1 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3 * doc/invoke.texi (-fstore-merging): Correct default optimization
4 levels at which it is enabled.
5 (-O): Move -fstore-merging from list to...
8 2017-01-25 Richard Biener <rguenther@suse.de>
11 * omp-expand.c: Include debug.h.
12 (expand_omp_taskreg): Make sure to generate early debug before
13 outlining anything from a function.
14 (expand_omp_target): Likewise.
15 (grid_expand_target_grid_body): Likewise.
17 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
20 * asan.c (get_translation_unit_decl): New function.
21 (asan_add_global): Extract modules file name from globals
22 TRANSLATION_UNIT_DECL name.
24 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
27 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
28 for long calls with APCS frame and VFP.
30 2017-01-24 David Malcolm <dmalcolm@redhat.com>
32 * cfg.c (original_copy_tables_initialized_p): New function.
33 * cfg.h (original_copy_tables_initialized_p): New decl.
34 * cfgrtl.c (relink_block_chain): Guard the call to
35 free_original_copy_tables with a call to
36 original_copy_tables_initialized_p.
37 * cgraph.h (symtab_node::native_rtl_p): New decl.
38 * cgraphunit.c (symtab_node::native_rtl_p): New function.
39 (symtab_node::needed_p): Don't assert for early assembly output
41 (cgraph_node::finalize_function): Set "force_output" for __RTL
43 (cgraph_node::analyze): Bail out early for __RTL functions.
44 (analyze_functions): Update assertion to support __RTL functions.
45 (cgraph_node::expand): Bail out early for __RTL functions.
46 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
48 * function.h (struct function): Update comment for field
50 * gimple-expr.c: Include "tree-pass.h".
51 (gimple_has_body_p): Return false for __RTL functions.
52 * Makefile.in (OBJS): Add run-rtl-passes.o.
53 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
55 (gcc::pass_manager::get_clean_slate): New accessor.
56 * passes.c: Include "insn-addr.h".
57 (should_skip_pass_p): Add logging. Update logic for running
58 "expand" to be compatible with both __GIMPLE and __RTL. Guard
59 property-provider override so it is only done for gimple passes.
61 (skip_pass): New function.
62 (execute_one_pass): Call skip_pass when skipping passes.
63 * read-md.c (md_reader::read_char): Support filtering
64 the input to a subset of line numbers.
65 (md_reader::md_reader): Initialize fields
66 m_first_line and m_last_line.
67 (md_reader::read_file_fragment): New function.
68 * read-md.h (md_reader::read_file_fragment): New decl.
69 (md_reader::m_first_line): New field.
70 (md_reader::m_last_line): New field.
71 * read-rtl-function.c (function_reader::create_function): Only
72 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
73 curr_properties. Set DECL_INITIAL to a dummy block.
74 (read_rtl_function_body_from_file_range): New function.
75 * read-rtl-function.h (read_rtl_function_body_from_file_range):
77 * run-rtl-passes.c: New file.
78 * run-rtl-passes.h: New file.
80 2017-01-24 Jeff Law <law@redhat.com>
82 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
85 2017-01-24 Bin Cheng <bin.cheng@arm.com>
87 PR tree-optimization/79159
88 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
89 (record_nonwrapping_iv): Improve boundary using above function if no
90 value range information.
92 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
93 Martin Jambor <mjambor@suse.cz>
95 * brig-builtins.def: New file.
96 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
97 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
98 (DEF_HSAIL_SAT_BUILTIN): Likewise.
99 (DEF_HSAIL_INTR_BUILTIN): Likewise.
100 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
101 * builtin-types.def (BT_INT8): New.
102 (BT_INT16): Likewise.
103 (BT_UINT8): Likewise.
104 (BT_UINT16): Likewise.
105 (BT_FN_ULONG): Likewise.
106 (BT_FN_UINT_INT): Likewise.
107 (BT_FN_UINT_ULONG): Likewise.
108 (BT_FN_UINT_LONG): Likewise.
109 (BT_FN_UINT_PTR): Likewise.
110 (BT_FN_ULONG_PTR): Likewise.
111 (BT_FN_INT8_FLOAT): Likewise.
112 (BT_FN_INT16_FLOAT): Likewise.
113 (BT_FN_UINT32_FLOAT): Likewise.
114 (BT_FN_UINT16_FLOAT): Likewise.
115 (BT_FN_UINT8_FLOAT): Likewise.
