1 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
4 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
5 endianess for subregs into account.
6 * lra-constraints.c (lra_constraints): Do risky transformations
7 always on the first iteration.
8 * lra-lives.c (check_pseudos_live_through_calls): Add arg
9 last_call_used_reg_set.
10 (process_bb_lives): Define and use last_call_used_reg_set.
11 * lra.c (lra): Always continue after lra_constraints on the first
14 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
16 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
18 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
20 2017-01-26 Jakub Jelinek <jakub@redhat.com>
22 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
23 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
24 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
25 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
26 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
27 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
28 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
29 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
30 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
32 2017-01-26 Marek Polacek <polacek@redhat.com>
35 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
36 for the third operand.
38 2017-01-26 Jakub Jelinek <jakub@redhat.com>
41 * omp-low.c (struct omp_context): Add simt_stmt field.
42 (scan_omp_for): Return omp_context *.
43 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
44 context to the _simt_ SIMD stmt.
45 (lower_omp_for): For combined SIMD with sibling _simt_
46 SIMD, make sure to use the same decls in _looptemp_
47 clauses as in the sibling.
49 2017-01-26 David Sherwood <david.sherwood@arm.com>
52 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
55 2017-01-26 Jakub Jelinek <jakub@redhat.com>
58 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
59 emit fld b; fld a; if possible.
61 * brig-builtins.def: Update copyright years.
62 * config/arm/arm_acle_builtins.def: Update copyright years.
64 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
67 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
68 constraint instead of o for the stxsd instruction.
70 2017-01-25 Carl Love <cel@us.ibm.com>
72 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
73 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
75 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
77 * doc/invoke.texi (C++ Dialect Options): Fix typo.
79 2017-01-25 Richard Biener <rguenther@suse.de>
81 PR tree-optimization/69264
82 * target.def (vector_alignment_reachable): Improve documentation.
83 * doc/tm.texi: Regenerate.
84 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
86 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
87 earlier changes with respect to TYPE_USER_ALIGN.
88 (vector_alignment_reachable_p): Likewise. Improve dumping.
90 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
93 * config/arm/arm.md (xordi3): Force constant operand into a register
96 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
98 * doc/invoke.texi (-fstore-merging): Correct default optimization
99 levels at which it is enabled.
100 (-O): Move -fstore-merging from list to...
103 2017-01-25 Richard Biener <rguenther@suse.de>
106 * omp-expand.c: Include debug.h.
107 (expand_omp_taskreg): Make sure to generate early debug before
108 outlining anything from a function.
109 (expand_omp_target): Likewise.
110 (grid_expand_target_grid_body): Likewise.
112 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
115 * asan.c (get_translation_unit_decl): New function.
116 (asan_add_global): Extract modules file name from globals
117 TRANSLATION_UNIT_DECL name.
119 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
122 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
123 for long calls with APCS frame and VFP.
125 2017-01-24 David Malcolm <dmalcolm@redhat.com>
127 * cfg.c (original_copy_tables_initialized_p): New function.
128 * cfg.h (original_copy_tables_initialized_p): New decl.
129 * cfgrtl.c (relink_block_chain): Guard the call to
130 free_original_copy_tables with a call to
131 original_copy_tables_initialized_p.
132 * cgraph.h (symtab_node::native_rtl_p): New decl.
133 * cgraphunit.c (symtab_node::native_rtl_p): New function.
134 (symtab_node::needed_p): Don't assert for early assembly output
136 (cgraph_node::finalize_function): Set "force_output" for __RTL
138 (cgraph_node::analyze): Bail out early for __RTL functions.
139 (analyze_functions): Update assertion to support __RTL functions.
140 (cgraph_node::expand): Bail out early for __RTL functions.
141 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
143 * function.h (struct function): Update comment for field
145 * gimple-expr.c: Include "tree-pass.h".
146 (gimple_has_body_p): Return false for __RTL functions.
147 * Makefile.in (OBJS): Add run-rtl-passes.o.
148 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
150 (gcc::pass_manager::get_clean_slate): New accessor.
151 * passes.c: Include "insn-addr.h".
152 (should_skip_pass_p): Add logging. Update logic for running
153 "expand" to be compatible with both __GIMPLE and __RTL. Guard
154 property-provider override so it is only done for gimple passes.
156 (skip_pass): New function.
157 (execute_one_pass): Call skip_pass when skipping passes.
158 * read-md.c (md_reader::read_char): Support filtering
159 the input to a subset of line numbers.
160 (md_reader::md_reader): Initialize fields
161 m_first_line and m_last_line.
162 (md_reader::read_file_fragment): New function.
163 * read-md.h (md_reader::read_file_fragment): New decl.
164 (md_reader::m_first_line): New field.
165 (md_reader::m_last_line): New field.
166 * read-rtl-function.c (function_reader::create_function): Only
167 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
168 curr_properties. Set DECL_INITIAL to a dummy block.
169 (read_rtl_function_body_from_file_range): New function.
170 * read-rtl-function.h (read_rtl_function_body_from_file_range):
172 * run-rtl-passes.c: New file.
173 * run-rtl-passes.h: New file.
175 2017-01-24 Jeff Law <law@redhat.com>
177 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
180 2017-01-24 Bin Cheng <bin.cheng@arm.com>
182 PR tree-optimization/79159
183 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
184 (record_nonwrapping_iv): Improve boundary using above function if no
185 value range information.
187 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
188 Martin Jambor <mjambor@suse.cz>
190 * brig-builtins.def: New file.
191 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
192 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
193 (DEF_HSAIL_SAT_BUILTIN): Likewise.
194 (DEF_HSAIL_INTR_BUILTIN): Likewise.
195 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
196 * builtin-types.def (BT_INT8): New.
197 (BT_INT16): Likewise.
198 (BT_UINT8): Likewise.
199 (BT_UINT16): Likewise.
200 (BT_FN_ULONG): Likewise.
201 (BT_FN_UINT_INT): Likewise.
202 (BT_FN_UINT_ULONG): Likewise.
203 (BT_FN_UINT_LONG): Likewise.
204 (BT_FN_UINT_PTR): Likewise.
205 (BT_FN_ULONG_PTR): Likewise.
206 (BT_FN_INT8_FLOAT): Likewise.
207 (BT_FN_INT16_FLOAT): Likewise.
208 (BT_FN_UINT32_FLOAT): Likewise.
209 (BT_FN_UINT16_FLOAT): Likewise.
210 (BT_FN_UINT8_FLOAT): Likewise.
211 (BT_FN_UINT64_FLOAT): Likewise.
212 (BT_FN_UINT16_UINT32): Likewise.
213 (BT_FN_UINT32_UINT16): Likewise.
214 (BT_FN_UINT16_UINT16_UINT16): Likewise.
215 (BT_FN_INT_PTR_INT): Likewise.
216 (BT_FN_UINT_PTR_UINT): Likewise.
217 (BT_FN_LONG_PTR_LONG): Likewise.
218 (BT_FN_ULONG_PTR_ULONG): Likewise.
219 (BT_FN_VOID_UINT64_UINT64): Likewise.
220 (BT_FN_UINT8_UINT8_UINT8): Likewise.
221 (BT_FN_INT8_INT8_INT8): Likewise.
222 (BT_FN_INT16_INT16_INT16): Likewise.
223 (BT_FN_INT_INT_INT): Likewise.
224 (BT_FN_UINT_FLOAT_UINT): Likewise.
225 (BT_FN_FLOAT_UINT_UINT): Likewise.
226 (BT_FN_ULONG_UINT_UINT): Likewise.
227 (BT_FN_ULONG_UINT_PTR): Likewise.
228 (BT_FN_ULONG_ULONG_ULONG): Likewise.
229 (BT_FN_UINT_UINT_UINT): Likewise.
230 (BT_FN_VOID_UINT_PTR): Likewise.
231 (BT_FN_UINT_UINT_PTR: Likewise.
232 (BT_FN_UINT32_UINT64_PTR): Likewise.
233 (BT_FN_INT_INT_UINT_UINT): Likewise.
234 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
235 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
236 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
237 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
238 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
239 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
240 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
241 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
242 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
243 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
244 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
245 * doc/frontends.texi: List BRIG FE.
246 * doc/install.texi (Testing): Add BRIG tesring requirements.
247 * doc/invoke.texi (Overall Options): Mention BRIG.
248 * doc/standards.texi (Standards): Doucment BRIG HSA version.
250 2017-01-24 Richard Biener <rguenther@suse.de>
253 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
255 2017-01-24 Martin Jambor <mjambor@suse.cz>
258 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
259 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
262 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
265 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
266 casts from signed to unsigned really don't have a range.
268 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
270 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
271 GMP_RNDx for compatiblity.
273 2017-01-24 Martin Liska <mliska@suse.cz>
276 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
277 that would prevent us to call alloca with -1 as argument.
279 2017-01-24 Jakub Jelinek <jakub@redhat.com>
281 * dwarf2out.c (output_compilation_unit_header, output_file_names):
282 Avoid -Wformat-security warning.
284 2017-01-23 Andrew Pinski <apinski@cavium.com>
286 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
289 2017-01-23 Martin Sebor <msebor@redhat.com>
292 * gimple-ssa-sprintf.c (warn_level): New global.
293 (format_integer): Use it here and throughout the rest of the file.
294 Use the same switch to compute sign as base.
295 (maybe_warn): New function.
296 (format_directive): Factor out warnings into maybe_warn.
297 Add debugging output. Use warn_level.
298 (add_bytes): Use warn_level.
299 (pass_sprintf_length::compute_format_length): Add debugging output.
300 (try_substitute_return_value): Same.
301 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
304 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
305 (struct fmtresult, format_integer, format_floating): Adjust.
306 (fmtresult::fmtresult): Set max correctly in two argument ctor.
307 (get_string_length, format_string,format_directive): Same.
308 (pass_sprintf_length::compute_format_length): Same.
309 (try_substitute_return_value): Simplify slightly.
312 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
313 (fmtresult::operator+=): Outlined.
314 (struct fmtresult): Add ctors.
315 (struct conversion_spec): Rename...
316 (struct directive): ...to this. Add and remove data members.
317 (directive::set_width, directive::set_precision): New functions.
318 (format_percent): Use fmtresult ctor.
