.
5638 PR middle-end/59261
5639 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5640 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5642 2014-02-04 Richard Biener <rguenther@suse.de>
5644 PR tree-optimization/60012
5645 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5646 TBAA disambiguation to all DDRs.
5648 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5650 PR target/59788
5651 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5652 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5654 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5656 PR ipa/59882
5657 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5659 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5661 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5662 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5664 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5666 PR ipa/59831
5667 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5668 to figure out targets of polymorphic calls with known decl.
5669 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5670 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5671 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5672 (get_polymorphic_call_info): ... here.
5673 (get_polymorphic_call_info_from_invariant): New function.
5675 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5677 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5678 lookup via vtable pointer; check for type consistency
5679 and turn inconsitent facts into UNREACHABLE.
5680 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5681 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5682 type inconsistent querries; return UNREACHABLE instead.
5684 2014-02-03 Richard Henderson <rth@twiddle.net>
5686 PR tree-opt/59924
5687 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5688 already processed this node.
5689 (normalize_one_pred_1): Pass along mark_set.
5690 (normalize_one_pred): Create and destroy a pointer_set_t.
5691 (normalize_one_pred_chain): Likewise.
5693 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5695 PR gcov-profile/58602
5696 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5698 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5700 PR ipa/59831
5701 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5702 -fno-devirtualize; try to devirtualize by the knowledge of
5703 virtual table pointer given by aggregate propagation.
5704 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5705 (ipa_print_node_jump_functions): Dump also offset that
5706 is relevant for polymorphic calls.
5707 (determine_known_aggregate_parts): Add arg_type parameter; use it
5708 instead of determining the type from pointer type.
5709 (ipa_compute_jump_functions_for_edge): Update call of
5710 determine_known_aggregate_parts.
5711 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5712 (gimple_get_virt_method_for_binfo): ... here; simplify using
5713 vtable_pointer_value_to_vtable.
5714 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5715 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5716 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5717 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5718 POINTER_PLUS_EXPR.
5719 (vtable_pointer_value_to_binfo): ... here.
5720 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5722 2014-02-03 Teresa Johnson <tejohnson@google.com>
5724 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5725 redef of outer loop index variable.
5727 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5729 PR c++/53017
5730 PR c++/59211
5731 * doc/extend.texi (Function Attributes): Typo.
5733 2014-02-03 Cong Hou <congh@google.com>
5735 PR tree-optimization/60000
5736 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5737 if the vectorized statement is a store. A store statement can only
5738 appear at the end of pattern statements.
5740 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5742 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5743 (ix86_option_override_internal): Default long double to 64-bit for
5744 32-bit Bionic and to 128-bit for 64-bit Bionic.
5746 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5747 TARGET_LONG_DOUBLE_128 is true.
5748 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5750 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5751 (mlong-double-64): Negate -mlong-double-128.
5752 (mlong-double-128): New option.
5754 * config/i386/i386-c.c (ix86_target_macros): Define
5755 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5757 * doc/invoke.texi: Document -mlong-double-128.
5759 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5761 PR rtl-optimization/60024
5762 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5764 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5766 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5768 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5770 PR rtl-optimization/57662
5771 * sel-sched.c (code_motion_path_driver): Do not mark already not
5772 existing blocks in the visiting bitmap.
5774 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5776 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
5777 on the insn being emitted.
5779 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
5780 Will Deacon <will.deacon@arm.com>
5782 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
5784 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5786 * config/arm/arm-tables.opt: Regenerate.
5788 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5790 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
5791 for vector types other than V16QImode.
5792 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
5793 define_expand, and call altivec_expand_vec_perm_le when producing
5794 code with little endian element order.
5795 (*altivec_vperm_<mode>_internal): New insn having previous
5796 behavior of altivec_vperm_<mode>.
5797 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
5798 altivec_expand_vec_perm_le when producing code with little endian
5799 element order.
5800 (*altivec_vperm_<mode>_uns_internal): New insn having previous
5801 behavior of altivec_vperm_<mode>_uns.
5803 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5805 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
5806 (altivec_vsumsws): Add handling for -maltivec=be with a little
5807 endian target.
5808 (altivec_vsumsws_direct): New.
5809 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
5810 gen_altivec_vsumsws.
5812 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5814 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
5815 vtable_pointer_value_to_binfo): New functions.
5816 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
5817 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
5819 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
5821 * config/nios2/nios2.md (load_got_register): Initialize GOT
5822 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
5823 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
5825 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5827 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
5828 preserverd by passthrough, do not propagate the type.
5830 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5832 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
5833 (mips_atomic_assign_expand_fenv): New function.
5834 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
5836 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5838 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
5839 (__builtin_mips_set_fcsr): Likewise.
5840 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
5841 MIPS_USI_FTYPE_VOID.
5842 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
5843 (mips16_expand_set_fcsr): Likewise.
5844 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
5845 (mips16_set_fcsr_stub): Likewise.
5846 (mips16_get_fcsr_one_only_stub): New class.
5847 (mips16_set_fcsr_one_only_stub): Likewise.
5848 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
5849 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
5850 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
5851 (hard_float): New availability predicate.
5852 (mips_builtins): Add get_fcsr and set_fcsr.
5853 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
5854 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
5855 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
5856 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
5857 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
5858 patterns.
5860 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5862 * config/mips/mips.c (mips_one_only_stub): New class.
5863 (mips_need_mips16_rdhwr_p): Replace with...
5864 (mips16_rdhwr_stub): ...this new variable.
5865 (mips16_stub_call_address): New function.
5866 (mips16_rdhwr_one_only_stub): New class.
5867 (mips_expand_thread_pointer): Use mips16_stub_call_address.
5868 (mips_output_mips16_rdhwr): Delete.
5869 (mips_finish_stub): New function.
5870 (mips_code_end): Use it to handle rdhwr stubs.
5872 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
5874 PR target/60017
5875 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
5876 when calculating size of integer atomic types.
5878 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
5880 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
5882 2014-02-01 Jakub Jelinek <jakub@redhat.com>
5884 PR tree-optimization/60003
5885 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
5886 * profile.c (branch_prob): Use gimple_call_builtin_p
5887 to check for BUILT_IN_SETJMP_RECEIVER.
5888 * tree-inline.c (copy_bb): Call notice_special_calls.
5890 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
5892 PR bootstrap/59985
5893 * lra-constraints.c (process_alt_operands): Update reload_sum only
5894 on the first pass.
5896 2014-01-31 Richard Henderson <rth@redhat.com>
5898 PR middle-end/60004
5899 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
5900 until after else_eh is processed.
5902 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
5904 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
5905 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
5906 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
5907 in smmintrin.h, remove them.
5908 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
5909 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
5910 * config/i386/i386.md (ROUND_SAE): Fix value.
5911 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
5912 (const48_operand): New.
5913 * config/i386/subst.md (round), (round_expand): Use
5914 const_4_or_8_to_11_operand.
5915 (round_saeonly), (round_saeonly_expand): Use const48_operand.
5917 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
5919 * config/i386/constraints.md (Yk): Swap meaning with k.
5920 * config/i386/i386.md (movhi_internal): Change Yk to k.
5921 (movqi_internal): Ditto.
5922 (*k<logic><mode>): Ditto.
5923 (*andhi_1): Ditto.
5924 (*andqi_1): Ditto.
5925 (kandn<mode>): Ditto.
5926 (*<code>hi_1): Ditto.
5927 (*<code>qi_1): Ditto.
5928 (kxnor<mode>): Ditto.
5929 (kortestzhi): Ditto.
5930 (kortestchi): Ditto.
5931 (kunpckhi): Ditto.
5932 (*one_cmplhi2_1): Ditto.
5933 (*one_cmplqi2_1): Ditto.
5934 * config/i386/sse.md (): Change k to Yk.
5935 (avx512f_load<mode>_mask): Ditto.
5936 (avx512f_blendm<mode>): Ditto.
5937 (avx512f_store<mode>_mask): Ditto.
5938 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
5939 (avx512f_storedqu<mode>_mask): Ditto.
5940 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
5941 Ditto.
5942 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
5943 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
5944 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
5945 (avx512f_maskcmp<mode>3): Ditto.
5946 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
5947 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
5948 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
5949 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
5950 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
5951 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
5952 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
5953 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
5954 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
5955 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
5956 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
5957 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
5958 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
5959 (vec_extract_lo_<mode>_maskm): Ditto.
5960 (vec_extract_hi_<mode>_maskm): Ditto.
5961 (avx512f_vternlog<mode>_mask): Ditto.
5962 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
5963 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
5964 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
5965 (avx512f_<code>v8div16qi2_mask): Ditto.
5966 (avx512f_<code>v8div16qi2_mask_store): Ditto.
5967 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
5968 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
5969 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
5970 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
5971 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5972 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5973 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5974 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5975 (avx512cd_maskb_vec_dupv8di): Ditto.
5976 (avx512cd_maskw_vec_dupv16si): Ditto.
5977 (avx512f_vpermi2var<mode>3_maskz): Ditto.
5978 (avx512f_vpermi2var<mode>3_mask): Ditto.
5979 (avx512f_vpermi2var<mode>3_mask): Ditto.
5980 (avx512f_vpermt2var<mode>3_maskz): Ditto.
5981 (*avx512f_gathersi<mode>): Ditto.
5982 (*avx512f_gathersi<mode>_2): Ditto.
5983 (*avx512f_gatherdi<mode>): Ditto.
5984 (*avx512f_gatherdi<mode>_2): Ditto.
5985 (*avx512f_scattersi<mode>): Ditto.
5986 (*avx512f_scatterdi<mode>): Ditto.