116 (BT_FN_UINT64_FLOAT): Likewise.
117 (BT_FN_UINT16_UINT32): Likewise.
118 (BT_FN_UINT32_UINT16): Likewise.
119 (BT_FN_UINT16_UINT16_UINT16): Likewise.
120 (BT_FN_INT_PTR_INT): Likewise.
121 (BT_FN_UINT_PTR_UINT): Likewise.
122 (BT_FN_LONG_PTR_LONG): Likewise.
123 (BT_FN_ULONG_PTR_ULONG): Likewise.
124 (BT_FN_VOID_UINT64_UINT64): Likewise.
125 (BT_FN_UINT8_UINT8_UINT8): Likewise.
126 (BT_FN_INT8_INT8_INT8): Likewise.
127 (BT_FN_INT16_INT16_INT16): Likewise.
128 (BT_FN_INT_INT_INT): Likewise.
129 (BT_FN_UINT_FLOAT_UINT): Likewise.
130 (BT_FN_FLOAT_UINT_UINT): Likewise.
131 (BT_FN_ULONG_UINT_UINT): Likewise.
132 (BT_FN_ULONG_UINT_PTR): Likewise.
133 (BT_FN_ULONG_ULONG_ULONG): Likewise.
134 (BT_FN_UINT_UINT_UINT): Likewise.
135 (BT_FN_VOID_UINT_PTR): Likewise.
136 (BT_FN_UINT_UINT_PTR: Likewise.
137 (BT_FN_UINT32_UINT64_PTR): Likewise.
138 (BT_FN_INT_INT_UINT_UINT): Likewise.
139 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
140 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
141 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
142 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
143 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
144 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
145 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
146 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
147 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
148 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
149 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
150 * doc/frontends.texi: List BRIG FE.
151 * doc/install.texi (Testing): Add BRIG tesring requirements.
152 * doc/invoke.texi (Overall Options): Mention BRIG.
153 * doc/standards.texi (Standards): Doucment BRIG HSA version.
155 2017-01-24 Richard Biener <rguenther@suse.de>
158 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
160 2017-01-24 Martin Jambor <mjambor@suse.cz>
163 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
164 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
167 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
170 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
171 casts from signed to unsigned really don't have a range.
173 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
175 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
176 GMP_RNDx for compatiblity.
178 2017-01-24 Martin Liska <mliska@suse.cz>
181 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
182 that would prevent us to call alloca with -1 as argument.
184 2017-01-24 Jakub Jelinek <jakub@redhat.com>
186 * dwarf2out.c (output_compilation_unit_header, output_file_names):
187 Avoid -Wformat-security warning.
189 2017-01-23 Andrew Pinski <apinski@cavium.com>
191 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
194 2017-01-23 Martin Sebor <msebor@redhat.com>
197 * gimple-ssa-sprintf.c (warn_level): New global.
198 (format_integer): Use it here and throughout the rest of the file.
199 Use the same switch to compute sign as base.
200 (maybe_warn): New function.
201 (format_directive): Factor out warnings into maybe_warn.
202 Add debugging output. Use warn_level.
203 (add_bytes): Use warn_level.
204 (pass_sprintf_length::compute_format_length): Add debugging output.
205 (try_substitute_return_value): Same.
206 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
209 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
210 (struct fmtresult, format_integer, format_floating): Adjust.
211 (fmtresult::fmtresult): Set max correctly in two argument ctor.
212 (get_string_length, format_string,format_directive): Same.
213 (pass_sprintf_length::compute_format_length): Same.
214 (try_substitute_return_value): Simplify slightly.
217 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
218 (fmtresult::operator+=): Outlined.
219 (struct fmtresult): Add ctors.
220 (struct conversion_spec): Rename...
221 (struct directive): ...to this. Add and remove data members.
222 (directive::set_width, directive::set_precision): New functions.
223 (format_percent): Use fmtresult ctor.
224 (get_width_and_precision): Remove.
225 (format_integer): Make naming changes. Avoid computing width and
227 (format_floating): Same. Adjust indentation.
228 (format_character, format_none): New functions.
229 (format_string): Moved character handling to format_character.
230 (format_directive): Remove arguments, change return type.
231 (parse_directive): New function.
232 (pass_sprintf_length::compute_format_length): Move directive
233 parsing to parse_directive.
235 2017-01-23 Jakub Jelinek <jakub@redhat.com>
237 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
238 (assign_assembler_name_if_needed): ... this.