319 (get_width_and_precision): Remove.
320 (format_integer): Make naming changes. Avoid computing width and
322 (format_floating): Same. Adjust indentation.
323 (format_character, format_none): New functions.
324 (format_string): Moved character handling to format_character.
325 (format_directive): Remove arguments, change return type.
326 (parse_directive): New function.
327 (pass_sprintf_length::compute_format_length): Move directive
328 parsing to parse_directive.
330 2017-01-23 Jakub Jelinek <jakub@redhat.com>
332 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
333 (assign_assembler_name_if_needed): ... this.
334 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
335 (assign_assembler_name_if_needed): ... this.
336 (free_lang_data_in_cgraph): Adjust callers.
337 * cgraphunit.c (cgraph_node::analyze): Likewise.
338 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
341 2017-01-23 Richard Biener <rguenther@suse.de>
343 PR tree-optimization/79088
344 PR tree-optimization/79188
345 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
346 resetting loop bounds after last path deletion. Reset loop
347 bounds of the target loop, make code match the comments.
348 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
349 Make sure loops need no fixups.
351 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
353 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
354 exponent support with double type for first argument.
355 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
356 type returned by __builtin_vec_extract_sig,
357 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
358 functions from "vector int" to "vector unsigned int" or from
359 "vector long long int" to "vector unsigned long long int".
360 Changed type returned by __builtin_vec_extract_exp,
361 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
362 functions from "vector int" to "vector unsigned int" or from
363 "vector long long int" to "vector unsigned long long int".
364 Changed return type of __builtin_vec_test_data_class,
365 __builtin_vec_test_data_class_sp, and
366 __builtin_vec_test_data_class_dp from "vector int" to
367 "vector bool int" or from "vector long long int" to "vector bool
368 long long int" and changed second argument type from "unsigned
369 int" to "int". Added new overloaded function forms "vector float
370 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
371 "vector float __builtin_vec_insert_exp_sp (vector float, vector
372 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
373 double, vector unsigned long long int)" and "vector double
374 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
375 long int)". Changed return type of
376 __builtin_scalar_test_data_class and
377 __builtin_scalar_test_data_class_sp and
378 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
379 int" and changed second argument from "unsigned int" to "int".
380 Changed type returned by __builtin_scalar_test_neg,
381 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
382 from "int" to "bool int". Added new overloaded function form
383 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
384 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
385 exponent double-precision with floating point first argument.
386 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
387 documentation of scalar_test_data_class, scalar_test_neg,
388 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
389 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
390 vec_test_data_class built-in functions to reflect refinements in
391 their type signatures.
393 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
395 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
397 (aarch64_elf_asm_destructor): Likewise.
399 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
401 PR rtl-optimization/78634
402 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
403 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
404 * ifcvt.c (noce_try_cmove): Add missing cost check.
406 PR rtl-optimization/71724
407 * combine.c (if_then_else_cond): Look for situations where it is
408 beneficial to undo the work of one of the recursive calls.
410 2017-01-23 Bin Cheng <bin.cheng@arm.com>
412 PR tree-optimization/70754
413 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
414 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
415 combined stmt before it if not NULL.
416 (combine_chains): Process refs reversely and compute dominance point
419 2017-01-23 Martin Liska <mliska@suse.cz>
421 PR tree-optimization/79196
422 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
423 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
425 (strlen_optimize_stmt): Call the renamed function.
427 2017-01-23 Michael Matz <matz@suse.de>
429 PR tree-optimization/78384
430 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
432 2017-01-23 Richard Biener <rguenther@suse.de>
434 PR tree-optimization/79186
435 * tree-vrp.c (register_new_assert_for): Make sure we've seen
436 both incoming edges before moving an assert.
438 2017-01-23 Martin Jambor <mjambor@suse.cz>
440 * ipa-prop.c (load_from_param_1): Removed.
441 (load_from_unmodified_param): Bits from load_from_param_1 put back
443 (load_from_param): Removed.
444 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
445 with stmt. Reverted back to use of load_from_unmodified_param.
447 2017-01-23 Martin Jambor <mjambor@suse.cz>
450 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
451 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
452 field a pointer to garbage collected vector, mark lattices and
453 ipcp_orig_node with GTY((skip)).
454 (ipa_get_param_count): Adjust to descriptors being a pointer.
455 (ipa_get_param): Likewise.
456 (ipa_get_type): Likewise.
457 (ipa_get_param_move_cost): Likewise.
458 (ipa_set_param_used): Likewise.
459 (ipa_get_controlled_uses): Likewise.
460 (ipa_set_controlled_uses): Likewise.
461 (ipa_is_param_used): Likewise.
462 (ipa_node_params_t): Move into garbage collector. New methods insert
464 (ipa_node_params_sum): Annotate wth GTY(()).
465 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
467 (ipa_load_from_parm_agg): Adjust declaration.
468 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
469 * ipa-profile.c (ipa_profile): Likewise.
470 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
471 (ipa_populate_param_decls): Make descriptors parameter garbage
473 (ipa_dump_param): Adjust to descriptors being a pointer.
474 (ipa_alloc_node_params): Likewise.
475 (ipa_initialize_node_params): Likewise.
476 (load_from_param_1): Make descriptors parameter garbage collected.
477 (load_from_unmodified_param): Likewise.
478 (load_from_param): Likewise.
479 (ipa_load_from_parm_agg): Likewise.
480 (ipa_node_params::~ipa_node_params): Removed.
481 (ipa_free_all_node_params): Remove call to delete operator.
482 (ipa_node_params_t::insert): New.
483 (ipa_node_params_t::remove): Likewise.
484 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
485 copy known_csts and known_contexts vectors.
486 (ipa_read_node_info): Adjust to descriptors being a pointer.
487 (ipcp_modif_dom_walker): Make m_descriptors field garbage
489 (ipcp_transform_function): Make descriptors variable garbage
492 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
494 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
495 * config/i386/avx512dqintrin.h: Ditto.
496 * config/i386/avx512fintrin.h: Ditto.
497 * gcc/config/i386/i386.c: Handle new builtins.
498 * config/i386/i386-builtin.def: Add new builtins.
499 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
500 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
502 2017-01-23 Jakub Jelinek <jakub@redhat.com>
503 Martin Liska <mliska@suse.cz>
505 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
506 * asan.c (asan_expand_poison_ifn): Support stores and use
507 appropriate ASAN report function.
508 * internal-fn.c (expand_ASAN_POISON_USE): New function.
509 * internal-fn.def (ASAN_POISON_USE): Declare.
510 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
511 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
512 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
513 ASAN_POISON calls w/o LHS.
514 * tree-ssa.c (execute_update_addresses_taken): Create clobber
515 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
516 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
517 * gimplify.c (asan_poison_variables): Add attribute
518 use_after_scope_memory to variables that really needs to live
520 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
521 having the attribute.
523 2017-01-23 Martin Liska <mliska@suse.cz>
525 * asan.c (create_asan_shadow_var): New function.
526 (asan_expand_poison_ifn): Likewise.
527 * asan.h (asan_expand_poison_ifn): New declaration.
528 * internal-fn.c (expand_ASAN_POISON): Likewise.
529 * internal-fn.def (ASAN_POISON): New builtin.
530 * sanopt.c (pass_sanopt::execute): Expand
531 asan_expand_poison_ifn.
532 * tree-inline.c (copy_decl_for_dup_finish): Make function
534 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
535 * tree-ssa.c (is_asan_mark_p): New function.
536 (execute_update_addresses_taken): Rewrite local variables
537 (identified just by use-after-scope as addressable) into SSA.
539 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
541 * doc/install.texi (Specific): opensource.apple.com uses https
542 now. Remove trailing slash.
544 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
546 * README.Portability: Remove note on an Irix compatibility issue.
548 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
550 * gcov.c (INCLUDE_ALGORITHM): Define.
551 (INCLUDE_VECTOR): Define.
552 No longer include <vector> and <algorithm> directly.
554 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
556 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
558 * doc/invoke.texi (Code Gen Options): Ditto.
560 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
563 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
565 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
567 rtl-optimization/79125
568 * cprop.c (local_cprop_pass): Handle cases where we make an
571 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
575 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
576 read from, for big endian.
578 2017-01-20 Jiong Wang <jiong.wang@arm.com>
580 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
581 register pauth builtins for LP64 only.
583 2017-01-20 Marek Polacek <polacek@redhat.com>
586 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
589 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
591 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
593 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
594 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
595 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
597 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
600 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
601 in big-endian mode when they are not a single duplicated value.
603 2017-01-20 Richard Biener <rguenther@suse.de>
605 * BASE-VER: Bump to 7.0.1.
607 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
609 * omp-low.c (omplow_simd_context): New struct. Use it...
610 (lower_rec_simd_input_clauses): ...here and...
611 (lower_rec_input_clauses): ...here to hold common data. Adjust all
612 references to idx, lane, max_vf, is_simt.
614 2017-01-20 Graham Markall <graham.markall@embecosm.com>
616 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
619 2017-01-20 Martin Jambor <mjambor@suse.cz>
621 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
622 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
624 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
625 (GTFILES): Rename hsa.c to hsa-common.c.
626 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
627 * hsa-dump.c: Likewise.
628 * hsa-gen.c: Likewise.
629 * hsa-regalloc.c: Likewise.
630 * ipa-hsa.c: Likewise.
631 * omp-expand.c: Likewise.
632 * omp-low.c: Likewise.
633 * toplev.c: Likewise.
635 2017-01-20 Marek Polacek <polacek@redhat.com>
638 * doc/invoke.texi: Document -Wduplicated-branches.
639 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
640 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
641 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
642 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
643 return 0 only when not OEP_LEXICOGRAPHIC.
644 (fold_build_cleanup_point_expr): Use the expression
645 location when building CLEANUP_POINT_EXPR.
646 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
647 * tree.c (add_expr): Handle error_mark_node.
649 2017-01-20 Martin Liska <mliska@suse.cz>
652 * tree-profile.c (init_ic_make_global_vars): Do not call
654 (gimple_init_gcov_profiler): Likewise.
656 2017-01-20 Martin Liska <mliska@suse.cz>
659 * cgraph.h (maybe_create_reference): Remove argument and
661 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
663 * ipa-cp.c (create_specialized_node): Likewise.