5987 (avx512f_compress<mode>_mask): Ditto.
5988 (avx512f_compressstore<mode>_mask): Ditto.
5989 (avx512f_expand<mode>_mask): Ditto.
5990 * config/i386/subst.md (mask): Change k to Yk.
5991 (mask_scalar_merge): Ditto.
5992 (sd): Ditto.
5994 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
5996 * doc/extend.texi (Vector Extensions): Document ?: in C++.
5998 2014-01-31 Richard Biener <rguenther@suse.de>
6000 PR middle-end/59990
6001 * builtins.c (fold_builtin_memory_op): Make sure to not
6002 use a floating-point mode or a boolean or enumeral type for
6003 the copy operation.
6005 2014-01-30 DJ Delorie <dj@redhat.com>
6007 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6008 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6009 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6010 whenever main() has an epilogue.
6012 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6014 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6015 unused variable "field".
6016 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6017 (vsx_mergeh_<mode>): Likewise.
6018 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6019 (altivec_vmrghh): Likewise.
6020 (altivec_vmrghw): Likewise.
6021 (altivec_vmrglb): Likewise.
6022 (altivec_vmrglh): Likewise.
6023 (altivec_vmrglw): Likewise.
6024 (altivec_vspltb): Add missing uses.
6025 (altivec_vsplth): Likewise.
6026 (altivec_vspltw): Likewise.
6027 (altivec_vspltsf): Likewise.
6029 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6031 PR target/59923
6032 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6033 frame related instructions.
6035 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6037 PR rtl-optimization/59959
6038 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6039 any reload of register whose subreg is invalid.
6041 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6043 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6044 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6045 Add missing return type - void.
6047 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6049 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6050 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6051 remove element index adjustment for endian (now handled in vsx.md
6052 and altivec.md).
6053 (altivec_expand_vec_perm_const): Use
6054 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6055 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6056 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6057 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6058 define_expand and a new define_insn *altivec_vspltb_internal;
6059 adjust for -maltivec=be on a little endian target.
6060 (altivec_vspltb_direct): New.
6061 (altivec_vsplth): Divide into a define_expand and a new
6062 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6063 little endian target.
6064 (altivec_vsplth_direct): New.
6065 (altivec_vspltw): Divide into a define_expand and a new
6066 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6067 little endian target.
6068 (altivec_vspltw_direct): New.
6069 (altivec_vspltsf): Divide into a define_expand and a new
6070 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6071 a little endian target.
6073 2014-01-30 Richard Biener <rguenther@suse.de>
6075 PR tree-optimization/59993
6076 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6077 can propagate form the earlier stmt and avoid the transform
6078 when the intermediate result is needed.
6080 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6082 * README.Portability: Fix typo.
6084 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6086 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6087 comparison_operator with ordered_comparison_operator.
6089 2014-01-30 Nick Clifton <nickc@redhat.com>
6091 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6092 Rename to mn10300_store_multiple_regs.
6093 * config/mn10300/mn10300.c: Likewise.
6094 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6095 store_multiple_regs.
6096 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6097 Call mn10300_store_multiple_regs.
6099 2014-01-30 Nick Clifton <nickc@redhat.com>
6100 DJ Delorie <dj@redhat.com>
6102 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6103 %fp 2 to keep registers after it properly word-aligned.
6104 (rl78_alloc_physical_registers_umul): Handle the case where both
6105 input operands are the same.
6107 2014-01-30 Richard Biener <rguenther@suse.de>
6109 PR tree-optimization/59903
6110 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6111 check properly.
6113 2014-01-30 Jason Merrill <jason@redhat.com>
6115 PR c++/59633
6116 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6118 PR c++/59645
6119 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6121 2014-01-30 Richard Biener <rguenther@suse.de>
6123 PR tree-optimization/59951
6124 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6126 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6128 PR target/59784
6129 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6130 SFmode to DFmode case.
6132 2014-01-29 DJ Delorie <dj@redhat.com>
6134 * config/msp430/msp430.opt (-minrt): New.
6135 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6136 if -minrt given.
6137 (ENDFILE_SPEC): Likewise.
6139 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6141 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6142 (estimate_function_body_sizes): Use it.
6144 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6146 PR c++/58561
6147 * dwarf2out.c (is_cxx_auto): New.
6148 (is_base_type): Use it.
6149 (gen_type_die_with_usage): Likewise.
6151 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6153 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6154 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6155 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6156 -maltivec=be with LE targets.
6157 (vsx_mergeh_<mode>): Likewise.
6158 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6159 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6160 (altivec_vmrghb): Replace with define_expand and new
6161 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6162 (altivec_vmrghb_direct): New define_insn.
6163 (altivec_vmrghh): Replace with define_expand and new
6164 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6165 (altivec_vmrghh_direct): New define_insn.
6166 (altivec_vmrghw): Replace with define_expand and new
6167 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6168 (altivec_vmrghw_direct): New define_insn.
6169 (*altivec_vmrghsf): Adjust for endianness.
6170 (altivec_vmrglb): Replace with define_expand and new
6171 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6172 (altivec_vmrglb_direct): New define_insn.
6173 (altivec_vmrglh): Replace with define_expand and new
6174 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6175 (altivec_vmrglh_direct): New define_insn.
6176 (altivec_vmrglw): Replace with define_expand and new
6177 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6178 (altivec_vmrglw_direct): New define_insn.
6179 (*altivec_vmrglsf): Adjust for endianness.
6180 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6181 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6182 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6183 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6184 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6185 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6186 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6187 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6189 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6191 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6192 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6193 whitespace.
6195 2014-01-29 Richard Biener <rguenther@suse.de>
6197 PR tree-optimization/58742
6198 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6199 associate_pointerplus_align.
6200 (associate_pointerplus_diff): New function.
6201 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6202 and associate_pointerplus_diff.
6204 2014-01-29 Richard Biener <rguenther@suse.de>
6206 * lto-streamer.h (LTO_major_version): Bump to 3.
6207 (LTO_minor_version): Reset to 0.
6209 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6211 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6212 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6213 (arm_file_start): Generate correct asm header for armv7ve.
6214 * config/arm/bpabi.h: Add multilib support for armv7ve.
6215 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6216 and cortex-a15 to armv7ve.
6217 * config/arm/t-aprofile: Add multilib support for armv7ve.
6218 * doc/invoke.texi: Document -march=armv7ve.
6220 2014-01-29 Richard Biener <rguenther@suse.de>
6222 PR tree-optimization/58742
6223 * tree-ssa-forwprop.c (associate_plusminus): Return true
6224 if we changed sth, defer EH cleanup to ...
6225 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6226 (simplify_mult): New function.
6228 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6230 PR middle-end/59917
6231 PR tree-optimization/59920
6232 * tree.c (build_common_builtin_nodes): Remove
6233 __builtin_setjmp_dispatcher initialization.
6234 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6235 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6236 instead of gsi_after_labels + manually skipping debug stmts.
6237 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6238 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6239 * tree-inline.c (copy_edges_for_bb): Remove
6240 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6241 argument. Ignore computed_goto_p stmts. Don't call
6242 make_abnormal_goto_edges. If a call might need abnormal edges
6243 for non-local gotos, see if it already has an edge to
6244 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6245 with true argument, don't do anything then, otherwise add
6246 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6247 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6248 caller.
6249 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6250 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6251 (lower_stmt): Don't set data->calls_builtin_setjmp.
6252 (lower_builtin_setjmp): Adjust comment.
6253 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6254 * tree-cfg.c (found_computed_goto): Remove.
6255 (factor_computed_gotos): Remove.
6256 (make_goto_expr_edges): Return bool, true for computed gotos.
6257 Don't call make_abnormal_goto_edges.
6258 (build_gimple_cfg): Don't set found_computed_goto, don't call
6259 factor_computed_gotos.
6260 (computed_goto_p): No longer static.
6261 (make_blocks): Don't set found_computed_goto.
6262 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6263 (make_edges): If make_goto_expr_edges returns true, push bb
6264 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6265 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6266 vector. Record mapping between bbs and OpenMP regions if there
6267 are any, adjust make_gimple_omp_edges caller. Call
6268 handle_abnormal_edges.
6269 (make_abnormal_goto_edges): Remove.
6270 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6271 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6272 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6273 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6274 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6275 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6276 filling *region also set *region_idx to (*region)->entry->index.
6278 PR other/58712
6279 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6280 For REGs set ORIGINAL_REGNO.
6282 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6284 * doc/md.texi: Mention that a target shouldn't implement
6285 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6286 than hi/lo pair.
6288 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6290 PR tree-optimization/59594
6291 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6292 a copy of the datarefs vector rather than the vector itself.
6294 2014-01-28 Jason Merrill <jason@redhat.com>
6296 PR c++/53756
6297 * dwarf2out.c (auto_die): New static.
6298 (gen_type_die_with_usage): Handle C++1y 'auto'.
6299 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6300 on definition.
6302 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6304 PR target/59672
6305 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6306 (SPEC_X32): Likewise.
6307 (SPEC_64): Likewise.
6308 * config/i386/i386.c (ix86_option_override_internal): Turn off
6309 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6310 for TARGET_16BIT.
6311 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6312 * config/i386/i386.h (TARGET_16BIT): New macro.
6313 (TARGET_16BIT_P): Likewise.
6314 * config/i386/i386.opt: Add m16.
6315 * doc/invoke.texi: Document -m16.
6317 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6319 PR preprocessor/59935
6320 * input.c (location_get_source_line): Bail out on when line number
6321 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6323 2014-01-28 Richard Biener <rguenther@suse.de>
6325 PR tree-optimization/58742
6326 * tree-ssa-forwprop.c (associate_plusminus): Handle
6327 pointer subtraction of the form (T)(P + A) - (T)P.