239 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
240 (assign_assembler_name_if_needed): ... this.
241 (free_lang_data_in_cgraph): Adjust callers.
242 * cgraphunit.c (cgraph_node::analyze): Likewise.
243 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
246 2017-01-23 Richard Biener <rguenther@suse.de>
248 PR tree-optimization/79088
249 PR tree-optimization/79188
250 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
251 resetting loop bounds after last path deletion. Reset loop
252 bounds of the target loop, make code match the comments.
253 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
254 Make sure loops need no fixups.
256 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
258 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
259 exponent support with double type for first argument.
260 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
261 type returned by __builtin_vec_extract_sig,
262 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
263 functions from "vector int" to "vector unsigned int" or from
264 "vector long long int" to "vector unsigned long long int".
265 Changed type returned by __builtin_vec_extract_exp,
266 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
267 functions from "vector int" to "vector unsigned int" or from
268 "vector long long int" to "vector unsigned long long int".
269 Changed return type of __builtin_vec_test_data_class,
270 __builtin_vec_test_data_class_sp, and
271 __builtin_vec_test_data_class_dp from "vector int" to
272 "vector bool int" or from "vector long long int" to "vector bool
273 long long int" and changed second argument type from "unsigned
274 int" to "int". Added new overloaded function forms "vector float
275 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
276 "vector float __builtin_vec_insert_exp_sp (vector float, vector
277 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
278 double, vector unsigned long long int)" and "vector double
279 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
280 long int)". Changed return type of
281 __builtin_scalar_test_data_class and
282 __builtin_scalar_test_data_class_sp and
283 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
284 int" and changed second argument from "unsigned int" to "int".
285 Changed type returned by __builtin_scalar_test_neg,
286 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
287 from "int" to "bool int". Added new overloaded function form
288 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
289 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
290 exponent double-precision with floating point first argument.
291 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
292 documentation of scalar_test_data_class, scalar_test_neg,
293 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
294 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
295 vec_test_data_class built-in functions to reflect refinements in
296 their type signatures.
298 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
300 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
302 (aarch64_elf_asm_destructor): Likewise.
304 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
306 PR rtl-optimization/78634
307 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
308 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
309 * ifcvt.c (noce_try_cmove): Add missing cost check.
311 PR rtl-optimization/71724
312 * combine.c (if_then_else_cond): Look for situations where it is
313 beneficial to undo the work of one of the recursive calls.
315 2017-01-23 Bin Cheng <bin.cheng@arm.com>
317 PR tree-optimization/70754
318 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
319 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
320 combined stmt before it if not NULL.
321 (combine_chains): Process refs reversely and compute dominance point
324 2017-01-23 Martin Liska <mliska@suse.cz>
326 PR tree-optimization/79196
327 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
328 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
330 (strlen_optimize_stmt): Call the renamed function.
332 2017-01-23 Michael Matz <matz@suse.de>
334 PR tree-optimization/78384
335 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
337 2017-01-23 Richard Biener <rguenther@suse.de>
339 PR tree-optimization/79186
340 * tree-vrp.c (register_new_assert_for): Make sure we've seen
341 both incoming edges before moving an assert.
343 2017-01-23 Martin Jambor <mjambor@suse.cz>
345 * ipa-prop.c (load_from_param_1): Removed.
346 (load_from_unmodified_param): Bits from load_from_param_1 put back
348 (load_from_param): Removed.
349 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
350 with stmt. Reverted back to use of load_from_unmodified_param.
352 2017-01-23 Martin Jambor <mjambor@suse.cz>
355 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
356 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
357 field a pointer to garbage collected vector, mark lattices and
358 ipcp_orig_node with GTY((skip)).
359 (ipa_get_param_count): Adjust to descriptors being a pointer.
360 (ipa_get_param): Likewise.
361 (ipa_get_type): Likewise.
362 (ipa_get_param_move_cost): Likewise.
363 (ipa_set_param_used): Likewise.
364 (ipa_get_controlled_uses): Likewise.
365 (ipa_set_controlled_uses): Likewise.
366 (ipa_is_param_used): Likewise.
367 (ipa_node_params_t): Move into garbage collector. New methods insert
369 (ipa_node_params_sum): Annotate wth GTY(()).
370 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
372 (ipa_load_from_parm_agg): Adjust declaration.
373 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
374 * ipa-profile.c (ipa_profile): Likewise.