664 * symtab.c (symtab_node::maybe_create_reference): Handle
665 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
667 2017-01-20 Martin Liska <mliska@suse.cz>
669 * read-rtl-function.c (function_reader::create_function): Use
670 build_decl instread of build_decl_stat.
672 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
674 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
675 * config/i386/avx512dqintrin.h: Ditto.
676 * config/i386/avx512fintrin.h: Ditto.
677 * config/i386/i386-builtin-types.def: Add new types.
678 * gcc/config/i386/i386.c: Handle new types.
679 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
680 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
681 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
682 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
683 (__builtin_ia32_kshiftridi): New.
684 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
686 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
690 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
691 define to rs6000_init_stack_protect_guard.
692 (rs6000_init_stack_protect_guard): New function.
694 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
695 Yunqiang Su <yunqiang.su@imgtec.com>
697 * config.gcc (supported_defaults): Add madd4.
698 (with_madd4): Add validation.
699 (all_defaults): Add madd4.
700 * config/mips/mips.opt (mmadd4): New option.
701 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
703 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
705 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
706 (ISA_HAS_FUSED_MADD4): Likewise.
707 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
708 * gcc/doc/install.texi (--with-madd4): Document the new option.
710 2017-01-19 Jiong Wang <jiong.wang@arm.com>
712 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
713 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
714 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
715 (aarch64_init_pauth_hint_builtins): New.
716 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
717 (aarch64_expand_builtin): Expand new builtins.
719 2017-01-19 Jiong Wang <jiong.wang@arm.com>
721 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
722 * combine-stack-adj.c (no_unhandled_cfa): Handle
723 REG_CFA_TOGGLE_RA_MANGLE.
724 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
725 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
726 info for return address signing.
727 (aarch64_expand_epilogue): Likewise.
729 2017-01-19 Jiong Wang <jiong.wang@arm.com>
731 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
732 * config/aarch64/aarch64-protos.h
733 (aarch64_return_address_signing_enabled): New declaration.
734 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
736 (aarch64_expand_prologue): Sign return address before it's pushed onto
738 (aarch64_expand_epilogue): Authenticate return address fetched from
740 (aarch64_override_options): Sanity check for ILP32 and ISA level.
741 (aarch64_attributes): New function attributes for "sign-return-address".
742 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
743 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
744 ("*do_return"): Generate combined instructions according to key index.
745 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
746 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
748 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
749 * config/aarch64/aarch64.opt (msign-return-address=): New.
750 * doc/extend.texi (AArch64 Function Attributes): Documents
751 "sign-return-address=".
752 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
754 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
756 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
757 overall option summary.
759 2017-01-19 Jiong Wang <jiong.wang@arm.com>
761 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
762 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
763 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
764 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
766 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
768 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
769 -mpower9-minmax by default for -mcpu=power9.
770 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
771 128-bit floating point.
773 2017-01-20 Alan Modra <amodra@gmail.com>
775 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
778 2017-01-20 Alan Modra <amodra@gmail.com>
781 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
782 for strcmp and strncmp from corresponding builtin decl.
784 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
786 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
787 instead of i386/rtems-64.h.
788 * config/i386/rtems-64.h: Remove.
790 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
794 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
796 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
798 2017-01-19 Tamar Christina <tamar.christina@arm.com>
800 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
801 Change int to HOST_WIDE_INT.
802 * config/aarch64/aarch64-protos.h
803 (aarch64_simd_gen_const_vector_dup): Likewise.
804 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
806 2017-01-19 David Malcolm <dmalcolm@redhat.com>
808 * langhooks-def.h (lhd_type_for_size): New decl.
809 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
810 * langhooks.c (lhd_type_for_size): New function, taken from
813 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
815 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
816 define_bypass for CR latency.
817 (power9-cracked-alu): Update bypass latency and remove power9-branch.
818 (power9-alu2): Add define_bypass for CR latency.
820 (power9-mul): Update insn latency.
821 (power9-mul-compare): Update insn latency, bypass latency and remove
824 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
826 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
828 * config/aarch64/aarch64.md
829 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
830 aarch64_nopcrelative_literal_loads.
831 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
833 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
835 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
837 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
839 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
842 * config.gcc (supported_defaults): Add lxc1-sxc1.
843 (with_lxc1_sxc1): Add validation.
844 (all_defaults): Add lxc1-sxc1.
845 * config/mips/mips.opt (mlxc1-sxc1): New option.
846 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
848 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
850 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
851 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
852 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
854 2017-01-19 Richard Biener <rguenther@suse.de>
856 PR tree-optimization/72488
857 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
858 sure to restore SSA info.
859 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
861 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
863 PR rtl-optimization/79121
864 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
865 of the inner type when shifting an extended value.
867 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
870 * symtab.c (symtab_node::equal_address_to): Fix comparing of
871 interposable aliases.
873 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
876 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
877 Use the evmergelohi instruction.
878 (mov_si<mode>_e500_subreg4_2_le): Likewise.
879 (mov_sitf_e500_subreg8_2_be): Likewise.
880 (mov_sitf_e500_subreg12_2_le): Likewise.
881 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
882 (mov_si<mode>_e500_subreg4_2_be): Likewise.
883 (mov_sitf_e500_subreg8_2_le): Likewise.
884 (mov_sitf_e500_subreg12_2_be): Likewise.
886 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
888 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
889 attribute from vecsimple to vecperm.
890 (altivec_vbpermq2): Likewise.
892 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
895 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
897 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
898 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
899 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
900 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
901 case where N arg is SIZE_MAX.
902 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
903 (cmpstrsi): Add pattern.
905 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
907 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
908 __builtin_vec_revb builtins.
909 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
910 built-in functions to support generation of the ISA 3.0 XXBR<x>
911 vector byte reverse instructions.
912 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
913 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
914 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
915 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
916 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
917 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
918 (P9V_BUILTIN_VEC_REVB): Likewise.
919 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
920 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
921 (p9_xxbrq_v16qi): Likewise.
922 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
923 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
924 (p9_xxbrh_v8hi): Likewise.
925 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
926 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
927 vec_revb built-in functions.
929 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
931 PR rtl-optimization/78952
932 * config/i386/i386.md (any_extract): New code iterator.
933 (*insvqi_2): Use any_extract for source operand.
934 (*insvqi_3): Use any_shiftrt for source operand.
936 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
938 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
940 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
942 2017-01-18 Matthias Klose <doko@ubuntu.com>
944 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
946 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
948 * config/rs6000/altivec.h (vec_bperm): Change #define.
949 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
950 (altivec_vbpermq2): New define_insn.
951 (altivec_vbpermd): Likewise.
952 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
955 (VBPERM): New polymorphic function interface.
956 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
957 Add entries for P9V_BUILTIN_VEC_VBPERM.
958 * doc/extend.texi: Add interfaces for vec_bperm.
960 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
962 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
963 first letter of error messages.
964 (s390_resolve_overloaded_builtin): Likewise.
965 * config/s390/s390.c (s390_expand_builtin): Likewise.
966 (s390_invalid_arg_for_unprototyped_fn): Likewise.
967 (s390_valid_target_attribute_inner_p): Likewise.
968 * config/s390/s390.md ("tabort"): Likewise.
970 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
972 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
973 (ISA_AVOID_DIV_HILO): New macro.
974 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
975 (ISA_HAS_DDIV): Likewise.
977 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
979 * doc/invoke.texi (fabi-version): Correct number of occurrences.
981 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
983 * doc/invoke.texi (fabi-version): Spelling fix.
985 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
988 * doc/invoke.texi (fabi-version): Mention mangling fix for
991 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
994 * doc/invoke.texi (fabi-version): Document discriminator mangling.
996 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
999 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
1000 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
1002 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
1004 (stack_protect_test): Ditto.
1005 * config/rs6000/rs6000.opt (mstack-protector-guard=,
1006 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
1008 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
1009 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
1010 -mstack-protector-guard-offset=.
1011 (RS/6000 and PowerPC Options): Ditto.
1013 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1015 * config/i386/i386.h (MASK_CLASS_P): New define.
1016 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
1017 there are no registers from different register sets also when
1018 mask registers are used. Update function comment.
1019 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
1020 to (*k/*r) and (*k/*km) alternatives.
1022 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
1024 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
1025 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
1026 (EH_RETURN_HANDLER_RTX): New define.
1027 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
1028 Force frame pointer in EH return functions.
1029 (aarch64_expand_epilogue): Add barrier for eh_return.
1030 (aarch64_final_eh_return_addr): Remove.
1031 (aarch64_eh_return_handler_rtx): New function.
1032 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
1034 (aarch64_eh_return_handler_rtx): New prototype.
1036 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1038 * config/rs6000/altivec.h (vec_rlmi): New #define.
1039 (vec_vrlnm): Likewise.
1040 (vec_rlnm): Likewise.
1041 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
1042 (UNSPEC_VRLNM): Likewise.
1043 (VIlong): New mode iterator.
1044 (altivec_vrl<VI_char>mi): New define_insn.
1045 (altivec_vrl<VI_char>nm): Likewise.
1046 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
1049 (RLNM): New polymorphic function entry.
1050 (VRLWMI): New monomorphic function entry.
1052 (RLMI): New polymorphic function entry.
1053 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
1054 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
1055 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
1058 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1061 * dwarf2out.c (field_byte_offset): Restore the
1062 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
1063 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
1064 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
1067 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1070 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
1072 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1075 * dwarf2out.c (add_data_member_location_attribute): For constant
1076 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
1077 instead of DW_AT_data_member_location, DW_AT_bit_offset and
1078 DW_AT_byte_size attributes.
1080 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1082 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
1083 after forcing to constant memory when the code model is medium.
1085 2017-01-17 Julia Koval <julia.koval@intel.com>
1088 * config/i386/avx512fintrin.h
1089 (_mm512_i32gather_ps): Change __addr type to void const*.
1090 (_mm512_mask_i32gather_ps): Ditto.
1091 (_mm512_i32gather_pd): Ditto.
1092 (_mm512_mask_i32gather_pd): Ditto.
1093 (_mm512_i64gather_ps): Ditto.
1094 (_mm512_mask_i64gather_ps): Ditto.
1095 (_mm512_i64gather_pd): Ditto.
1096 (_mm512_mask_i64gather_pd): Ditto.