6329 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6331 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6332 at const_int_cost.
6334 2014-01-28 Richard Biener <rguenther@suse.de>
6336 Revert
6337 2014-01-28 Richard Biener <rguenther@suse.de>
6339 PR rtl-optimization/45364
6340 PR rtl-optimization/59890
6341 * var-tracking.c (local_get_addr_clear_given_value): Handle
6342 already cleared slot.
6343 (val_reset): Handle not allocated local_get_addr_cache.
6344 (vt_find_locations): Use post-order on the inverted CFG.
6346 2014-01-28 Richard Biener <rguenther@suse.de>
6348 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6350 2014-01-28 Richard Biener <rguenther@suse.de>
6352 PR rtl-optimization/45364
6353 PR rtl-optimization/59890
6354 * var-tracking.c (local_get_addr_clear_given_value): Handle
6355 already cleared slot.
6356 (val_reset): Handle not allocated local_get_addr_cache.
6357 (vt_find_locations): Use post-order on the inverted CFG.
6359 2014-01-28 Alan Modra <amodra@gmail.com>
6361 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6362 * configure.ac <recursive call for build != host>: Define
6363 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6364 and LD_FOR_BUILD too.
6365 * configure: Regenerate.
6367 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6369 * config/i386/i386.c (get_builtin_code_for_version): Separate
6370 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6371 Broadwell from Haswell.
6373 2014-01-27 Steve Ellcey <sellcey@mips.com>
6375 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6376 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6377 * config/mips/mips.c (mips_option_override): Change setting
6378 of TARGET_DSP.
6379 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6380 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6381 Change from Mask to Var.
6383 2014-01-27 Jeff Law <law@redhat.com>
6385 * ipa-inline.c (inline_small_functions): Fix typo.
6387 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6389 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6390 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6391 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6392 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6393 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6394 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6395 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6396 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6397 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6398 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6399 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6400 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6401 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6402 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6403 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6404 (_mm512_storeu_epi64): Ditto.
6405 (_mm512_cmpge_epi32_mask): Ditto.
6406 (_mm512_cmpge_epu32_mask): Ditto.
6407 (_mm512_cmpge_epi64_mask): Ditto.
6408 (_mm512_cmpge_epu64_mask): Ditto.
6409 (_mm512_cmple_epi32_mask): Ditto.
6410 (_mm512_cmple_epu32_mask): Ditto.
6411 (_mm512_cmple_epi64_mask): Ditto.
6412 (_mm512_cmple_epu64_mask): Ditto.
6413 (_mm512_cmplt_epi32_mask): Ditto.
6414 (_mm512_cmplt_epu32_mask): Ditto.
6415 (_mm512_cmplt_epi64_mask): Ditto.
6416 (_mm512_cmplt_epu64_mask): Ditto.
6417 (_mm512_cmpneq_epi32_mask): Ditto.
6418 (_mm512_cmpneq_epu32_mask): Ditto.
6419 (_mm512_cmpneq_epi64_mask): Ditto.
6420 (_mm512_cmpneq_epu64_mask): Ditto.
6421 (_mm512_expand_pd): Ditto.
6422 (_mm512_expand_ps): Ditto.
6423 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6424 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6425 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6426 * config/i386/i386.c (ix86_builtins): Add
6427 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6428 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6429 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6430 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6431 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6432 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6433 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6434 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6435 IX86_BUILTIN_PMOVUSQW512_MEM.
6436 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6437 __builtin_ia32_pmovsqd512mem_mask,
6438 __builtin_ia32_pmovqd512mem_mask,
6439 __builtin_ia32_pmovusqw512mem_mask,
6440 __builtin_ia32_pmovsqw512mem_mask,
6441 __builtin_ia32_pmovqw512mem_mask,
6442 __builtin_ia32_pmovusdw512mem_mask,
6443 __builtin_ia32_pmovsdw512mem_mask,
6444 __builtin_ia32_pmovdw512mem_mask,
6445 __builtin_ia32_pmovqb512mem_mask,
6446 __builtin_ia32_pmovusqb512mem_mask,
6447 __builtin_ia32_pmovsqb512mem_mask,
6448 __builtin_ia32_pmovusdb512mem_mask,
6449 __builtin_ia32_pmovsdb512mem_mask,
6450 __builtin_ia32_pmovdb512mem_mask.
6451 (bdesc_args): Add __builtin_ia32_expanddf512,
6452 __builtin_ia32_expandsf512.
6453 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6454 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6455 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6456 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6457 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6458 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6459 (avx512f_<code>v8div16qi2_mask_store): This.
6460 (avx512f_expand<mode>): New.
6462 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6464 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6465 New.
6466 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6467 (_mm512_prefetch_i32scatter_pd): Ditto.
6468 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6469 (_mm512_prefetch_i64scatter_pd): Ditto.
6470 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6471 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6472 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6473 (_mm512_prefetch_i32scatter_ps): Ditto.
6474 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6475 (_mm512_prefetch_i64scatter_ps): Ditto.
6476 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6477 * config/i386/i386-builtin-types.def: Define
6478 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6479 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6480 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6481 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6482 IX86_BUILTIN_SCATTERPFQPD.
6483 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6484 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6485 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6486 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6487 __builtin_ia32_scatterpfqps.
6488 (ix86_expand_builtin): Expand new built-ins.
6489 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6490 fix memory access data type.
6491 (*avx512pf_gatherpf<mode>_mask): Ditto.
6492 (*avx512pf_gatherpf<mode>): Ditto.
6493 (avx512pf_scatterpf<mode>): Ditto.
6494 (*avx512pf_scatterpf<mode>_mask): Ditto.
6495 (*avx512pf_scatterpf<mode>): Ditto.
6496 (GATHER_SCATTER_SF_MEM_MODE): New.
6497 (avx512pf_gatherpf<mode>df): Ditto.
6498 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6499 (*avx512pf_scatterpf<mode>df): Ditto.
6501 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6503 PR bootstrap/59934
6504 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6505 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6506 reached.
6508 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6510 * common/config/arm/arm-common.c
6511 (arm_rewrite_mcpu): Handle multiple names.
6512 * config/arm/arm.h
6513 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6515 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6517 * gimple-builder.h (create_gimple_tmp): Delete.
6519 2014-01-27 Christian Bruel <christian.bruel@st.com>
6521 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6522 words comparisons.
6524 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6526 * config/pa/pa.md (call): Generate indirect long calls to non-local
6527 functions when outputing 32-bit code.
6528 (call_value): Likewise except for special call to buggy powf function.
6530 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6531 portable runtime and PIC indirect calls.
6532 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6533 and PIC call sequences. Use ldo instead of blr to set return register
6534 in PIC call sequence.
6536 2014-01-25 Walter Lee <walt@tilera.com>
6538 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6539 avoid clobbering a live register.
6541 2014-01-25 Walter Lee <walt@tilera.com>
6543 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6544 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6545 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6546 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6548 2014-01-25 Walter Lee <walt@tilera.com>
6550 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6551 arguments on even registers.
6552 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6553 STACK_BOUNDARY.
6554 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6555 (BIGGEST_ALIGNMENT): Ditto.
6556 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6558 2014-01-25 Walter Lee <walt@tilera.com>
6560 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6561 insns before bundling.
6562 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6564 2014-01-25 Walter Lee <walt@tilera.com>
6566 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6567 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6568 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6570 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6572 * config/mips/constraints.md (kl): Delete.
6573 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6574 define expands, using...
6575 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6576 instructions for MIPS16.
6577 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6578 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6580 2014-01-25 Walter Lee <walt@tilera.com>
6582 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6583 (clzdi2): Ditto.
6584 (ffsdi2): Ditto.
6586 2014-01-25 Walter Lee <walt@tilera.com>
6588 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6589 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6591 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6593 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6594 Handle XOR.
6596 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6598 * print-rtl.c (in_call_function_usage): New var.
6599 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6600 EXPR_LIST mode as mode and not as reg note name.
6602 PR middle-end/59561
6603 * cfgloopmanip.c (copy_loop_info): If
6604 loop->warned_aggressive_loop_optimizations, make sure
6605 the flag is set in target loop too.
6607 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6609 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6610 flag_cilkplus.
6611 * builtins.def: Likewise.
6612 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6613 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6614 * ira.c (ira_setup_eliminable_regset): Likewise.
6615 * omp-low.c (gate_expand_omp): Likewise.
6616 (execute_lower_omp): Likewise.
6617 (diagnose_sb_0): Likewise.
6618 (gate_diagnose_omp_blocks): Likewise.
6619 (simd_clone_clauses_extract): Likewise.
6620 (gate): Likewise.
6622 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6624 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6625 correction for little endian...
6626 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6627 here.
6629 2014-01-24 Jeff Law <law@redhat.com>
6631 PR tree-optimization/59919
6632 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6633 for non-returning calls.
6635 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6637 * common/config/aarch64/aarch64-common.c
6638 (aarch64_rewrite_mcpu): Handle multiple names.
6639 * config/aarch64/aarch64.h
6640 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6642 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6644 * input.c (add_file_to_cache_tab): Handle the case where fopen
6645 returns NULL.
6647 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6649 PR target/59929
6650 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6651 from push operand if code of push isn't PRE_DEC.
6653 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6655 PR target/59909
6656 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6657 -mquad-memory-atomic. Update -mquad-memory documentation to say
6658 it is only used for non-atomic loads/stores.
6660 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6661 -mquad-memory or -mquad-memory-atomic switches.