375 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
376 (ipa_populate_param_decls): Make descriptors parameter garbage
378 (ipa_dump_param): Adjust to descriptors being a pointer.
379 (ipa_alloc_node_params): Likewise.
380 (ipa_initialize_node_params): Likewise.
381 (load_from_param_1): Make descriptors parameter garbage collected.
382 (load_from_unmodified_param): Likewise.
383 (load_from_param): Likewise.
384 (ipa_load_from_parm_agg): Likewise.
385 (ipa_node_params::~ipa_node_params): Removed.
386 (ipa_free_all_node_params): Remove call to delete operator.
387 (ipa_node_params_t::insert): New.
388 (ipa_node_params_t::remove): Likewise.
389 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
390 copy known_csts and known_contexts vectors.
391 (ipa_read_node_info): Adjust to descriptors being a pointer.
392 (ipcp_modif_dom_walker): Make m_descriptors field garbage
394 (ipcp_transform_function): Make descriptors variable garbage
397 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
399 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
400 * config/i386/avx512dqintrin.h: Ditto.
401 * config/i386/avx512fintrin.h: Ditto.
402 * gcc/config/i386/i386.c: Handle new builtins.
403 * config/i386/i386-builtin.def: Add new builtins.
404 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
405 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
407 2017-01-23 Jakub Jelinek <jakub@redhat.com>
408 Martin Liska <mliska@suse.cz>
410 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
411 * asan.c (asan_expand_poison_ifn): Support stores and use
412 appropriate ASAN report function.
413 * internal-fn.c (expand_ASAN_POISON_USE): New function.
414 * internal-fn.def (ASAN_POISON_USE): Declare.
415 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
416 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
417 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
418 ASAN_POISON calls w/o LHS.
419 * tree-ssa.c (execute_update_addresses_taken): Create clobber
420 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
421 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
422 * gimplify.c (asan_poison_variables): Add attribute
423 use_after_scope_memory to variables that really needs to live
425 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
426 having the attribute.
428 2017-01-23 Martin Liska <mliska@suse.cz>
430 * asan.c (create_asan_shadow_var): New function.
431 (asan_expand_poison_ifn): Likewise.
432 * asan.h (asan_expand_poison_ifn): New declaration.
433 * internal-fn.c (expand_ASAN_POISON): Likewise.
434 * internal-fn.def (ASAN_POISON): New builtin.
435 * sanopt.c (pass_sanopt::execute): Expand
436 asan_expand_poison_ifn.
437 * tree-inline.c (copy_decl_for_dup_finish): Make function
439 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
440 * tree-ssa.c (is_asan_mark_p): New function.
441 (execute_update_addresses_taken): Rewrite local variables
442 (identified just by use-after-scope as addressable) into SSA.
444 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
446 * doc/install.texi (Specific): opensource.apple.com uses https
447 now. Remove trailing slash.
449 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
451 * README.Portability: Remove note on an Irix compatibility issue.
453 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
455 * gcov.c (INCLUDE_ALGORITHM): Define.
456 (INCLUDE_VECTOR): Define.
457 No longer include <vector> and <algorithm> directly.
459 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
461 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
463 * doc/invoke.texi (Code Gen Options): Ditto.
465 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
468 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
470 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
472 rtl-optimization/79125
473 * cprop.c (local_cprop_pass): Handle cases where we make an
476 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
480 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
481 read from, for big endian.
483 2017-01-20 Jiong Wang <jiong.wang@arm.com>
485 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
486 register pauth builtins for LP64 only.
488 2017-01-20 Marek Polacek <polacek@redhat.com>
491 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
494 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
496 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
498 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
499 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
500 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
502 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
505 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
506 in big-endian mode when they are not a single duplicated value.
508 2017-01-20 Richard Biener <rguenther@suse.de>
510 * BASE-VER: Bump to 7.0.1.
512 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
514 * omp-low.c (omplow_simd_context): New struct. Use it...
515 (lower_rec_simd_input_clauses): ...here and...
516 (lower_rec_input_clauses): ...here to hold common data. Adjust all
517 references to idx, lane, max_vf, is_simt.
519 2017-01-20 Graham Markall <graham.markall@embecosm.com>
521 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
524 2017-01-20 Martin Jambor <mjambor@suse.cz>
526 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
527 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
529 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
530 (GTFILES): Rename hsa.c to hsa-common.c.
531 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
532 * hsa-dump.c: Likewise.
533 * hsa-gen.c: Likewise.