1097 (_mm512_i32gather_epi32): Ditto.
1098 (_mm512_mask_i32gather_epi32): Ditto.
1099 (_mm512_i32gather_epi64): Ditto.
1100 (_mm512_mask_i32gather_epi64): Ditto.
1101 (_mm512_i64gather_epi32): Ditto.
1102 (_mm512_mask_i64gather_epi32): Ditto.
1103 (_mm512_i64gather_epi64): Ditto.
1104 (_mm512_mask_i64gather_epi64): Ditto.
1105 (_mm512_i32scatter_ps): Change __addr type to void*.
1106 (_mm512_mask_i32scatter_ps): Ditto.
1107 (_mm512_i32scatter_pd): Ditto.
1108 (_mm512_mask_i32scatter_pd): Ditto.
1109 (_mm512_i64scatter_ps): Ditto.
1110 (_mm512_mask_i64scatter_ps): Ditto.
1111 (_mm512_i64scatter_pd): Ditto.
1112 (_mm512_mask_i64scatter_pd): Ditto.
1113 (_mm512_i32scatter_epi32): Ditto.
1114 (_mm512_mask_i32scatter_epi32): Ditto.
1115 (_mm512_i32scatter_epi64): Ditto.
1116 (_mm512_mask_i32scatter_epi64): Ditto.
1117 (_mm512_i64scatter_epi32): Ditto.
1118 (_mm512_mask_i64scatter_epi32): Ditto.
1119 (_mm512_i64scatter_epi64): Ditto.
1120 (_mm512_mask_i64scatter_epi64): Ditto.
1121 * config/i386/avx512pfintrin.h
1122 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1123 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1124 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1125 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1126 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1127 (_mm512_prefetch_i32scatter_ps): Ditto.
1128 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1129 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1130 (_mm512_prefetch_i64scatter_pd): Ditto.
1131 (_mm512_prefetch_i64scatter_ps): Ditto.
1132 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1133 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1134 * config/i386/avx512vlintrin.h
1135 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1136 (_mm_mmask_i32gather_ps): Ditto.
1137 (_mm256_mmask_i32gather_pd): Ditto.
1138 (_mm_mmask_i32gather_pd): Ditto.
1139 (_mm256_mmask_i64gather_ps): Ditto.
1140 (_mm_mmask_i64gather_ps): Ditto.
1141 (_mm256_mmask_i64gather_pd): Ditto.
1142 (_mm_mmask_i64gather_pd): Ditto.
1143 (_mm256_mmask_i32gather_epi32): Ditto.
1144 (_mm_mmask_i32gather_epi32): Ditto.
1145 (_mm256_mmask_i32gather_epi64): Ditto.
1146 (_mm_mmask_i32gather_epi64): Ditto.
1147 (_mm256_mmask_i64gather_epi32): Ditto.
1148 (_mm_mmask_i64gather_epi32): Ditto.
1149 (_mm256_mmask_i64gather_epi64): Ditto.
1150 (_mm_mmask_i64gather_epi64): Ditto.
1151 (_mm256_i32scatter_ps): Change __addr type to void*.
1152 (_mm256_mask_i32scatter_ps): Ditto.
1153 (_mm_i32scatter_ps): Ditto.
1154 (_mm_mask_i32scatter_ps): Ditto.
1155 (_mm256_i32scatter_pd): Ditto.
1156 (_mm256_mask_i32scatter_pd): Ditto.
1157 (_mm_i32scatter_pd): Ditto.
1158 (_mm_mask_i32scatter_pd): Ditto.
1159 (_mm256_i64scatter_ps): Ditto.
1160 (_mm256_mask_i64scatter_ps): Ditto.
1161 (_mm_i64scatter_ps): Ditto.
1162 (_mm_mask_i64scatter_ps): Ditto.
1163 (_mm256_i64scatter_pd): Ditto.
1164 (_mm256_mask_i64scatter_pd): Ditto.
1165 (_mm_i64scatter_pd): Ditto.
1166 (_mm_mask_i64scatter_pd): Ditto.
1167 (_mm256_i32scatter_epi32): Ditto.
1168 (_mm256_mask_i32scatter_epi32): Ditto.
1169 (_mm_i32scatter_epi32): Ditto.
1170 (_mm_mask_i32scatter_epi32): Ditto.
1171 (_mm256_i32scatter_epi64): Ditto.
1172 (_mm256_mask_i32scatter_epi64): Ditto.
1173 (_mm_i32scatter_epi64): Ditto.
1174 (_mm_mask_i32scatter_epi64): Ditto.
1175 (_mm256_i64scatter_epi32): Ditto.
1176 (_mm256_mask_i64scatter_epi32): Ditto.
1177 (_mm_i64scatter_epi32): Ditto.
1178 (_mm_mask_i64scatter_epi32): Ditto.
1179 (_mm256_i64scatter_epi64): Ditto.
1180 (_mm256_mask_i64scatter_epi64): Ditto.
1181 (_mm_i64scatter_epi64): Ditto.
1182 (_mm_mask_i64scatter_epi64): Ditto.
1183 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1184 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1185 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1186 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1187 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1188 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1189 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1190 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1191 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1192 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1193 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1194 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1195 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1196 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1197 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1198 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1199 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1200 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1201 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1202 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1203 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1204 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1205 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1206 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1207 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1208 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1209 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1210 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1211 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1212 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1213 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1214 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1215 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1216 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1217 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1218 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1219 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1220 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1221 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1222 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1223 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1224 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1225 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1226 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1227 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1228 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1229 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1230 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1231 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1232 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1233 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1234 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1235 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1236 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1237 definitions accordingly.
1239 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1240 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1243 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1246 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1249 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1250 hard regs for inner regno.
1252 2017-01-17 Martin Liska <mliska@suse.cz>
1255 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1256 assumption and add comment.
1258 2017-01-17 Nathan Sidwell <nathan@acm.org>
1260 * ipa-visibility.c (localize_node): New function, broken out of ...
1261 (function_and_variable_visibility): ... here. Call it.
1263 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1266 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1267 correctly set frequency of oudgoing edge.
1268 (duplicate_thread_path): Fix profile updating.
1270 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1273 * configure.ac: Add GCC_BASE_VER.
1274 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1275 version from BASE-VER file.
1276 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1277 (gcc.o): Depend on $(BASEVER).
1278 * common.opt (dumpfullversion): New option.
1279 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1280 * doc/invoke.texi: Document -dumpfullversion.
1281 * doc/install.texi: Document --with-gcc-major-version-only.
1282 * configure: Regenerated.
1284 2017-01-17 Richard Biener <rguenther@suse.de>
1286 PR tree-optimization/71433
1287 * tree-vrp.c (register_new_assert_for): Merge same asserts
1288 on all incoming edges.
1289 (process_assert_insertions_for): Handle insertions at the
1292 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1294 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1295 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1297 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1300 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1303 2017-01-17 Alan Modra <amodra@gmail.com>
1306 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1307 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1308 symbolic stack limit when pic.
1310 2017-01-16 Martin Sebor <msebor@redhat.com>
1312 PR tree-optimization/78608
1313 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1315 2017-01-16 Jeff Law <law@redhat.com>
1318 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1319 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1320 for several include directories that may be relative to sysroot.
1321 * config/i386/x-mingw32 (gplus_includedir): Define.
1322 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1323 (native_system_includedir): Likewise.
1324 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1325 override if TARGET_SYSTEM_ROOT is defined.
1326 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1328 PR tree-optimization/79090
1329 PR tree-optimization/33562
1330 PR tree-optimization/61912
1331 PR tree-optimization/77485
1332 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1333 and computed trims into the dump file.
1335 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1337 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1339 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1342 * gimplify.c (gimplify_init_constructor): If want_value and
1343 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1347 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1348 sequence. Formatting fixes.
1349 (doloop_optimize): Formatting fixes.
1352 * gcc.c (debug_level_greater_than_spec_func): New function.
1353 (static_spec_functions): Add debug-level-gt spec function.
1354 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
1356 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1357 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1358 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
1359 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
1360 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
1361 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
1363 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
1365 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
1366 QImode fixups to general and mask registers only.
1368 2017-01-16 Carl Love <cel@us.ibm.com>
1370 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
1371 for built-in functions
1372 vector signed char vec_nabs (vector signed char)
1373 vector signed short vec_nabs (vector signed short)
1374 vector signed int vec_nabs (vector signed int)
1375 vector signed long long vec_nabs (vector signed long long)
1376 vector float vec_nabs (vector float)
1377 vector double vec_nabs (vector double)
1378 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
1380 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
1381 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
1382 * doc/extend.texi: Update the documentation file for the new built-in
1385 2017-01-16 Martin Sebor <msebor@redhat.com>
1387 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
1390 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1392 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
1393 UNSPEC_VSX__XXSPLTD to require special splat handling.
1395 2017-01-16 David Malcolm <dmalcolm@redhat.com>
1398 * system.h: Poison strndup.
1400 2017-01-16 Alan Modra <amodra@gmail.com>
1403 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
1406 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
1408 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
1410 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
1412 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
1413 call recog here. Assert that INSN_CODE (insn) is non-negative.
1415 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1418 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
1420 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
1421 in the currently scheduled RTL fragment.
1423 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1425 PR rtl-optimization/78751
1426 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
1429 2017-01-14 Jeff Law <law@redhat.com>
1431 PR tree-optimization/79090
1432 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
1433 variable length stores.
1434 (compute_trims): Delete dead assignment to *trim_tail.
1435 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
1438 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
1440 PR rtl-optimization/78626
1441 PR rtl-optimization/78727
1442 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
1443 of a block, and split such blocks after everything else is finished.
1445 2017-01-14 Alan Modra <amodra@gmail.com>
1448 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
1449 target legitimate_combined_insn.
1450 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
1451 (rs6000_legitimate_combined_insn): New function.
1452 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
1454 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
1455 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
1456 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
1458 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
1460 * doc/frontends.texi (G++ and GCC): Remove references to Java.
1462 2017-01-13 Jeff Law <law@redhat.com>
1464 PR tree-optimization/33562
1465 PR tree-optimization/61912
1466 PR tree-optimization/77485
1467 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
1469 (delete_dead_assignment): Likewise.
1470 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
1471 statement to delete_dead_call and delete_dead_assignment.