6663 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6664 -mquad-memory-atomic to ISA 2.07 support.
6666 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6667 to separate support of normal quad word memory operations (ldq, stq)
6668 from the atomic quad word memory operations.
6670 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6671 support to separate non-atomic quad word operations from atomic
6672 quad word operations. Disable non-atomic quad word operations in
6673 little endian mode so that we don't have to swap words after the
6674 load and before the store.
6675 (quad_load_store_p): Add comment about atomic quad word support.
6676 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6677 options printed with -mdebug=reg.
6679 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6680 -mquad-memory-atomic as the test for whether we have quad word
6681 atomic instructions.
6682 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6683 or -mp8-vector are used, allow byte/half-word atomic operations.
6685 * config/rs6000/sync.md (load_lockedti): Insure that the address
6686 is a proper indexed or indirect address for the lqarx instruction.
6687 On little endian systems, swap the hi/lo registers after the lqarx
6688 instruction.
6689 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6690 insure the address is valid for the lqarx instruction.
6691 (store_conditionalti): Insure that the address is a proper indexed
6692 or indirect address for the stqcrx. instruction. On little endian
6693 systems, swap the hi/lo registers before doing the stqcrx.
6694 instruction.
6695 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6696 insure the address is valid for the stqcrx. instruction.
6698 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6699 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6700 type of quad memory support is available.
6702 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6704 PR regression/59915
6705 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6706 there is a danger of looping.
6708 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6710 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6711 force flag_ira_loop_pressure if set via command line.
6713 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6715 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6716 (ashr_simd): New builtin handling DI mode.
6717 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6718 (aarch64_sshr_simddi): New match pattern.
6719 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6720 (vshrd_n_s64): Likewise.
6721 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6723 2014-01-23 Nick Clifton <nickc@redhat.com>
6725 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6726 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6727 favour of mcu specific scripts.
6728 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6729 430x multilibs.
6731 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6732 Alex Velenko <Alex.Velenko@arm.com>
6734 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6735 (vaddv_s16): Likewise.
6736 (vaddv_s32): Likewise.
6737 (vaddv_u8): Likewise.
6738 (vaddv_u16): Likewise.
6739 (vaddv_u32): Likewise.
6740 (vaddvq_s8): Likewise.
6741 (vaddvq_s16): Likewise.
6742 (vaddvq_s32): Likewise.
6743 (vaddvq_s64): Likewise.
6744 (vaddvq_u8): Likewise.
6745 (vaddvq_u16): Likewise.
6746 (vaddvq_u32): Likewise.
6747 (vaddvq_u64): Likewise.
6748 (vaddv_f32): Likewise.
6749 (vaddvq_f32): Likewise.
6750 (vaddvq_f64): Likewise.
6751 (vmaxv_f32): Likewise.
6752 (vmaxv_s8): Likewise.
6753 (vmaxv_s16): Likewise.
6754 (vmaxv_s32): Likewise.
6755 (vmaxv_u8): Likewise.
6756 (vmaxv_u16): Likewise.
6757 (vmaxv_u32): Likewise.
6758 (vmaxvq_f32): Likewise.
6759 (vmaxvq_f64): Likewise.
6760 (vmaxvq_s8): Likewise.
6761 (vmaxvq_s16): Likewise.
6762 (vmaxvq_s32): Likewise.
6763 (vmaxvq_u8): Likewise.
6764 (vmaxvq_u16): Likewise.
6765 (vmaxvq_u32): Likewise.
6766 (vmaxnmv_f32): Likewise.
6767 (vmaxnmvq_f32): Likewise.
6768 (vmaxnmvq_f64): Likewise.
6769 (vminv_f32): Likewise.
6770 (vminv_s8): Likewise.
6771 (vminv_s16): Likewise.
6772 (vminv_s32): Likewise.
6773 (vminv_u8): Likewise.
6774 (vminv_u16): Likewise.
6775 (vminv_u32): Likewise.
6776 (vminvq_f32): Likewise.
6777 (vminvq_f64): Likewise.
6778 (vminvq_s8): Likewise.
6779 (vminvq_s16): Likewise.
6780 (vminvq_s32): Likewise.
6781 (vminvq_u8): Likewise.
6782 (vminvq_u16): Likewise.
6783 (vminvq_u32): Likewise.
6784 (vminnmv_f32): Likewise.
6785 (vminnmvq_f32): Likewise.
6786 (vminnmvq_f64): Likewise.
6788 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6790 * config/aarch64/aarch64-simd.md
6791 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
6792 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
6793 (*aarch64_mul3_elt<mode>): Likewise.
6794 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
6795 (*aarch64_mul3_elt_to_64v2df): Likewise.
6796 (*aarch64_mla_elt<mode>): Likewise.
6797 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6798 (*aarch64_mls_elt<mode>): Likewise.
6799 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6800 (*aarch64_fma4_elt<mode>): Likewise.
6801 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6802 (*aarch64_fma4_elt_to_64v2df): Likewise.
6803 (*aarch64_fnma4_elt<mode>): Likewise.
6804 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6805 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6806 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6807 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6808 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6809 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6810 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6811 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6812 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6814 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
6816 * config/aarch64/aarch64-simd.md
6817 (aarch64_be_checked_get_lane<mode>): New define_expand.
6818 * config/aarch64/aarch64-simd-builtins.def
6819 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
6820 New builtin definition.
6821 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
6822 Use new safe be builtin.
6824 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6826 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
6827 New define_insn.
6828 (aarch64_be_st1<mode>): Likewise.
6829 (aarch_ld1<VALL:mode>): Define_expand modified.
6830 (aarch_st1<VALL:mode>): Likewise.
6831 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
6832 (UNSPEC_ST1): Likewise.
6834 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
6836 * config/microblaze/microblaze.md: Add trap insn and attribute
6838 2014-01-23 Dodji Seketeli <dodji@redhat.com>
6840 PR preprocessor/58580
6841 * input.h (location_get_source_line): Take an additional line_size
6842 parameter.
6843 (void diagnostics_file_cache_fini): Declare new function.
6844 * input.c (struct fcache): New type.
6845 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
6846 New static constants.
6847 (diagnostic_file_cache_init, total_lines_num)
6848 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
6849 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
6850 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
6851 (get_next_line, read_next_line, goto_next_line, read_line_num):
6852 New static function definitions.
6853 (diagnostic_file_cache_fini): New function.
6854 (location_get_source_line): Take an additional output line_len
6855 parameter. Re-write using lookup_or_add_file_to_cache_tab and
6856 read_line_num.
6857 * diagnostic.c (diagnostic_finish): Call
6858 diagnostic_file_cache_fini.
6859 (adjust_line): Take an additional input parameter for the length
6860 of the line, rather than calculating it with strlen.
6861 (diagnostic_show_locus): Adjust the use of
6862 location_get_source_line and adjust_line with respect to their new
6863 signature. While displaying a line now, do not stop at the first
6864 null byte. Rather, display the zero byte as a space and keep
6865 going until we reach the size of the line.
6866 * Makefile.in: Add vec.o to OBJS-libcommon
6868 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6869 Ilya Tocar <ilya.tocar@intel.com>
6871 * config/i386/avx512fintrin.h (_mm512_kmov): New.
6872 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
6873 (__builtin_ia32_kmov16): Ditto.
6874 * config/i386/i386.md (UNSPEC_KMOV): New.
6875 (kmovw): Ditto.
6877 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6879 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
6880 (_mm512_storeu_si512): Ditto.
6882 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
6884 PR target/52125
6885 * rtl.h (get_referenced_operands): Declare.
6886 * recog.c (get_referenced_operands): New function.
6887 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
6888 operands have been referenced when recording LO_SUM references.
6890 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
6892 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
6894 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
6896 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
6897 Enable for generic and recent AMD targets.
6899 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
6901 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
6902 ARG_SIZE note when adjustment was eliminated.
6904 2014-01-22 Jeff Law <law@redhat.com>
6906 PR tree-optimization/59597
6907 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
6908 in file. Accept new argument REGISTERING and use it to modify
6909 dump output appropriately.
6910 (register_jump_thread): Corresponding changes.
6911 (mark_threaded_blocks): Reinstate code to cancel unprofitable
6912 thread paths involving joiner blocks. Add code to dump cancelled
6913 jump threading paths.
6915 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
6917 PR rtl-optimization/59477
6918 * lra-constraints.c (inherit_in_ebb): Process call for living hard
6919 regs. Update reloads_num and potential_reload_hard_regs for all insns.
6921 2014-01-22 Tom Tromey <tromey@redhat.com>
6923 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
6924 PARAMS.
6925 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
6927 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
6929 PR rtl-optimization/59896
6930 * lra-constraints.c (process_alt_operands): Check unused note for
6931 matched operands of insn with no output reloads.
6933 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
6935 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
6936 (mips_move_from_gpr_cost): Likewise.
6938 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
6940 PR rtl-optimization/59858
6941 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
6942 ira_class_hard_regs_num.
6943 (process_alt_operands): Increase reject for dying matched operand.
6945 2014-01-21 Jakub Jelinek <jakub@redhat.com>
6947 PR target/59003
6948 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
6949 smaller than size, perform several stores or loads and stores
6950 at dst + count - size to store or copy all of size bytes, rather
6951 than just last modesize bytes.
6953 2014-01-20 DJ Delorie <dj@redhat.com>
6955 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
6956 that CLOBBERs are REGs before propogating their values.
6958 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
6960 PR middle-end/59789
6961 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
6962 (cgraph_inline_failed_type): New function.
6963 * cgraph.h (DEFCIFCODE): Add type.