534 * hsa-regalloc.c: Likewise.
535 * ipa-hsa.c: Likewise.
536 * omp-expand.c: Likewise.
537 * omp-low.c: Likewise.
538 * toplev.c: Likewise.
540 2017-01-20 Marek Polacek <polacek@redhat.com>
543 * doc/invoke.texi: Document -Wduplicated-branches.
544 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
545 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
546 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
547 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
548 return 0 only when not OEP_LEXICOGRAPHIC.
549 (fold_build_cleanup_point_expr): Use the expression
550 location when building CLEANUP_POINT_EXPR.
551 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
552 * tree.c (add_expr): Handle error_mark_node.
554 2017-01-20 Martin Liska <mliska@suse.cz>
557 * tree-profile.c (init_ic_make_global_vars): Do not call
559 (gimple_init_gcov_profiler): Likewise.
561 2017-01-20 Martin Liska <mliska@suse.cz>
564 * cgraph.h (maybe_create_reference): Remove argument and
566 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
568 * ipa-cp.c (create_specialized_node): Likewise.
569 * symtab.c (symtab_node::maybe_create_reference): Handle
570 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
572 2017-01-20 Martin Liska <mliska@suse.cz>
574 * read-rtl-function.c (function_reader::create_function): Use
575 build_decl instread of build_decl_stat.
577 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
579 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
580 * config/i386/avx512dqintrin.h: Ditto.
581 * config/i386/avx512fintrin.h: Ditto.
582 * config/i386/i386-builtin-types.def: Add new types.
583 * gcc/config/i386/i386.c: Handle new types.
584 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
585 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
586 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
587 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
588 (__builtin_ia32_kshiftridi): New.
589 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
591 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
595 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
596 define to rs6000_init_stack_protect_guard.
597 (rs6000_init_stack_protect_guard): New function.
599 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
600 Yunqiang Su <yunqiang.su@imgtec.com>
602 * config.gcc (supported_defaults): Add madd4.
603 (with_madd4): Add validation.
604 (all_defaults): Add madd4.
605 * config/mips/mips.opt (mmadd4): New option.
606 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
608 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
610 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
611 (ISA_HAS_FUSED_MADD4): Likewise.
612 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
613 * gcc/doc/install.texi (--with-madd4): Document the new option.
615 2017-01-19 Jiong Wang <jiong.wang@arm.com>
617 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
618 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
619 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
620 (aarch64_init_pauth_hint_builtins): New.
621 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
622 (aarch64_expand_builtin): Expand new builtins.
624 2017-01-19 Jiong Wang <jiong.wang@arm.com>
626 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
627 * combine-stack-adj.c (no_unhandled_cfa): Handle
628 REG_CFA_TOGGLE_RA_MANGLE.
629 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
630 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
631 info for return address signing.
632 (aarch64_expand_epilogue): Likewise.
634 2017-01-19 Jiong Wang <jiong.wang@arm.com>
636 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
637 * config/aarch64/aarch64-protos.h
638 (aarch64_return_address_signing_enabled): New declaration.
639 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
641 (aarch64_expand_prologue): Sign return address before it's pushed onto
643 (aarch64_expand_epilogue): Authenticate return address fetched from
645 (aarch64_override_options): Sanity check for ILP32 and ISA level.
646 (aarch64_attributes): New function attributes for "sign-return-address".
647 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
648 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
649 ("*do_return"): Generate combined instructions according to key index.
650 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
651 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
653 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
654 * config/aarch64/aarch64.opt (msign-return-address=): New.
655 * doc/extend.texi (AArch64 Function Attributes): Documents
656 "sign-return-address=".
657 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
659 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
661 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
662 overall option summary.
664 2017-01-19 Jiong Wang <jiong.wang@arm.com>
666 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
667 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
668 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
669 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
671 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
673 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
674 -mpower9-minmax by default for -mcpu=power9.
675 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
676 128-bit floating point.
678 2017-01-20 Alan Modra <amodra@gmail.com>
680 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
683 2017-01-20 Alan Modra <amodra@gmail.com>
686 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
687 for strcmp and strncmp from corresponding builtin decl.
689 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
691 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
692 instead of i386/rtems-64.h.
693 * config/i386/rtems-64.h: Remove.
695 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
699 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
701 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
703 2017-01-19 Tamar Christina <tamar.christina@arm.com>
705 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
706 Change int to HOST_WIDE_INT.
707 * config/aarch64/aarch64-protos.h
708 (aarch64_simd_gen_const_vector_dup): Likewise.