1473 2017-01-13 David Malcolm <dmalcolm@redhat.com>
1476 * substring-locations.c (format_warning_va): Strengthen case 1 so
1477 that both endpoints of the substring must be within the format
1478 range for just the substring to be printed.
1480 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
1482 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
1483 * config/i386/i386.c (ix86_target_string): Add missing options
1484 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
1485 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
1486 flags_other and ix86_target_other to flags2_other. Display unknown
1488 (ix86_valid_target_attribute_inner_p): Add missing options and
1489 reorder options by implied ISAs, as in ix86_target_string.
1491 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1493 * hash-table.h (hash_table::too_empty_p): New function.
1494 (hash_table::expand): Use it.
1495 (hash_table::traverse): Likewise.
1496 (hash_table::empty_slot): Use sizeof (value_type) instead of
1497 sizeof (PTR) to convert bytes to elements. Shrink the table
1498 if the current size is excessive for the current number of
1501 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1503 * ira-costs.c (record_reg_classes): Break from the inner loop
1504 early once alt_fail is known to be true. Update outer loop
1505 handling accordingly.
1507 2017-01-13 Jeff Law <law@redhat.com>
1509 * tree-ssa-dse.c (decrement_count): New function.
1510 (increment_start_addr, maybe_trim_memstar_call): Likewise.
1511 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
1512 when we know the partially dead statement is a mem* function.
1514 PR tree-optimization/61912
1515 PR tree-optimization/77485
1516 * tree-ssa-dse.c: Include expr.h.
1517 (maybe_trim_constructor_store): New function.
1518 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
1520 PR tree-optimization/33562
1521 PR tree-optimization/61912
1522 PR tree-optimization/77485
1523 * doc/invoke.texi: Document new dse-max-object-size param.
1524 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
1525 * tree-ssa-dse.c: Include params.h.
1526 (dse_store_status): New enum.
1527 (initialize_ao_ref_for_dse): New, partially extracted from
1529 (valid_ao_ref_for_dse, normalize_ref): New.
1530 (setup_live_bytes_from_ref, compute_trims): Likewise.
1531 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
1532 (maybe_trim_partially_dead_store): Likewise.
1533 (maybe_trim_complex_store): Likewise.
1534 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
1535 Track what bytes live from the original store. Return tri-state
1536 for dead, partially dead or live.
1537 (dse_dom_walker): Add constructor, destructor and new private members.
1538 (delete_dead_call, delete_dead_assignment): New extracted from
1540 (dse_optimize_stmt): Make a member of dse_dom_walker.
1541 Use initialize_ao_ref_for_dse.
1543 PR tree-optimization/33562
1544 PR tree-optimization/61912
1545 PR tree-optimization/77485
1546 * sbitmap.h (bitmap_count_bits): Prototype.
1547 (bitmap_clear_range, bitmap_set_range): Likewise.
1548 * sbitmap.c (bitmap_clear_range): New function.
1549 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
1551 2017-01-13 Martin Liska <mliska@suse.cz>
1554 * function.c (set_cfun): Add new argument force.
1555 * function.h (set_cfun): Likewise.
1556 * ipa-inline-transform.c (inline_call): Use the function when
1557 strict alising from is dropped for function we inline to.
1559 2017-01-13 Richard Biener <rguenther@suse.de>
1561 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
1562 for dumping GIMPLE INTEGER_CSTs.
1564 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1566 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
1567 to 201112L since C++17.
1569 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
1572 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
1573 if -fsanitize=kernel-address is present.
1575 2017-01-13 Richard Biener <rguenther@suse.de>
1577 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
1578 as _Literal ( type ) number in case usual suffixes do not
1579 preserve all information.
1581 2017-01-13 Richard Biener <rguenther@suse.de>
1583 PR tree-optimization/77283
1584 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
1585 and ssa-iterators.h.
1586 (is_feasible_trace): Implement a cost model based on joiner
1589 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1592 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
1593 char or short to __float128/_Float128 directly.
1595 2017-01-12 Martin Sebor <msebor@redhat.com>
1597 to -Wformat-overflow.
1598 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
1599 (min_bytes_remaining): Same.
1600 (get_string_length): Same.
1601 (format_string): Same.
1602 (format_directive): Same.
1604 (pass_sprintf_length::handle_gimple_call): Same.
1606 2017-01-12 Jakub Jelinek <jakub@redhat.com>
1608 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
1609 info.nowrite calls with no lhs that can't throw. Return bool
1610 whether gsi_remove has been called or not.
1611 (pass_sprintf_length::handle_gimple_call): Return bool whether
1612 try_substitute_return_value called gsi_remove. Formatting fix.
1613 (pass_sprintf_length::execute): Don't use gsi_remove if
1614 handle_gimple_call returned true.
1617 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
1618 be removed due to side-effects, don't remove following barrier nor
1619 turn the successor edge into fallthru edge.
1621 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1624 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
1625 element-reversing loads and stores as not swappable.
1627 2017-01-12 Nathan Sidwell <nathan@acm.org>
1628 Nicolai Stange <nicstange@gmail.com>
1630 * combine.c (try_combine): Don't ignore result of overlap checking
1631 loop. Combine overlap & asm check into single loop.
1633 2017-01-12 Richard Biener <rguenther@suse.de>
1635 * tree-pretty-print.c (dump_generic_node): Provide -gimple
1636 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
1638 2017-01-12 Richard Biener <rguenther@suse.de>
1640 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
1641 and TS_TARGET_OPTION directly derive from TS_BASE.
1642 * tree-core.h (tree_optimization_option): Derive from tree_base.
1643 (tree_target_option): Likewise.
1645 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
1647 * config/i386/i386.c (memory_address_length): Increase len
1648 only when rip_relative_addr_p returns false.
1650 2017-01-11 Julia Koval <julia.koval@intel.com>
1652 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
1653 (OPTION_MASK_ISA_SGX_SET): New.
1654 (ix86_handle_option): Handle OPT_msgx.
1655 * config.gcc: Added sgxintrin.h.
1656 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
1657 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
1658 * config/i386/i386.c (ix86_target_string): Add -msgx.
1660 (ix86_option_override_internal): Handle new options.
1661 (ix86_valid_target_attribute_inner_p): Add sgx.
1662 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
1663 * config/i386/i386.opt: Add msgx.
1664 * config/i386/sgxintrin.h: New file.
1665 * config/i386/x86intrin.h: Add sgxintrin.h.
1667 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1670 * fold-const.c (maybe_nonzero_address): Return 1 for function
1672 (tree_single_nonzero_warnv_p): Don't handle function local objects
1676 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
1679 2017-01-11 David Malcolm <dmalcolm@redhat.com>
1682 * opts.c: Include "spellcheck.h"
1683 (struct string_fragment): New struct.
1684 (struct edit_distance_traits<const string_fragment &>): New
1686 (get_closest_sanitizer_option): New function.
1687 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
1689 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1691 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
1693 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
1694 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
1695 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
1696 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
1697 for initial die_offset if dwarf_split_debug_info.
1698 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
1699 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
1701 (output_skeleton_debug_sections): Formatting fix. Use
1702 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
1703 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
1705 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
1707 * config/arm/cortex-a53.md: Add bypasses for
1709 (cortex_a53_r2f): Only use for transfers.
1710 (cortex_a53_f2r): Likewise.
1711 (cortex_a53_r2f_cvt): Add reservation for conversions.
1712 (cortex_a53_f2r_cvt): Likewise.
1714 2017-01-11 Tamar Christina <tamar.christina@arm.com>
1716 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
1717 to all inlined functions, change static to extern.
1719 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
1722 * config/arm/arm.c (legitimize_pic_address): Handle reference to
1724 (arm_assemble_integer): Likewise.
1726 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
1728 * config.gcc: Use new awk script to check CPU, FPU and architecture
1729 parameters for --with-... options.
1730 * config/arm/parsecpu.awk: New file
1731 * config/arm/arm-cpus.in: New file.
1732 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
1734 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
1736 * config/arm/t-arm: Update dependency rules.
1737 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
1738 of processing .def files.
1739 * config/arm/genopt.sh: Deleted.
1740 * config/arm/gentune.sh: Deleted.
1741 * config/arm/arm-cores.def: Deleted.
1742 * config/arm/arm-arches.def: Deleted.
1743 * config/arm/arm-fpus.def: Deleted.
1744 * config/arm/arm-tune.md: Regenerated.
1745 * config/arm/arm-tables.opt: Regenerated.
1746 * config/arm/arm-cpu.h: New generated file.
1747 * config/arm/arm-cpu-data.h: New generated file.
1748 * config/arm/arm-cpu-cdata.h: New generated file.
1750 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
1753 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
1755 (input_varpool_node): Unpack dynamically_initialized bit.
1757 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
1759 PR rtl-optimization/79032
1760 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
1761 the alignment of the adjusted memory reference against that of MODE,
1762 instead of the alignment of the original memory reference.
1764 2017-01-11 Martin Jambor <mjambor@suse.cz>
1766 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
1768 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
1769 decorated functions.
1771 2017-01-11 Richard Biener <rguenther@suse.de>
1773 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
1774 set range/nonnull info for PHI results. Do not set it on
1775 stmts marked for removal.
1777 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
1779 * expr.c (store_field): In the bitfield case, fetch the return value
1780 from the registers before applying a single big-endian adjustment.
1781 Always do a final load for a BLKmode value not larger than a word.
1783 2017-01-10 David Malcolm <dmalcolm@redhat.com>
1786 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
1787 that we correctly handle column numbers greater than
1788 LINE_MAP_MAX_COLUMN_NUMBER.
1790 2017-01-10 Martin Sebor <msebor@redhat.com>
1793 * gimple-ssa-sprintf.c (get_destination_size): Call
1794 {init,fini}object_sizes.
1795 * tree-object-size.c (addr_object_size): Adjust.
1796 (pass_through_call): Adjust.
1797 (pass_object_sizes::execute): Adjust.
1798 * tree-object-size.h (fini_object_sizes): Declare.
1800 2017-01-10 Martin Sebor <msebor@redhat.com>
1802 PR tree-optimization/78775
1803 * builtins.c (get_size_range): Move...
1804 * calls.c: ...to here.
1805 (alloc_max_size): Accept zero argument.