6964 (cgraph_inline_failed_type_t): New enum.
6965 (cgraph_inline_failed_type): New prototype.
6966 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
6967 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
6968 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
6969 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
6970 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
6971 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
6972 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
6973 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
6974 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
6975 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
6976 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
6977 OPTIMIZATION_MISMATCH.
6978 * tree-inline.c (expand_call_inline): Emit errors during
6979 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
6981 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
6983 PR target/59685
6984 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
6985 mode attribute in insn output.
6987 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
6989 * output.h (output_constant): Delete.
6990 * varasm.c (output_constant): Make private.
6992 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
6994 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
6996 2014-01-20 Jakub Jelinek <jakub@redhat.com>
6998 PR middle-end/59860
6999 * tree.h (fold_builtin_strcat): New prototype.
7000 * builtins.c (fold_builtin_strcat): No longer static. Add len
7001 argument, if non-NULL, don't call c_strlen. Optimize
7002 directly into __builtin_memcpy instead of __builtin_strcpy.
7003 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7004 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7006 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7008 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7009 for SImode_address_operand operands, having only a REG argument.
7011 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7013 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7014 loader name using mbig-endian.
7015 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7017 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7019 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7020 (-mtune): Likewise.
7021 (-mcpu): Likewise.
7023 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7025 * config/aarch64/aarch64-protos.h
7026 (aarch64_cannot_change_mode_class_ptr): Declare.
7027 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7028 aarch64_cannot_change_mode_class_ptr): New.
7029 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7030 backend hook aarch64_cannot_change_mode_class.
7032 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7034 * common/config/aarch64/aarch64-common.c
7035 (aarch64_handle_option): Don't handle any option order logic here.
7036 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7037 selected_cpu, warn on architecture version mismatch.
7038 (aarch64_override_options): Fix parsing order for option strings.
7040 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7041 Iain Sandoe <iain@codesourcery.com>
7043 PR bootstrap/59496
7044 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7045 warning. Amend comment to reflect current functionality.
7047 2014-01-20 Richard Biener <rguenther@suse.de>
7049 PR middle-end/59860
7050 * builtins.c (fold_builtin_strcat): Remove case better handled
7051 by tree-ssa-strlen.c.
7053 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7055 * config/aarch64/aarch64.opt
7056 (mcpu, march, mtune): Make case-insensitive.
7058 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7060 PR target/59880
7061 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7062 if operands[1] is a REG or ZERO_EXTEND of a REG.
7064 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7066 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7068 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7070 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7071 long non-pic millicode calls.
7073 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7075 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7077 2014-01-19 Kito Cheng <kito@0xlab.org>
7079 * builtins.c (expand_movstr): Check movstr expand done or fail.
7081 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7082 H.J. Lu <hongjiu.lu@intel.com>
7084 PR target/59379
7085 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7086 to DImode for zero-extended addresses.
7088 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7090 PR rtl-optimization/57763
7091 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7092 on the new indirect jump_insn and increment LABEL_NUSES (label).
7094 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7096 PR bootstrap/59580
7097 PR bootstrap/59583
7098 * config.gcc (x86_archs): New variable.
7099 (x86_64_archs): Likewise.
7100 (x86_cpus): Likewise.
7101 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7102 --with-arch/--with-cpu= options.
7103 Support --with-arch=/--with-cpu={nehalem,westmere,
7104 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7106 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7108 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7109 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7111 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7113 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7115 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7117 PR target/58944
7118 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7119 clear cpp_get_options (parse_in)->warn_unused_macros for
7120 ix86_target_macros_internal with cpp_define.
7122 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7124 * jump.c (delete_related_insns): Keep (use (insn))s.
7125 * reorg.c (redundant_insn): Check for barriers too.
7127 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7129 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7131 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7133 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7134 call to $$dyncall when TARGET_LONG_CALLS is true.
7136 2014-01-17 Jeff Law <law@redhat.com>
7138 * ree.c (combine_set_extension): Temporarily disable test for
7139 changing number of hard registers.
7141 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7143 PR middle-end/58125
7144 * ipa-inline-analysis.c (inline_free_summary):
7145 Do not free summary of aliases.
7147 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7149 PR middle-end/59706
7150 * gimplify.c (gimplify_expr): Use create_tmp_var
7151 instead of create_tmp_var_raw. If cond doesn't have
7152 integral type, don't add the IFN_ANNOTATE builtin at all.
7154 2014-01-17 Martin Jambor <mjambor@suse.cz>
7156 PR ipa/59736
7157 * ipa-cp.c (prev_edge_clone): New variable.
7158 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7159 Also resize prev_edge_clone vector.
7160 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7161 (ipcp_edge_removal_hook): New function.
7162 (ipcp_driver): Register ipcp_edge_removal_hook.
7164 2014-01-17 Andrew Pinski <apinski@cavium.com>
7165 Steve Ellcey <sellcey@mips.com>
7167 PR target/59462
7168 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7169 of operator mode.
7171 2014-01-17 Jeff Law <law@redhat.com>
7173 PR middle-end/57904
7174 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7175 so that pass_ccp runs first.
7177 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7179 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7180 (ix86_adjust_cost): Use !TARGET_XXX.
7181 (do_reorder_for_imul): Likewise.
7182 (swap_top_of_ready_list): Likewise.
7183 (ix86_sched_reorder): Likewise.
7185 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7187 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7188 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7189 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7190 (intel_memset): New. Duplicate slm_memset.
7191 (intel_cost): New. Duplicate slm_cost.
7192 (m_INTEL): New macro.
7193 (processor_target_table): Add "intel".
7194 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7195 with PROCESSOR_INTEL for "intel".
7196 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7197 PROCESSOR_SILVERMONT.
7198 (ix86_issue_rate): Likewise.
7199 (ix86_adjust_cost): Likewise.
7200 (ia32_multipass_dfa_lookahead): Likewise.
7201 (swap_top_of_ready_list): Likewise.
7202 (ix86_sched_reorder): Likewise.
7203 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7204 instead of TARGET_OPT_AGU.
7205 * config/i386/i386.h (TARGET_INTEL): New.
7206 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7207 (processor_type): Add PROCESSOR_INTEL.
7208 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7209 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7211 2014-01-17 Marek Polacek <polacek@redhat.com>
7213 PR c/58346
7214 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7215 size is zero.
7217 2014-01-17 Richard Biener <rguenther@suse.de>
7219 PR tree-optimization/46590
7220 * opts.c (default_options_table): Add entries for
7221 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7222 all enabled at -O1 but not for -Og.
7223 * common.opt (fbranch-count-reg): Remove Init(1).
7224 (fmove-loop-invariants): Likewise.
7225 (ftree-pta): Likewise.
7227 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7229 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7230 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7231 decls to at least the GCC 4.8 used alignments.
7233 PR fortran/59440
7234 * tree-nested.c (convert_nonlocal_reference_stmt,
7235 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7236 of GIMPLE_BIND stmts, adjust associated decls.
7238 2014-01-17 Richard Biener <rguenther@suse.de>
7240 PR tree-optimization/46590
7241 * vec.h (vec<>::bseach): New member function implementing
7242 binary search according to C89 bsearch.
7243 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7244 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7245 bitmap pointer again. Make accesses_in_loop a flat array.
7246 (mem_ref_obstack): New global.
7247 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7248 (mark_ref_stored): Likewise.
7249 (ref_indep_loop_p_2): Likewise.
7250 (set_ref_stored_in_loop): New helper function.
7251 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7252 (memref_free): Adjust.
7253 (record_mem_ref_loc): Simplify.
7254 (gather_mem_refs_stmt): Adjust.
7255 (sort_locs_in_loop_postorder_cmp): New function.
7256 (analyze_memory_references): Sort accesses_in_loop after
7257 loop postorder number.
7258 (find_ref_loc_in_loop_cmp): New function.
7259 (for_all_locs_in_loop): Find relevant cluster of locs in
7260 accesses_in_loop and iterate without recursion.
7261 (execute_sm): Avoid uninit warning.
7262 (struct ref_always_accessed): Simplify.
7263 (ref_always_accessed::operator ()): Likewise.
7264 (ref_always_accessed_p): Likewise.
7265 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7266 loop postorder numbers here.
7267 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7268 numbers.
7270 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7272 PR c++/57945
7273 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7274 on decls for which assemble_alias has been called.
7276 2014-01-17 Nick Clifton <nickc@redhat.com>
7278 * config/msp430/msp430.opt: (mcpu): New option.
7279 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7280 (msp430_option_override): Parse target_cpu. If the MCU name
7281 matches a generic string, clear target_mcu.
7282 (msp430_attr): Allow numeric interrupt values up to 63.
7283 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7284 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7285 option.
7286 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7287 Add mcpu matches.
7288 * config/msp430/msp430.md (popm): Use %J rather than %I.
7289 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7290 (addhi_cy_i): Use immediate_operand for operand 2.
7291 * doc/invoke.texi: Document -mcpu option.
7293 2014-01-17 Richard Biener <rguenther@suse.de>
7295 PR rtl-optimization/38518
7296 * df.h (df_analyze_loop): Declare.
7297 * df-core.c: Include cfgloop.h.
7298 (df_analyze_1): Split out main part of df_analyze.
7299 (df_analyze): Adjust.
7300 (loop_inverted_post_order_compute): New function.
7301 (loop_post_order_compute): Likewise.
7302 (df_analyze_loop): New function avoiding whole-function
7303 postorder computes.
7304 * loop-invariant.c (find_defs): Use df_analyze_loop.
7305 (find_invariants): Adjust.