709 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
711 2017-01-19 David Malcolm <dmalcolm@redhat.com>
713 * langhooks-def.h (lhd_type_for_size): New decl.
714 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
715 * langhooks.c (lhd_type_for_size): New function, taken from
718 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
720 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
721 define_bypass for CR latency.
722 (power9-cracked-alu): Update bypass latency and remove power9-branch.
723 (power9-alu2): Add define_bypass for CR latency.
725 (power9-mul): Update insn latency.
726 (power9-mul-compare): Update insn latency, bypass latency and remove
729 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
731 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
733 * config/aarch64/aarch64.md
734 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
735 aarch64_nopcrelative_literal_loads.
736 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
738 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
740 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
742 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
744 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
747 * config.gcc (supported_defaults): Add lxc1-sxc1.
748 (with_lxc1_sxc1): Add validation.
749 (all_defaults): Add lxc1-sxc1.
750 * config/mips/mips.opt (mlxc1-sxc1): New option.
751 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
753 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
755 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
756 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
757 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
759 2017-01-19 Richard Biener <rguenther@suse.de>
761 PR tree-optimization/72488
762 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
763 sure to restore SSA info.
764 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
766 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
768 PR rtl-optimization/79121
769 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
770 of the inner type when shifting an extended value.
772 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
775 * symtab.c (symtab_node::equal_address_to): Fix comparing of
776 interposable aliases.
778 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
781 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
782 Use the evmergelohi instruction.
783 (mov_si<mode>_e500_subreg4_2_le): Likewise.
784 (mov_sitf_e500_subreg8_2_be): Likewise.
785 (mov_sitf_e500_subreg12_2_le): Likewise.
786 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
787 (mov_si<mode>_e500_subreg4_2_be): Likewise.
788 (mov_sitf_e500_subreg8_2_le): Likewise.
789 (mov_sitf_e500_subreg12_2_be): Likewise.
791 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
793 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
794 attribute from vecsimple to vecperm.
795 (altivec_vbpermq2): Likewise.
797 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
800 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
802 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
803 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
804 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
805 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
806 case where N arg is SIZE_MAX.
807 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
808 (cmpstrsi): Add pattern.
810 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
812 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
813 __builtin_vec_revb builtins.
814 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
815 built-in functions to support generation of the ISA 3.0 XXBR<x>
816 vector byte reverse instructions.
817 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
818 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
819 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
820 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
821 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
822 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
823 (P9V_BUILTIN_VEC_REVB): Likewise.
824 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
825 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
826 (p9_xxbrq_v16qi): Likewise.
827 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
828 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
829 (p9_xxbrh_v8hi): Likewise.
830 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
831 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
832 vec_revb built-in functions.
834 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
836 PR rtl-optimization/78952
837 * config/i386/i386.md (any_extract): New code iterator.
838 (*insvqi_2): Use any_extract for source operand.
839 (*insvqi_3): Use any_shiftrt for source operand.
841 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
843 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
845 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
847 2017-01-18 Matthias Klose <doko@ubuntu.com>
849 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
851 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
853 * config/rs6000/altivec.h (vec_bperm): Change #define.
854 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
855 (altivec_vbpermq2): New define_insn.
856 (altivec_vbpermd): Likewise.
857 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
860 (VBPERM): New polymorphic function interface.
861 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
862 Add entries for P9V_BUILTIN_VEC_VBPERM.
863 * doc/extend.texi: Add interfaces for vec_bperm.
865 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
867 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
868 first letter of error messages.
869 (s390_resolve_overloaded_builtin): Likewise.
870 * config/s390/s390.c (s390_expand_builtin): Likewise.
871 (s390_invalid_arg_for_unprototyped_fn): Likewise.
872 (s390_valid_target_attribute_inner_p): Likewise.
873 * config/s390/s390.md ("tabort"): Likewise.
875 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
877 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
878 (ISA_AVOID_DIV_HILO): New macro.
879 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
880 (ISA_HAS_DDIV): Likewise.
882 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
884 * doc/invoke.texi (fabi-version): Correct number of occurrences.
886 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
888 * doc/invoke.texi (fabi-version): Spelling fix.
890 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
893 * doc/invoke.texi (fabi-version): Mention mangling fix for
896 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
899 * doc/invoke.texi (fabi-version): Document discriminator mangling.