1806 (operand_signed_p): Remove.
1807 (maybe_warn_alloc_args_overflow): Call get_size_range.
1808 * calls.h (get_size_range): Declare.
1810 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
1812 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
1813 from TI's devices.csv file as of September 2016.
1814 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
1816 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
1818 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
1819 * doc/invoke.texi: Likewise.
1820 * doc/md.texi: Likewise.
1821 * doc/objc.texi: Likewise.
1823 2017-01-10 Joshua Conner <joshconner@google.com>
1825 * config/arm/fuchsia-elf.h: New file.
1826 * config/fuchsia.h: New file.
1827 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
1828 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
1830 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
1832 2016-01-10 Richard Biener <rguenther@suse.de>
1834 PR tree-optimization/79034
1835 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
1836 Propagate out degenerate PHIs in the joiner.
1838 2017-01-10 Martin Liska <mliska@suse.cz>
1840 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
1841 (sort_congruence_classes_by_decl_uid): Likewise.
1842 (sort_congruence_class_groups_by_decl_uid): Likewise.
1843 (sem_item_optimizer::merge_classes): Sort class, groups in these
1844 classes and members in the groups by DECL_UID of declarations.
1845 This would make merge operations stable.
1847 2017-01-10 Martin Liska <mliska@suse.cz>
1849 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
1850 usage of m_classes_vec.
1851 (sem_item_optimizer::~sem_item_optimizer): Likewise.
1852 (sem_item_optimizer::get_group_by_hash): Likewise.
1853 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
1854 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
1855 (sem_item_optimizer::verify_classes): Likewise.
1856 (sem_item_optimizer::process_cong_reduction): Likewise.
1857 (sem_item_optimizer::dump_cong_classes): Likewise.
1858 (sem_item_optimizer::merge_classes): Likewise.
1859 * ipa-icf.h (congruence_class_hash): Rename from
1860 congruence_class_group_hash. Remove declaration of m_classes_vec.
1862 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
1864 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
1865 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
1866 * config.gcc: Add avx512vpopcntdqintrin.h.
1867 * config/i386/avx512vpopcntdqintrin.h: New.
1868 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
1869 * config/i386/i386-builtin-types.def: Add new types.
1870 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
1871 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
1872 __builtin_ia32_vpopcountq_v8di_mask): New.
1873 * config/i386/i386-c.c (ix86_target_macros_internal): Define
1874 __AVX512VPOPCNTDQ__.
1875 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
1876 (PTA_AVX512VPOPCNTDQ): Define.
1877 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
1878 TARGET_AVX512VPOPCNTDQ_P): Define.
1879 * config/i386/i386.opt: Add mavx512vpopcntdq.
1880 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
1881 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
1883 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
1886 * predict.def (PRED_CALL): Set to 67.
1888 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
1890 * expr.c (store_field): In the bitfield case, if the value comes from
1891 a function call and is of an aggregate type returned in registers, do
1892 not modify the field mode; extract the value in all cases if the mode
1893 is BLKmode and the size is not larger than a word.
1895 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
1898 * config/i386/cpuid.h: Fix undefined behavior.
1900 2017-01-04 Jeff Law <law@redhat.com>
1902 PR tree-optimization/79007
1903 PR tree-optimization/67955
1904 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
1905 conservative for pt.null when flag_non_call_exceptions is on.
1907 2017-01-09 Jakub Jelinek <jakub@redhat.com>
1909 PR translation/79019
1910 PR translation/79020
1911 * params.def (PARAM_INLINE_MIN_SPEEDUP,
1912 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
1913 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
1915 * config/avr/avr.opt (maccumulate-args): Likewise.
1916 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
1917 * common.opt (freport-bug): Likewise.
1918 * cif-code.def (CIF_FINAL_ERROR): Likewise.
1919 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
1920 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
1921 translatable string.
1922 * config/i386/i386.c (function_value_32): Likewise.
1923 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
1924 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
1926 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
1927 * common/config/msp430/msp430-common.c (msp430_handle_option):
1929 * symtab.c (symtab_node::verify_base): Likewise.
1930 * opts.c (set_debug_level): Likewise.
1931 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
1932 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
1933 missing whitespace to translatable strings.
1934 * config/avr/avr.md (bswapsi2): Fix typo in comment.
1935 * config/sh/superh.h: Likewise.
1936 * config/i386/xopintrin.h: Likewise.
1937 * config/i386/znver1.md: Likewise.
1938 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
1939 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
1940 * double-int.h (struct double_int): Likewise.
1941 * double-int.c (div_and_round_double): Likewise.
1942 * wide-int.cc: Likewise.
1943 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1944 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
1945 * cfgcleanup.c (crossjumps_occured): Renamed to ...
1946 (crossjumps_occurred): ... this.
1947 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
1950 PR tree-optimization/78899
1951 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
1952 returning bool return struct loop *, NULL for failure and the new
1954 (versionable_outer_loop_p): Don't version outer loop if it has
1955 dont_vectorized bit set.
1956 (tree_if_conversion): When versioning outer loop, ensure
1957 tree_if_conversion is performed also on the inner loop of the
1958 non-vectorizable outer loop copy.
1959 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
1960 LOOP_VECTORIZED in inner loop of the scalar outer loop and
1961 prevent vectorization of it.
1962 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
1963 the outer loop vectorization of the non-scalar version is attempted
1964 before vectorization of the inner loop in scalar version. If
1965 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
1966 vectorization of its inner loop.
1967 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
1968 has 2 inner loops, rename also on edges from bb whose single pred
1969 is outer_loop->header. Fix typo in function comment.
1971 2017-01-09 Martin Sebor <msebor@redhat.com>
1974 * asan.c (asan_emit_stack_protection): Increase local buffer size
1975 to avoid snprintf truncation warning.
1977 2017-01-09 Andrew Pinski <apinski@cavium.com>
1979 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
1980 to reference thunderx2t99 for the tuning structure
1981 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
1983 (thunderx2t99_extra_costs): This.
1984 * config/aarch64/aarch64-tune.md: Regenerate.
1985 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
1986 (vulcan_addrcost_table): This.
1987 (vulcan_regmove_cost): Rename to ...
1988 (thunderx2t99_regmove_cost): This.
1989 (vulcan_vector_cost): Rename to ...
1990 (thunderx2t99_vector_cost): this.
1991 (vulcan_branch_cost): Rename to ...
1992 (thunderx2t99_branch_cost): This.
1993 (vulcan_tunings): Rename to ...
1994 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
1995 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
1997 2017-01-09 Martin Jambor <mjambor@suse.cz>
2001 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
2002 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
2003 (propagate_vr_accross_jump_function): Use the above function for all
2004 value range computations for pass-through jump functions and type
2005 converasion from explicit value range values.
2006 (ipcp_propagate_stage): Do not attempt to deduce types of formal
2007 parameters from TYPE_ARG_TYPES.
2008 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
2009 (ipa_write_node_info): Stream type of the actual argument.
2010 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
2012 2017-01-09 Martin Liska <mliska@suse.cz>
2015 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
2016 (lookup_compiler): Do not show error message with have_E.
2018 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2020 PR tree-optimization/78938
2021 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
2022 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
2023 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
2024 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
2027 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2029 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
2032 2017-01-09 Richard Biener <rguenther@suse.de>
2034 PR tree-optimization/78997
2035 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
2036 name condition properly.
2038 2017-01-09 Richard Biener <rguenther@suse.de>
2041 * dwarf2out.c (is_cxx): New overload with context.
2042 (is_naming_typedef_decl): Use it.
2044 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2046 * invoke.texi (Option Summary): Correct spacing in option lists
2047 and add line breaks to fix over-long lines.
2049 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2053 * extend.texi (Common Variable Attributes): Add xref to GCC
2054 Internals manual to explain mode attribute keywords.
2056 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2059 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
2060 and Preprocessor Options.
2061 (Options for Linking): Document -pthread here....
2062 (RS/6000 and PowerPC Options): ...not here.
2063 (Solaris 2 Options): ...or here.
2064 * doc/cppopts.texi: Document -pthread.
2066 2017-01-08 Martin Sebor <msebor@redhat.com>
2069 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
2070 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
2071 New member functions.
2072 (format_directive): Used them.
2074 (pass_sprintf_length::handle_gimple_call): Same.
2075 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
2076 to avoid truncation for any argument.
2077 (extract_affine_mul): Same.
2078 * tree.c (get_file_function_name): Same.
2080 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2083 * predict.def (PRED_INDIR_CALL): Set to 86.
2085 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2087 PR preprocessor/54124
2088 * doc/cppopts.texi: Reformat -d subtable to list the full name
2089 of the options. Add cross-reference to the docs for the general
2090 compiler -d options.
2091 * doc/invoke.texi (Developer Options): Add cross-reference to the
2092 preprocessor-specific -d option documentation.
2094 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2096 PR preprocessor/13498
2097 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2098 redudant material, and reflect new command-line options.
2099 (System Headers): Likewise.
2101 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2103 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2104 -isystem, and -idirafter. Copy-edit.
2105 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2106 default for -ftrack-macro-expansion. Delete obsolete and
2107 badly-formatted implementation details about -fdebug-cpp output.
2108 * doc/cppwarnopts.texi: Copy-edit.
2110 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2113 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2114 that the transition from a max line width >= 1<<10 to narrower
2115 lines works correctly.
2117 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2119 * doc/options.texi (PerFunction): New.
2120 * opt-functions.awk (switch_flags): Map both Optimization and
2121 PerFunction to CL_OPTIMIZATION.
2122 * opth-gen.awk: Test for PerFunction flag along with
2124 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2125 it only when the latter is present. Skip those that don't in
2126 the hash function generator.
2127 * common.opt (fvar-tracking): Mark as PerFunction instead of
2129 (fvar-tracking-assignments): Likewise.
2130 (fvar-tracking-assignments-toggle): Likewise.
2131 (fvar-tracking-uninit): Likewise.
2133 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2135 PR translation/79018
2136 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2139 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2142 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2143 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2144 TARGET_LONG_JUMP_TABLE_OFFSETS.
2145 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2146 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2147 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2148 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2149 * config/m68k/m68k.md (tablejump expander): Likewise.
2150 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2151 TARGET_LONG_JUMP_TABLE_OFFSETS.