7306 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7308 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7310 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7311 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7313 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7315 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7316 traversal when removing references.
7318 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7320 PR ipa/59775
7321 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7323 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7325 PR middle-end/56791
7326 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7327 pushing a reload for an autoinc when we had previously reloaded an
7328 inner part of the address.
7330 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7332 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7333 field.
7334 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7335 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7336 when not giving up or versioning for alias only because of
7337 loop->safelen.
7338 (vect_analyze_data_ref_dependences): Set to true.
7339 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7340 is a GIMPLE_PHI.
7341 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7342 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7343 to the condition.
7345 PR middle-end/58344
7346 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7348 PR target/59839
7349 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7350 operand 0 predicate for gathers, use a new pseudo as subtarget.
7352 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7354 PR middle-end/59609
7355 * lra-constraints.c (process_alt_operands): Add printing debug info.
7356 Check absence of input/output reloads for matched operands too.
7358 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7360 PR rtl-optimization/59835
7361 * ira.c (ira_init_register_move_cost): Increase cost for
7362 impossible modes.
7364 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7366 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7368 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7370 PR target/59780
7371 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7372 non-register objects. Use gen_(high/low)part more consistently.
7373 Fix assertions.
7375 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7377 PR target/59844
7378 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7379 endian support, remove tests for WORDS_BIG_ENDIAN.
7380 (p8_mfvsrd_3_<mode>): Likewise.
7381 (reload_gpr_from_vsx<mode>): Likewise.
7382 (reload_gpr_from_vsxsf): Likewise.
7383 (p8_mfvsrd_4_disf): Likewise.
7385 2014-01-16 Richard Biener <rguenther@suse.de>
7387 PR rtl-optimization/46590
7388 * lcm.c (compute_antinout_edge): Use postorder iteration.
7389 (compute_laterin): Use inverted postorder iteration.
7391 2014-01-16 Nick Clifton <nickc@redhat.com>
7393 PR middle-end/28865
7394 * varasm.c (output_constant): Return the number of bytes actually
7395 emitted.
7396 (output_constructor_array_range): Update the field size with the
7397 number of bytes emitted by output_constant.
7398 (output_constructor_regular_field): Likewise. Also do not
7399 complain if the total number of bytes emitted is now greater
7400 than the expected fieldpos.
7401 * output.h (output_constant): Update prototype and descriptive comment.
7403 2014-01-16 Marek Polacek <polacek@redhat.com>
7405 PR middle-end/59827
7406 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7407 it is error_mark_node.
7409 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7411 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7412 VALID_AVX256_REG_OR_OI_MODE.
7414 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7416 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7417 current procedure should be profiled.
7419 2014-01-15 Andrew Pinski <apinski@cavium.com>
7421 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7422 of moving from/to the STACK_REG register class.
7424 2014-01-15 Richard Henderson <rth@redhat.com>
7426 PR debug/54694
7427 * reginfo.c (global_regs_decl): Globalize.
7428 * rtl.h (global_regs_decl): Declare.
7429 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7430 reserved via global_regs.
7432 2014-01-15 Teresa Johnson <tejohnson@google.com>
7434 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7436 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7438 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7439 and vmulosh rather than call gen_vec_widen_smult_*.
7440 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7441 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7442 (vec_widen_smult_even_v16qi): Likewise.
7443 (vec_widen_umult_even_v8hi): Likewise.
7444 (vec_widen_smult_even_v8hi): Likewise.
7445 (vec_widen_umult_odd_v16qi): Likewise.
7446 (vec_widen_smult_odd_v16qi): Likewise.
7447 (vec_widen_umult_odd_v8hi): Likewise.
7448 (vec_widen_smult_odd_v8hi): Likewise.
7449 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7450 vmuloub rather than call gen_vec_widen_umult_*.
7451 (vec_widen_umult_lo_v16qi): Likewise.
7452 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7453 vmulosb rather than call gen_vec_widen_smult_*.
7454 (vec_widen_smult_lo_v16qi): Likewise.
7455 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7456 rather than call gen_vec_widen_umult_*.
7457 (vec_widen_umult_lo_v8hi): Likewise.
7458 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7459 rather than call gen_vec_widen_smult_*.
7460 (vec_widen_smult_lo_v8hi): Likewise.
7462 2014-01-15 Jeff Law <law@redhat.com>
7464 PR tree-optimization/59747
7465 * ree.c (find_and_remove_re): Properly handle case where a second
7466 eliminated extension requires widening a copy created for elimination
7467 of a prior extension.
7468 (combine_set_extension): Ensure that the number of hard regs needed
7469 for a destination register does not change when we widen it.
7471 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7473 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7474 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7475 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7476 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7477 (avr-*-rtems*): Likewise.
7478 (bfin*-rtems*): Likewise.
7479 (moxie-*-rtems*): Likewise.
7480 (h8300-*-rtems*): Likewise.
7481 (i[34567]86-*-rtems*): Likewise.
7482 (lm32-*-rtems*): Likewise.
7483 (m32r-*-rtems*): Likewise.
7484 (m68k-*-rtems*): Likewise.
7485 (microblaze*-*-rtems*): Likewise.
7486 (mips*-*-rtems*): Likewise.
7487 (powerpc-*-rtems*): Likewise.
7488 (sh-*-rtems*): Likewise.
7489 (sparc-*-rtems*): Likewise.
7490 (sparc64-*-rtems*): Likewise.
7491 (v850-*-rtems*): Likewise.
7492 (m32c-*-rtems*): Likewise.
7494 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7496 PR rtl-optimization/59511
7497 * ira.c (ira_init_register_move_cost): Use memory costs for some
7498 cases of register move cost calculations.
7499 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7500 instead of BB frequency.
7501 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7502 * lra-assigns.c (find_hard_regno_for): Ditto.
7504 2014-01-15 Richard Biener <rguenther@suse.de>
7506 PR tree-optimization/59822
7507 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7508 (vectorizable_load): Use it to hoist defs of uses of invariant
7509 loads out of the loop.
7511 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7512 Kugan Vivekanandarajah <kuganv@linaro.org>
7514 PR target/59695
7515 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7516 truncation.
7518 2014-01-15 Richard Biener <rguenther@suse.de>
7520 PR rtl-optimization/59802
7521 * lcm.c (compute_available): Use inverted postorder to seed
7522 the initial worklist.
7524 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7526 PR target/59803
7527 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7528 ADDR_REGS for invalid symrefs in non-PIC code.
7530 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7532 PR other/58712
7533 * builtins.c (determine_block_size): Initialize *probable_max_size
7534 even if len_rtx is CONST_INT.
7536 2014-01-14 Andrew Pinski <apinski@cavium.com>
7538 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7539 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7540 (cortexa53_tunings): Likewise.
7541 (aarch64_sched_issue_rate): New function.
7542 (TARGET_SCHED_ISSUE_RATE): Define.
7544 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7546 * ira-costs.c (find_costs_and_classes): Add missed
7547 ira_init_register_move_cost_if_necessary.
7549 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7551 PR target/59787
7552 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7554 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7556 PR target/59794
7557 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7558 to indicate if type is used for function return value. Warn ABI
7559 change if the vector mode isn't available for function return value.
7560 (ix86_function_arg_advance): Pass false to type_natural_mode.
7561 (ix86_function_arg): Likewise.
7562 (ix86_gimplify_va_arg): Likewise.
7563 (function_arg_32): Don't warn ABI change.
7564 (ix86_function_value): Pass true to type_natural_mode.
7565 (ix86_return_in_memory): Likewise.
7566 (ix86_struct_value_rtx): Removed.
7567 (TARGET_STRUCT_VALUE_RTX): Likewise.
7569 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7571 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7572 converting a conditional jump into a conditional return.
7574 2014-01-14 Richard Biener <rguenther@suse.de>
7576 PR tree-optimization/58921
7577 PR tree-optimization/59006
7578 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7579 hoisting invariant stmts.
7580 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7581 invariant loads on the preheader edge if possible.
7583 2014-01-14 Joey Ye <joey.ye@arm.com>
7585 * doc/plugin.texi (Building GCC plugins): Update to C++.
7587 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7589 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7590 (_mm_rcp28_round_ss): Ditto.
7591 (_mm_rsqrt28_round_sd): Ditto.
7592 (_mm_rsqrt28_round_ss): Ditto.
7593 (_mm_rcp28_sd): Ditto.
7594 (_mm_rcp28_ss): Ditto.
7595 (_mm_rsqrt28_sd): Ditto.
7596 (_mm_rsqrt28_ss): Ditto.
7597 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7598 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7599 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7600 (IX86_BUILTIN_RCP28SD): Ditto.
7601 (IX86_BUILTIN_RCP28SS): Ditto.
7602 (IX86_BUILTIN_RSQRT28SD): Ditto.
7603 (IX86_BUILTIN_RSQRT28SS): Ditto.
7604 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7605 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7606 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7607 (ix86_expand_special_args_builtin): Expand new FTYPE.
7608 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7609 (srcp14<mode>): Make insn unary.
7610 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7611 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7612 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7613 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7614 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7615 Fix rounding: make it SAE only.
7616 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7617 Ditto.
7618 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7619 Ditto.
7620 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7621 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7622 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7623 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7624 (round_saeonly_mask_scalar_operand4): Ditto.
7625 (round_saeonly_mask_scalar_op3): Ditto.
7626 (round_saeonly_mask_scalar_op4): Ditto.
7628 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7630 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7631 Implement -maltivec=be for vec_insert and vec_extract.
7633 2014-01-10 DJ Delorie <dj@redhat.com>
7635 * config/msp430/msp430.md (call_internal): Don't allow memory
7636 references with SP as the base register.