901 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
904 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
905 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
907 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
909 (stack_protect_test): Ditto.
910 * config/rs6000/rs6000.opt (mstack-protector-guard=,
911 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
913 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
914 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
915 -mstack-protector-guard-offset=.
916 (RS/6000 and PowerPC Options): Ditto.
918 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
920 * config/i386/i386.h (MASK_CLASS_P): New define.
921 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
922 there are no registers from different register sets also when
923 mask registers are used. Update function comment.
924 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
925 to (*k/*r) and (*k/*km) alternatives.
927 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
929 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
930 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
931 (EH_RETURN_HANDLER_RTX): New define.
932 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
933 Force frame pointer in EH return functions.
934 (aarch64_expand_epilogue): Add barrier for eh_return.
935 (aarch64_final_eh_return_addr): Remove.
936 (aarch64_eh_return_handler_rtx): New function.
937 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
939 (aarch64_eh_return_handler_rtx): New prototype.
941 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
943 * config/rs6000/altivec.h (vec_rlmi): New #define.
944 (vec_vrlnm): Likewise.
945 (vec_rlnm): Likewise.
946 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
947 (UNSPEC_VRLNM): Likewise.
948 (VIlong): New mode iterator.
949 (altivec_vrl<VI_char>mi): New define_insn.
950 (altivec_vrl<VI_char>nm): Likewise.
951 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
954 (RLNM): New polymorphic function entry.
955 (VRLWMI): New monomorphic function entry.
957 (RLMI): New polymorphic function entry.
958 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
959 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
960 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
963 2017-01-17 Jakub Jelinek <jakub@redhat.com>
966 * dwarf2out.c (field_byte_offset): Restore the
967 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
968 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
969 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
972 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
975 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
977 2017-01-17 Jakub Jelinek <jakub@redhat.com>
980 * dwarf2out.c (add_data_member_location_attribute): For constant
981 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
982 instead of DW_AT_data_member_location, DW_AT_bit_offset and
983 DW_AT_byte_size attributes.
985 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
987 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
988 after forcing to constant memory when the code model is medium.
990 2017-01-17 Julia Koval <julia.koval@intel.com>
993 * config/i386/avx512fintrin.h
994 (_mm512_i32gather_ps): Change __addr type to void const*.
995 (_mm512_mask_i32gather_ps): Ditto.
996 (_mm512_i32gather_pd): Ditto.
997 (_mm512_mask_i32gather_pd): Ditto.
998 (_mm512_i64gather_ps): Ditto.
999 (_mm512_mask_i64gather_ps): Ditto.
1000 (_mm512_i64gather_pd): Ditto.
1001 (_mm512_mask_i64gather_pd): Ditto.
1002 (_mm512_i32gather_epi32): Ditto.
1003 (_mm512_mask_i32gather_epi32): Ditto.
1004 (_mm512_i32gather_epi64): Ditto.
1005 (_mm512_mask_i32gather_epi64): Ditto.
1006 (_mm512_i64gather_epi32): Ditto.
1007 (_mm512_mask_i64gather_epi32): Ditto.
1008 (_mm512_i64gather_epi64): Ditto.
1009 (_mm512_mask_i64gather_epi64): Ditto.
1010 (_mm512_i32scatter_ps): Change __addr type to void*.
1011 (_mm512_mask_i32scatter_ps): Ditto.
1012 (_mm512_i32scatter_pd): Ditto.
1013 (_mm512_mask_i32scatter_pd): Ditto.
1014 (_mm512_i64scatter_ps): Ditto.
1015 (_mm512_mask_i64scatter_ps): Ditto.
1016 (_mm512_i64scatter_pd): Ditto.
1017 (_mm512_mask_i64scatter_pd): Ditto.
1018 (_mm512_i32scatter_epi32): Ditto.
1019 (_mm512_mask_i32scatter_epi32): Ditto.
1020 (_mm512_i32scatter_epi64): Ditto.
1021 (_mm512_mask_i32scatter_epi64): Ditto.
1022 (_mm512_i64scatter_epi32): Ditto.
1023 (_mm512_mask_i64scatter_epi32): Ditto.
1024 (_mm512_i64scatter_epi64): Ditto.
1025 (_mm512_mask_i64scatter_epi64): Ditto.
1026 * config/i386/avx512pfintrin.h
1027 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1028 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1029 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1030 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1031 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1032 (_mm512_prefetch_i32scatter_ps): Ditto.