2152 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2153 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2155 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2156 David Holsgrove <david.holsgrove@xilinx.com>
2158 * common/config/microblaze/microblaze-common.c
2159 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2160 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2162 * config/microblaze/microblaze.c (microblaze_must_save_register)
2163 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2165 (microblaze_eh_return): New function.
2166 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2167 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2168 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2169 * config/microblaze/microblaze.md (eh_return): New pattern.
2171 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2173 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2174 GCC_DIAGNOSTIC_STRINGIFY): Define.
2176 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2178 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2180 * config/arm/arm.md (<mcrr>): New.
2182 * config/arm/arm.c (arm_arch5te): New.
2183 (arm_option_override): Set arm_arch5te.
2184 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2186 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2187 (arm_mcrr_qualifiers): ... this. New.
2188 (MRRC_QUALIFIERS): Define to...
2189 (arm_mrrc_qualifiers): ... this. New.
2190 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2192 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2193 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2194 (MRRCI, mrrc, MRRC): New.
2195 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2196 VUNSPEC_MRRC2): New.
2198 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2200 * config/arm/arm.md (<mcr>): New.
2202 * config/arm/arm.c (arm_coproc_builtin_available): Add
2203 support for mcr, mrc, mcr2 and mrc2.
2204 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2205 (arm_mcr_qualifiers): ... this. New.
2206 (MRC_QUALIFIERS): Define to ...
2207 (arm_mrc_qualifiers): ... this. New.
2208 (MCR_QUALIFIERS): Define to ...
2209 (arm_mcr_qualifiers): ... this. New.
2210 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2212 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2213 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2214 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2217 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2219 * config/arm/arm.md (*ldc): New.
2223 * config/arm/arm.c (arm_coproc_builtin_available): Add
2224 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2225 (arm_coproc_ldc_stc_legitimate_address): New.
2226 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2227 'qualifier_const_pointer'.
2228 (LDC_QUALIFIERS): Define to...
2229 (arm_ldc_qualifiers): ... this. New.
2230 (STC_QUALIFIERS): Define to...
2231 (arm_stc_qualifiers): ... this. New.
2232 * config/arm/arm-protos.h
2233 (arm_coproc_ldc_stc_legitimate_address): New.
2234 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2235 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2236 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2237 stc2, stcl, stc2l): New.
2238 * config/arm/constraints.md (Uz): New.
2239 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2240 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2241 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2242 VUNSPEC_STC2L): New.
2244 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2246 * config/arm/arm.md (<cdp>): New.
2247 * config/arm/arm.c (neon_const_bounds): Rename this ...
2248 (arm_const_bounds): ... this.
2249 (arm_coproc_builtin_available): New.
2250 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2251 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2252 (CDP_QUALIFIERS): Define to...
2253 (arm_cdp_qualifiers): ... this. New.
2255 (arm_expand_builtin_args): Add case for 6 arguments.
2256 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2257 (arm_const_bounds): ... this.
2258 (arm_coproc_builtin_available): New.
2259 * config/arm/arm_acle.h (__arm_cdp): New.
2261 * config/arm/arm_acle_builtins.def (cdp): New.
2263 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2264 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2266 * config/arm/types.md (coproc): New.
2267 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2268 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2269 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2270 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2272 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2274 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2275 (UBINOP_QUALIFIERS): New.
2277 (acle_builtin_data): New. Change comment.
2278 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2279 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2280 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2281 arm_acle_builtins.def.
2282 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2283 (arm_init_acle_builtins): New.
2284 (CRC32_BUILTIN): Remove.
2285 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2286 crc32cb, crc32ch and crc32cw.
2287 (arm_init_crc32_builtins): Remove.
2288 (arm_init_builtins): Use arm_init_acle_builtins rather
2289 than arm_init_crc32_builtins.
2290 (arm_expand_acle_builtin): New.
2291 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2292 * config/arm/arm_acle_builtins.def: New.
2294 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2296 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2297 (arm_builtin_datum): ... this.
2298 (arm_init_neon_builtin): Rename to ...
2299 (arm_init_builtin): ... this. Add a new parameters PREFIX
2300 and USE_SIG_IN_NAME.
2301 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2302 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2303 'arm_builtin_datum'.
2304 (arm_init_vfp_builtins): Likewise.
2305 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2306 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2307 (arm_expand_neon_args): Rename to ...
2308 (arm_expand_builtin_args): ... this. Rename builtin_arg
2309 enum values and differentiate between ARG_BUILTIN_MEMORY
2310 and ARG_BUILTIN_NEON_MEMORY.
2311 (arm_expand_neon_builtin_1): Rename to ...
2312 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2313 values, arm_expand_builtin_args and add bool parameter NEON.
2314 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2315 (arm_expand_vfp_builtin): Likewise.
2316 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2318 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2321 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2322 * predict.c (tree_estimate_probability_bb): Reverse direction of
2323 polymorphic call predictor.
2325 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2327 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2328 (determine_pass_name_match): ...this new function and...
2329 (should_skip_pass_p): ...this new function.
2331 2017-01-06 Nathan Sidwell <nathan@acm.org>
2333 * ipa-visibility.c (function_and_variable_visibility): Reformat
2334 comments and long lines. Remove extrneous if.
2335 * symtab.c (symtab_node::make_decl_local): Fix code format.
2336 (symtab_node::set_section_for_node): Fix comment typo.
2338 2017-01-06 Martin Liska <mliska@suse.cz>
2341 * lra-constraints.c: Rename invariant to lra_invariant.
2342 * predict.c (set_even_probabilities): Initialize e to NULL.
2344 2017-01-05 Martin Sebor <msebor@redhat.com>
2346 PR tree-optimization/78910
2347 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2348 (format_integer): Correct off-by-one error in the handling
2349 of precision with negative numbers in signed conversions..
2351 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
2353 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
2355 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2357 PR tree-optimization/71016
2358 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
2359 factor_out_conditional_conversion. Formatting fix.
2360 (factor_out_conditional_conversion): Add cond_stmt argument.
2361 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
2362 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
2365 2017-01-05 David Malcolm <dmalcolm@redhat.com>
2367 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
2368 read-rtl-function.o, and selftest-rtl.o.
2369 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
2370 (selftest::aarch64_test_loading_full_dump): New function.
2371 (selftest::aarch64_run_selftests): New function.
2372 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
2373 selftest::aarch64_run_selftests.
2374 * config/i386/i386.c
2375 (selftest::ix86_test_loading_dump_fragment_1): New function.
2376 (selftest::ix86_test_loading_call_insn): New function.
2377 (selftest::ix86_test_loading_full_dump): New function.
2378 (selftest::ix86_test_loading_unspec): New function.
2379 (selftest::ix86_run_selftests): Call the new functions.
2380 * emit-rtl.c (maybe_set_max_label_num): New function.
2381 * emit-rtl.h (maybe_set_max_label_num): New decl.
2382 * function.c (instantiate_decls): Guard call to
2383 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
2384 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
2386 * gensupport.c (gen_reader::gen_reader): Pass "false"
2387 for new "compact" param of rtx_reader.
2388 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
2389 rather than an empty string for NULL strings.
2390 * read-md.c: Potentially include config.h rather than bconfig.h.
2391 Wrap include of errors.h with #ifdef GENERATOR_FILE.
2392 (have_error): New global, copied from errors.c.
2393 (md_reader::read_name): Rename to...
2394 (md_reader::read_name_1): ...this, adding "out_loc" param,
2395 and converting "missing name or number" to returning false, rather
2397 (md_reader::read_name): Reimplement in terms of read_name_1.
2398 (md_reader::read_name_or_nil): New function.
2399 (md_reader::read_string): Handle "(nil)" by returning NULL.
2400 (md_reader::md_reader): Add new param "compact".
2401 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
2402 (md_reader::read_file): New method.
2403 * read-md.h (md_reader::md_reader): Add new param "compact".
2404 (md_reader::read_file): New method.
2405 (md_reader::is_compact): New accessor.
2406 (md_reader::read_name): Convert return type from void to file_location.
2407 (md_reader::read_name_or_nil): New decl.
2408 (md_reader::read_name_1): New decl.
2409 (md_reader::m_compact): New field.
2410 (noop_reader::noop_reader): Pass "false" for new "compact" param
2412 (rtx_reader::rtx_reader): Add new "compact" param.
2413 (rtx_reader::read_rtx_operand): Make virtual and convert return
2414 type from void to rtx.
2415 (rtx_reader::read_until): New decl.
2416 (rtx_reader::handle_any_trailing_information): New virtual function.
2417 (rtx_reader::postprocess): New virtual function.
2418 (rtx_reader::finalize_string): New virtual function.
2419 (rtx_reader::m_in_call_function_usage): New field.
2420 (rtx_reader::m_reuse_rtx_by_id): New field.
2421 * read-rtl-function.c: New file.
2422 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
2423 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
2424 (selftest::verify_three_block_rtl_cfg): New decl.
2425 * read-rtl-function.h: New file.
2426 * read-rtl.c: Potentially include config.h rather than bconfig.h.
2427 For host, include function.h, memmodel.h, and emit-rtl.h.
2428 (one_time_initialization): New function.
2429 (struct compact_insn_name): New struct.
2430 (compact_insn_names): New array.
2431 (find_code): Handle insn codes in compact dumps.
2432 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
2433 (bind_subst_iter_and_attr): Likewise.
2434 (add_condition_to_string): Likewise.
2435 (add_condition_to_rtx): Likewise.
2436 (apply_attribute_uses): Likewise.
2437 (add_current_iterators): Likewise.
2438 (apply_iterators): Likewise.
2439 (initialize_iterators): Guard usage of apply_subst_iterator with
2440 #ifdef GENERATOR_FILE.
2441 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
2442 (md_reader::read_mapping): Likewise.
2443 (add_define_attr_for_define_subst): Likewise.
2444 (add_define_subst_attr): Likewise.
2445 (read_subst_mapping): Likewise.
2446 (check_code_iterator): Likewise.
2447 (rtx_reader::read_rtx): Likewise. Move one-time initialization
2449 (one_time_initialization): New function.
2450 (rtx_reader::read_until): New method.
2451 (read_flags): New function.
2452 (parse_reg_note_name): New function.