7637 (call_value_internal): Likewise.
7638 * config/msp430/constraints.md (Yc): New. For memory references
7639 that don't use SP as a base register.
7641 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7642 "an integer without a # prefix"
7643 * config/msp430/msp430.md (epilogue_helper): Use it.
7645 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7647 PR target/59617
7648 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7649 AVX512F gather builtins.
7650 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7651 on gather decls with INTEGER_TYPE masktype.
7652 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7653 directly into the builtin rather than hoisting it before loop.
7655 PR tree-optimization/59387
7656 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7657 (scev_const_prop): If folded_casts and type has undefined overflow,
7658 use force_gimple_operand instead of force_gimple_operand_gsi and
7659 for each added stmt if it is assign with
7660 arith_code_with_undefined_signed_overflow, call
7661 rewrite_to_defined_overflow.
7662 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7663 gimple-fold.h instead.
7664 (arith_code_with_undefined_signed_overflow,
7665 rewrite_to_defined_overflow): Moved to ...
7666 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7667 rewrite_to_defined_overflow): ... here. No longer static.
7668 Include gimplify-me.h.
7669 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7670 rewrite_to_defined_overflow): New prototypes.
7672 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7674 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7676 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7678 * builtins.c (get_object_alignment_2): Minor tweak.
7679 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7681 2014-01-13 Christian Bruel <christian.bruel@st.com>
7683 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7684 optimized non constant lengths.
7686 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7688 PR libgomp/59194
7689 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7690 load as __atomic_load_N if possible.
7692 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7694 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7695 target parameter.
7696 (rs6000_expand_builtin): Adjust call.
7698 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7700 PR target/58115
7701 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7702 * config/rs6000/rs6000.c: Include target-globals.h.
7703 (rs6000_set_current_function): Instead of doing target_reinit
7704 unconditionally, use save_target_globals_default_opts and
7705 restore_target_globals.
7707 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7708 FPSCR.
7709 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7710 (rs6000_expand_builtin): Handle mffs and mtfsf.
7711 (rs6000_init_builtins): Define mffs and mtfsf.
7712 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7713 (rs6000_mffs): New pattern.
7714 (rs6000_mtfsf): New pattern.
7716 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7718 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7719 Start narrowing with START. Apply candidate-use pair
7720 and check overall cost in narrowing.
7721 (iv_ca_prune): Pass new argument.
7723 2014-01-10 Jeff Law <law@redhat.com>
7725 PR middle-end/59743
7726 * ree.c (combine_reaching_defs): Ensure the defining statement
7727 occurs before the extension when optimizing extensions with
7728 different source and destination hard registers.
7730 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7732 PR ipa/58585
7733 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7734 vtables into the type inheritance graph.
7736 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7738 PR rtl-optimization/59754
7739 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7740 modes in the REGNO != REGNO case.
7742 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7744 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7746 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7748 PR tree-optimization/59745
7749 * tree-predcom.c (tree_predictive_commoning_loop): Call
7750 free_affine_expand_cache if giving up because components is NULL.
7752 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7753 GC in payload of target_globals struct instead of allocating them on
7754 the heap and the larger structs separately using GC.
7755 * target-globals.h (struct target_globals): Make regs, hard_regs,
7756 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7757 of GTY((skip)) and change type to void *.
7758 (reset_target_globals): Cast loads from those fields to corresponding
7759 types.
7761 2014-01-10 Steve Ellcey <sellcey@mips.com>
7763 PR plugins/59335
7764 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7765 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7766 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7768 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7770 PR target/59744
7771 * aarch64-modes.def (CC_Zmode): New flags mode.
7772 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7773 represents an equality.
7774 (aarch64_get_condition_code): Handle CC_Zmode.
7775 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
7777 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7779 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
7780 extraction in good case.
7782 2014-01-10 Richard Biener <rguenther@suse.de>
7784 PR tree-optimization/59374
7785 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
7786 checking after SLP discovery. Mark stmts not participating
7787 in any SLP instance properly.
7789 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7791 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
7792 when handling a SET rtx.
7794 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7796 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
7797 (cortex-a57): Likewise.
7798 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
7800 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7802 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
7803 non-iwmmxt builtins.
7805 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7807 PR ipa/58252
7808 PR ipa/59226
7809 * ipa-devirt.c record_target_from_binfo): Take as argument
7810 stack of binfos and lookup matching one for virtual inheritance.
7811 (possible_polymorphic_call_targets_1): Update.
7813 2014-01-10 Huacai Chen <chenhc@lemote.com>
7815 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
7816 kernel strings for Loongson-2E/2F/3A.
7818 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7820 PR middle-end/59670
7821 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
7822 is_gimple_call before calling gimple_call_internal_p.
7824 2014-01-09 Steve Ellcey <sellcey@mips.com>
7826 * Makefile.in (TREE_FLOW_H): Remove.
7827 (TREE_SSA_H): Add file names from tree-flow.h.
7828 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
7829 * tree.h: Remove tree-flow.h reference.
7830 * hash-table.h: Remove tree-flow.h reference.
7831 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
7832 reference with tree-ssa-loop.h.
7834 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7836 * doc/invoke.texi: Add -maltivec={be,le} options, and document
7837 default element-order behavior for -maltivec.
7838 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
7839 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
7840 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
7841 when targeting big endian, at least for now.
7842 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
7844 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7846 PR middle-end/47735
7847 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
7848 var satisfies use_register_for_decl, just take into account type
7849 alignment, rather than decl alignment.
7851 PR tree-optimization/59622
7852 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
7853 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
7854 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
7855 Don't devirtualize for inplace at all. For targets.length () == 1,
7856 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
7858 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7860 * config/i386/i386.md (cpu): Remove the unused btver1.
7862 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7864 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
7866 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7868 PR target/58115
7869 * tree-core.h (struct target_globals): New forward declaration.
7870 (struct tree_target_option): Add globals field.
7871 * tree.h (TREE_TARGET_GLOBALS): Define.
7872 (prepare_target_option_nodes_for_pch): New prototype.
7873 * target-globals.h (struct target_globals): Define even if
7874 !SWITCHABLE_TARGET.
7875 * tree.c (prepare_target_option_node_for_pch,
7876 prepare_target_option_nodes_for_pch): New functions.
7877 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
7878 * config/i386/i386.c: Include target-globals.h.
7879 (ix86_set_current_function): Instead of doing target_reinit
7880 unconditionally, use save_target_globals_default_opts and
7881 restore_target_globals.
7883 2014-01-09 Richard Biener <rguenther@suse.de>
7885 PR tree-optimization/59715
7886 * tree-cfg.h (split_critical_edges): Declare.
7887 * tree-cfg.c (split_critical_edges): Export.
7888 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
7890 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
7892 * cfgexpand.c (expand_stack_vars): Optionally disable
7893 asan stack protection.
7894 (expand_used_vars): Likewise.
7895 (partition_stack_vars): Likewise.
7896 * asan.c (asan_emit_stack_protection): Optionally disable
7897 after return stack usage.
7898 (instrument_derefs): Optionally disable memory access instrumentation.
7899 (instrument_builtin_call): Likewise.
7900 (instrument_strlen_call): Likewise.
7901 (asan_protect_global): Optionally disable global variables protection.
7902 * doc/invoke.texi: Added doc for new options.
7903 * params.def: Added new options.
7904 * params.h: Likewise.
7906 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7908 PR rtl-optimization/59724
7909 * ifcvt.c (cond_exec_process_if_block): Don't call
7910 flow_find_head_matching_sequence with 0 longest_match.
7911 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
7912 non-active insns if !stop_after.
7913 (try_head_merge_bb): Revert 2014-01-07 changes.
7915 2014-01-08 Jeff Law <law@redhat.com>
7917 * ree.c (get_sub_rtx): New function, extracted from...
7918 (merge_def_and_ext): Here.
7919 (combine_reaching_defs): Use get_sub_rtx.
7921 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
7923 * cgraph.h (varpool_variable_node): Do not choke on null node.
7925 2014-01-08 Catherine Moore <clm@codesourcery.com>
7927 * config/mips/mips.md (simple_return): Attempt to use JRC
7928 for microMIPS.
7929 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
7931 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
7933 PR rtl-optimization/59137
7934 * reorg.c (steal_delay_list_from_target): Call update_block for
7935 elided insns.
7936 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
7938 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7940 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
7941 two duplicate entries.
7943 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
7945 Revert:
7946 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
7948 * config/mips/mips.c (mips_truncated_op_cost): New function.
7949 (mips_rtx_costs): Adjust test for BADDU.
7950 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
7952 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
7954 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
7955 (*baddu_si): ...this new pattern.
7957 2014-01-08 Jakub Jelinek <jakub@redhat.com>
7959 PR ipa/59722
7960 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
7962 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
7964 PR middle-end/57748
7965 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
7966 inner_reference_p.
7967 (expand_expr, expand_normal): Adjust.
7968 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
7969 inner_reference_p. Use inner_reference_p to expand inner references.
7970 (store_expr): Adjust.
7971 * cfgexpand.c (expand_call_stmt): Adjust.
7973 2014-01-08 Rong Xu <xur@google.com>
7975 * gcov-io.c (gcov_var): Move from gcov-io.h.
7976 (gcov_position): Ditto.
7977 (gcov_is_error): Ditto.
7978 (gcov_rewrite): Ditto.
7979 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
7980 only part to libgcc/libgcov.h.
7982 2014-01-08 Marek Polacek <polacek@redhat.com>
7984 PR middle-end/59669
7985 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
7987 2014-01-08 Marek Polacek <polacek@redhat.com>
7989 PR sanitizer/59667
7990 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
7992 2014-01-08 Jakub Jelinek <jakub@redhat.com>
7994 PR rtl-optimization/59649
7995 * stor-layout.c (get_mode_bounds): For BImode return
7996 0 and STORE_FLAG_VALUE.
7998 2014-01-08 Richard Biener <rguenther@suse.de>
8000 PR middle-end/59630
8001 * gimple.h (is_gimple_builtin_call): Remove.
8002 (gimple_builtin_call_types_compatible_p): New.
8003 (gimple_call_builtin_p): New overload.
8004 * gimple.c (is_gimple_builtin_call): Remove.
8005 (validate_call): Rename to ...
8006 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8007 check return types.
8008 (validate_type): New static function.
8009 (gimple_call_builtin_p): New overload and adjust.
8010 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8011 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8012 (gimple_fold_stmt_to_constant_1): Likewise.
8013 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8015 2014-01-08 Richard Biener <rguenther@suse.de>
8017 PR middle-end/59471
8018 * gimplify.c (gimplify_expr): Gimplify register-register type
8019 VIEW_CONVERT_EXPRs to separate stmts.
8021 2014-01-07 Jeff Law <law@redhat.com>
8023 PR middle-end/53623
8024 * ree.c (combine_set_extension): Handle case where source
8025 and destination registers in an extension insn are different.
8026 (combine_reaching_defs): Allow source and destination registers
8027 in extension to be different under limited circumstances.
8028 (add_removable_extension): Remove restriction that the
8029 source and destination registers in the extension are the same.
8030 (find_and_remove_re): Emit a copy from the extension's
8031 destination to its source after the defining insn if
8032 the source and destination registers are different.
8034 PR middle-end/59285
8035 * ifcvt.c (merge_if_block): If we are merging a block with more than
8036 one successor with a block with no successors, remove any BARRIER
8037 after the second block.
8039 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8041 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8043 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8045 PR target/59652
8046 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8047 for 14-bit register offsets when INT14_OK_STRICT is false.
8049 2014-01-07 Roland Stigge <stigge@antcom.de>
8050 Michael Meissner <meissner@linux.vnet.ibm.com>
8052 PR 57386/target
8053 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8054 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8056 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8058 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8059 -mcpu.
8061 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8063 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8064 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8065 rtx is const0_rtx or not.
8067 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8069 PR target/58115
8070 * target-globals.c (save_target_globals): Remove this_fn_optab
8071 handling.
8072 * toplev.c: Include optabs.h.
8073 (target_reinit): Temporarily restore the global options if another
8074 set of options are in force.
8076 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8078 PR rtl-optimization/58668
8079 * cfgcleanup.c (flow_find_cross_jump): Don't count
8080 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8081 to determine what is counted.
8082 (flow_find_head_matching_sequence): Use active_insn_p to determine
8083 what is counted.
8084 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8085 counting change.
8086 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8087 determine what is counted.
8089 PR tree-optimization/59643
8090 * tree-predcom.c (split_data_refs_to_components): If one dr is
8091 read and one write, determine_offset fails and the write isn't
8092 in the bad component, just put the read into the bad component.
8094 2014-01-07 Mike Stump <mikestump@comcast.net>
8095 Jakub Jelinek <jakub@redhat.com>
8097 PR pch/59436
8098 * tree-core.h (struct tree_optimization_option): Change optabs
8099 type from unsigned char * to void *.
8100 * optabs.c (init_tree_optimization_optabs): Adjust
8101 TREE_OPTIMIZATION_OPTABS initialization.
8103 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8105 PR target/59644
8106 * config/i386/i386.h (struct machine_function): Add
8107 no_drap_save_restore field.
8108 * config/i386/i386.c (ix86_save_reg): Use
8109 !cfun->machine->no_drap_save_restore instead of
8110 crtl->stack_realign_needed.
8111 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8112 this function clears frame_pointer_needed. Set
8113 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8114 and DRAP reg is needed.
8116 2014-01-06 Marek Polacek <polacek@redhat.com>
8118 PR c/57773
8119 * doc/implement-c.texi: Mention that other integer types are
8120 permitted as bit-field types in strictly conforming mode.
8122 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8124 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8125 is newly allocated.
8127 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8129 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8131 2014-01-06 Martin Jambor <mjambor@suse.cz>
8133 PR ipa/59008
8134 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8135 to int.
8136 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8138 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8140 PR debug/59350
8141 PR debug/59510
8142 * var-tracking.c (add_stores): Preserve the value of the source even if
8143 we don't record the store.
8145 2014-01-06 Terry Guo <terry.guo@arm.com>
8147 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8149 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8151 PR bootstrap/59541
8152 * config/darwin.c (darwin_function_section): Adjust return values to
8153 correspond to optimisation changes made in r206070.
8155 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8157 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8158 from prefetch_block tune setting.
8159 (nocona_cost): Correct size of prefetch block to 64.
8161 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8163 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8164 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8165 used to save the static chain register in the computation of the offset
8166 from which the FP registers need to be restored.
8168 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8170 PR tree-optimization/59519
8171 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8172 ICE if get_current_def (current_new_name) is already non-NULL, as long
8173 as it is a phi result of some other phi in *new_exit_bb that has
8174 the same argument.
8176 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8177 or vmovdqu* for misaligned_operand.
8178 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8179 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8180 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8181 aligned_mem for AVX512F masked aligned load and store builtins and for
8182 non-temporal moves.
8184 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8186 PR tree-optimization/59651
8187 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8188 Address range for negative step should be added by TYPE_SIZE_UNIT.
8190 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8192 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8193 overlapping registers also for registers other than the stack pointer.
8195 2014-01-03 Marek Polacek <polacek@redhat.com>
8197 PR other/59661
8198 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8199 __builtin_FILE.
8201 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8203 PR target/59625
8204 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8205 asm goto as jump.
8207 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8208 (push splitter): Use <P:MODE_SIZE> instead of
8209 GET_MODE_SIZE (<P:MODE>mode).
8210 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8211 (mov -1, reg peephole2): Likewise.
8212 * config/i386/sse.md (*mov<mode>_internal,
8213 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8214 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8215 *<code><mode>3, *andnot<mode>3<mask_name>,
8216 <mask_codefor><code><mode>3<mask_name>): Likewise.
8217 * config/i386/subst.md (mask_mode512bit_condition,
8218 sd_mask_mode512bit_condition): Likewise.
8220 2014-01-02 Xinliang David Li <davidxl@google.com>
8222 PR tree-optimization/59303
8223 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8224 (dump_predicates): Better output format.
8225 (pred_equal_p): New function.
8226 (is_neq_relop_p): Ditto.
8227 (is_neq_zero_form_p): Ditto.
8228 (pred_expr_equal_p): Ditto.
8229 (pred_neg_p): Ditto.
8230 (simplify_pred): Ditto.
8231 (simplify_preds_2): Ditto.
8232 (simplify_preds_3): Ditto.
8233 (simplify_preds_4): Ditto.
8234 (simplify_preds): Ditto.
8235 (push_pred): Ditto.
8236 (push_to_worklist): Ditto.
8237 (get_pred_info_from_cmp): Ditto.
8238 (is_degenerated_phi): Ditto.
8239 (normalize_one_pred_1): Ditto.
8240 (normalize_one_pred): Ditto.
8241 (normalize_one_pred_chain): Ditto.
8242 (normalize_preds): Ditto.
8243 (normalize_cond_1): Remove function.
8244 (normalize_cond): Ditto.
8245 (is_gcond_subset_of): Ditto.
8246 (is_subset_of_any): Ditto.
8247 (is_or_set_subset_of): Ditto.
8248 (is_and_set_subset_of): Ditto.
8249 (is_norm_cond_subset_of): Ditto.
8250 (pred_chain_length_cmp): Ditto.
8251 (convert_control_dep_chain_into_preds): Type change.
8252 (find_predicates): Ditto.
8253 (find_def_preds): Ditto.
8254 (destroy_predicates_vecs): Ditto.
8255 (find_matching_predicates_in_rest_chains): Ditto.
8256 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8257 (is_pred_expr_subset): Ditto.
8258 (is_pred_chain_subset_of): Ditto.
8259 (is_included_in): Ditto.
8260 (is_superset_of): Ditto.
8262 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8264 Update copyright years.
8266 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8268 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8269 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8270 config/arc/arc.md, config/arc/arc.opt,
8271 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8272 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8273 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8274 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8275 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8276 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8277 vtable-verify.c, vtable-verify.h: Use the standard form for the
8278 copyright notice.
8280 2014-01-02 Tobias Burnus <burnus@net-b.de>
8282 * gcc.c (process_command): Update copyright notice dates.
8283 * gcov-dump.c: Ditto.
8284 * gcov.c: Ditto.
8285 * doc/cpp.texi: Bump @copying's copyright year.
8286 * doc/cppinternals.texi: Ditto.
8287 * doc/gcc.texi: Ditto.
8288 * doc/gccint.texi: Ditto.
8289 * doc/gcov.texi: Ditto.
8290 * doc/install.texi: Ditto.
8291 * doc/invoke.texi: Ditto.
8293 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8295 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8297 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8299 * config/i386/sse.md (*mov<mode>_internal): Guard
8300 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8302 PR rtl-optimization/59647
8303 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8304 new_rtx into UNSIGNED_FLOAT rtxes.
8305 \f
8306 Copyright (C) 2014 Free Software Foundation, Inc.
8308 Copying and distribution of this file, with or without modification,
8309 are permitted in any medium without royalty provided the copyright
8310 notice and this notice are preserved.