2453 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
2454 Handle reuse_rtx ids.
2455 Wrap iterator lookup within #ifdef GENERATOR_FILE.
2456 Add parsing support for RTL dumps, mirroring the special-cases in
2457 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
2458 values, and calling handle_any_trailing_information.
2459 (rtx_reader::read_rtx_operand): Convert return type from void
2460 to rtx, returning return_rtx. Handle case 'e'. Call
2461 finalize_string on XSTR and XTMPL fields.
2462 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
2463 "(nil)" values were omitted. Call the postprocess vfunc on the
2465 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
2466 class ctor. Initialize m_in_call_function_usage. Call
2467 one_time_initialization.
2468 * rtl-tests.c (selftest::test_uncond_jump): Call
2469 set_new_first_and_last_insn.
2470 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
2471 * selftest-rtl.c: New file.
2472 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
2473 (selftest::get_insn_by_uid): New decl.
2474 * selftest-run-tests.c (selftest::run_tests): Call
2475 read_rtl_function_c_tests.
2476 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
2477 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
2480 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
2482 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
2483 operands in a special way. Assert that pos+len <= mode precision.
2485 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2487 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
2488 3 argument Alias with unlimited for the negative form.
2489 (fno-vect-cost-model): Removed.
2491 2017-01-05 Martin Liska <mliska@suse.cz>
2493 * hsa-gen.c (gen_hsa_divmod): New function.
2494 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
2496 2017-01-05 Martin Liska <mliska@suse.cz>
2499 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
2502 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2504 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
2505 small constant length operands.
2507 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2509 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
2510 between loop iterations.
2512 2017-01-05 Martin Liska <mliska@suse.cz>
2515 * gimplify.c (gimplify_decl_expr): Compare to
2516 asan_poisoned_variables instread of checking flags.
2517 (gimplify_target_expr): Likewise.
2518 (gimplify_expr): Likewise.
2519 (gimplify_function_tree): Conditionally initialize
2520 asan_poisoned_variables.
2522 2017-01-04 Jeff Law <law@redhat.com>
2524 PR tree-optimizatin/78812
2525 * rtl.h (contains_mem_rtx_p): Prototype.
2526 * ifcvt.c (containts_mem_rtx_p): Move from here to...
2527 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
2528 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
2529 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
2530 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
2532 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2534 * input.c (assert_char_at_range): Default-initialize actual_range.
2536 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2538 * df-scan.c (df_ref_create_structure): Make regno unsigned,
2539 to match the caller.
2541 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2543 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
2544 insns after final jump in test to emit dummy move.
2546 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2548 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
2549 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
2551 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2553 * multiple_target.c (create_dispatcher_calls): Init e_next.
2554 * tree-ssa-loop-split.c (split_loop): Init border.
2555 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
2558 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
2563 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
2564 (altivec_register_operand): Do not return true if the operand
2565 contains a SUBREG mixing SImode and SFmode.
2566 (vsx_register_operand): Likewise.
2567 (vsx_reg_sfsubreg_ok): New predicate.
2568 (vfloat_operand): Do not return true if the operand contains a
2569 SUBREG mixing SImode and SFmode.
2570 (vint_operand): Likewise.
2571 (vlogical_operand): Likewise.
2572 (gpc_reg_operand): Likewise.
2573 (int_reg_operand): Likewise.
2574 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
2575 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
2576 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
2578 (rs6000_emit_move_si_sf_subreg): New helper function.
2579 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
2580 fixup SUBREGs involving SImode and SFmode.
2581 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
2582 numbers for the new peephole2 optimization.
2583 (peephole2 for SFmode unions): New peephole2 to optimize cases in
2584 the GLIBC math library that do AND/IOR/XOR operations on single
2585 precision floating point.
2586 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
2587 target macros to say whether we need to avoid SUBREGs mixing
2589 (TARGET_ALLOW_SF_SUBREG): Likewise.
2590 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
2591 (UNSPEC_SI_FROM_SF): Likewise.
2592 (iorxor): Change spacing.
2593 (and_ior_xor): New iterator for AND, IOR, and XOR.
2594 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
2595 (movdi_from_sf_zero_ext): Likewise.
2596 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
2597 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
2598 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
2599 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
2600 (fms<mode>4): Likewise.
2601 (fnma<mode>4): Likewise.
2602 (fnms<mode>4): Likewise.
2603 (nfma<mode>4): Likewise.
2604 (nfms<mode>4): Likewise.
2606 2017-01-04 Marek Polacek <polacek@redhat.com>
2609 * doc/invoke.texi: Document -Wpointer-compare.
2611 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2613 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
2616 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
2617 descriptions for -gdwarf-5 and emit them as uleb128 instead of
2620 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
2623 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
2624 documentation of the powerpc_popcntb_ok attribute.
2625 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
2626 code to issue warning messages if a requested CPU configuration is
2627 not supported by the binary (assembler and loader) toolchain.
2628 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
2629 made to define a built-in function that has been disabled.
2630 (paired_init_builtins): Add assertion to prevent ICE if attempt is
2631 made to define a built-in function that has been disabled.
2632 (altivec_init_builtins): Add comment explaining why definition
2633 of the DST built-in functions is not preceded by an assertion
2634 check. Add assertions to prevent ICE if attempts are made to
2635 define an altivec predicate or an abs* built-in function that has
2637 (htm_init_builtins): Add comment explaining why definition of the
2638 htm built-in functions is not preceded by an assertion check.
2640 2017-01-04 Jeff Law <law@redhat.com>
2642 PR tree-optimizatin/67955
2643 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
2644 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
2645 the points-to solution does not include pt_null. Use DECL_PT_UID
2648 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
2650 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
2651 Use gen_int_mode instead of gen_lopwart for const_int operands.
2653 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2655 PR tree-optimization/71563
2656 * match.pd: Simplify X << Y into X if Y is known to be 0 or
2657 out of range value - has low bits known to be zero.
2659 2017-01-04 Alan Modra <amodra@gmail.com>
2661 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
2662 * configure: Regenerate.
2663 * config.in: Regenerate.
2665 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2668 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
2669 a substring of the message, but strcmp with the whole message. Ifdef
2670 ENABLE_NLS, translate the message first using dgettext.
2672 2017-01-03 Jeff Law <law@redhat.com>
2674 PR tree-optimizatin/78856
2675 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
2676 (mark_threaded_blocks): Remove code to truncate thread paths that
2677 cross multiple loop headers. Instead invalidate the cached loop
2678 iteration information and handle case of a thread path walking
2679 into an irreducible region.
2681 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2684 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
2685 assertions. Add support for doing the signbit if the IEEE 128-bit
2686 floating point value is in a GPR.
2687 * config/rs6000/rs6000.md (Fsignbit): Delete.
2688 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
2689 Update the length attribute if the value is in a GPR.
2690 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
2691 the sign or zero extension instruction, since the value is always 0/1.
2692 (signbit<mode>2_dm2): Delete using <Fsignbit>.
2695 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
2696 extracting SImode to a GPR register so that we can generate a
2697 store, limit the vector to be in a traditional Altivec register
2698 for the vextuwrx instruction.
2700 2017-01-03 Ian Lance Taylor <iant@google.com>
2702 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
2704 2017-01-03 Martin Sebor <msebor@redhat.com>
2706 PR tree-optimization/78696
2707 * gimple-ssa-sprintf.c (format_floating): Correct handling of
2708 precision. Use MPFR for %f for greater fidelity. Correct handling
2710 (pass_sprintf_length::compute_format_length): Set width and precision
2711 specified by asrerisk to void_node for vararg functions.
2712 (try_substitute_return_value): Adjust dump output.
2714 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
2716 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
2718 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
2720 * doc/invoke.texi (SPARC options): Document -mlra as the default.
2721 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
2722 -mlra/-mno-lra was passed to the compiler.
2724 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
2726 PR rtl-optimization/65618
2727 * emit-rtl.c (try_split): Move initialization of "before" and
2728 "after" to just before the call to emit_insn_after_setloc.
2730 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
2732 * doc/md.texi (Standard Names): Remove reference to Java frontend.
2734 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
2736 * dwarf2out.c (gen_enumeration_type_die): When
2737 -gno-strict-dwarf, add a DW_AT_encoding attribute.
2739 2017-01-03 Jakub Jelinek <jakub@redhat.com>
2741 PR tree-optimization/78965
2742 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
2743 Change first argument from const call_info & to call_info &. For %n
2744 set info.nowrite to false.
2747 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
2748 possibly throwing calls.
2750 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
2751 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
2752 and fns handling, rather than in a separate case SSA_NAME.
2754 2017-01-02 Jeff Law <law@redhat.com>
2756 * config/darwin-driver.c (darwin_driver_init): Const-correctness
2757 fixes for first_period and second_period variables.
2759 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
2762 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
2763 (*insvqi_1): New insn pattern.
2764 (*insvqi_1_mem_rex64): Ditto.
2766 (*insvqi_3): Rename from *insvqi.
2768 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
2770 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
2772 * doc/cfg.texi (Edges): Remove reference to Java.
2773 (Maintaining the CFG): Ditto.
2775 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2778 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
2779 transparent aliases.
2781 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2784 * predict.def (PRED_CALL): Update hitrate.
2785 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
2786 * predict.c (tree_estimate_probability_bb): Split CALL predictor
2787 into direct/indirect/polymorphic variants.
2789 2017-01-01 Jakub Jelinek <jakub@redhat.com>
2791 Update copyright years.
2793 * gcc.c (process_command): Update copyright notice dates.
2794 * gcov-dump.c (print_version): Ditto.
2795 * gcov.c (print_version): Ditto.
2796 * gcov-tool.c (print_version): Ditto.
2797 * gengtype.c (create_file): Ditto.
2798 * doc/cpp.texi: Bump @copying's copyright year.
2799 * doc/cppinternals.texi: Ditto.
2800 * doc/gcc.texi: Ditto.
2801 * doc/gccint.texi: Ditto.
2802 * doc/gcov.texi: Ditto.
2803 * doc/install.texi: Ditto.
2804 * doc/invoke.texi: Ditto.
2806 Copyright (C) 2017 Free Software Foundation, Inc.
2808 Copying and distribution of this file, with or without modification,
2809 are permitted in any medium without royalty provided the copyright
2810 notice and this notice are preserved.