1033 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1034 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1035 (_mm512_prefetch_i64scatter_pd): Ditto.
1036 (_mm512_prefetch_i64scatter_ps): Ditto.
1037 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1038 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1039 * config/i386/avx512vlintrin.h
1040 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1041 (_mm_mmask_i32gather_ps): Ditto.
1042 (_mm256_mmask_i32gather_pd): Ditto.
1043 (_mm_mmask_i32gather_pd): Ditto.
1044 (_mm256_mmask_i64gather_ps): Ditto.
1045 (_mm_mmask_i64gather_ps): Ditto.
1046 (_mm256_mmask_i64gather_pd): Ditto.
1047 (_mm_mmask_i64gather_pd): Ditto.
1048 (_mm256_mmask_i32gather_epi32): Ditto.
1049 (_mm_mmask_i32gather_epi32): Ditto.
1050 (_mm256_mmask_i32gather_epi64): Ditto.
1051 (_mm_mmask_i32gather_epi64): Ditto.
1052 (_mm256_mmask_i64gather_epi32): Ditto.
1053 (_mm_mmask_i64gather_epi32): Ditto.
1054 (_mm256_mmask_i64gather_epi64): Ditto.
1055 (_mm_mmask_i64gather_epi64): Ditto.
1056 (_mm256_i32scatter_ps): Change __addr type to void*.
1057 (_mm256_mask_i32scatter_ps): Ditto.
1058 (_mm_i32scatter_ps): Ditto.
1059 (_mm_mask_i32scatter_ps): Ditto.
1060 (_mm256_i32scatter_pd): Ditto.
1061 (_mm256_mask_i32scatter_pd): Ditto.
1062 (_mm_i32scatter_pd): Ditto.
1063 (_mm_mask_i32scatter_pd): Ditto.
1064 (_mm256_i64scatter_ps): Ditto.
1065 (_mm256_mask_i64scatter_ps): Ditto.
1066 (_mm_i64scatter_ps): Ditto.
1067 (_mm_mask_i64scatter_ps): Ditto.
1068 (_mm256_i64scatter_pd): Ditto.
1069 (_mm256_mask_i64scatter_pd): Ditto.
1070 (_mm_i64scatter_pd): Ditto.
1071 (_mm_mask_i64scatter_pd): Ditto.
1072 (_mm256_i32scatter_epi32): Ditto.
1073 (_mm256_mask_i32scatter_epi32): Ditto.
1074 (_mm_i32scatter_epi32): Ditto.
1075 (_mm_mask_i32scatter_epi32): Ditto.
1076 (_mm256_i32scatter_epi64): Ditto.
1077 (_mm256_mask_i32scatter_epi64): Ditto.
1078 (_mm_i32scatter_epi64): Ditto.
1079 (_mm_mask_i32scatter_epi64): Ditto.
1080 (_mm256_i64scatter_epi32): Ditto.
1081 (_mm256_mask_i64scatter_epi32): Ditto.
1082 (_mm_i64scatter_epi32): Ditto.
1083 (_mm_mask_i64scatter_epi32): Ditto.
1084 (_mm256_i64scatter_epi64): Ditto.
1085 (_mm256_mask_i64scatter_epi64): Ditto.
1086 (_mm_i64scatter_epi64): Ditto.
1087 (_mm_mask_i64scatter_epi64): Ditto.
1088 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1089 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1090 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1091 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1092 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1093 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1094 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1095 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1096 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1097 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1098 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1099 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1100 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1101 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1102 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1103 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1104 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1105 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1106 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1107 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1108 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1109 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1110 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1111 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1112 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1113 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1114 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1115 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1116 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1117 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1118 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1119 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1120 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1121 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1122 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1123 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1124 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1125 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1126 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1127 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1128 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1129 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1130 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1131 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1132 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1133 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1134 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1135 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1136 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1137 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1138 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1139 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1140 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1141 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1142 definitions accordingly.
1144 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1145 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1148 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1151 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1154 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1155 hard regs for inner regno.
1157 2017-01-17 Martin Liska <mliska@suse.cz>
1160 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1161 assumption and add comment.
1163 2017-01-17 Nathan Sidwell <nathan@acm.org>
1165 * ipa-visibility.c (localize_node): New function, broken out of ...
1166 (function_and_variable_visibility): ... here. Call it.
1168 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1171 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1172 correctly set frequency of oudgoing edge.
1173 (duplicate_thread_path): Fix profile updating.
1175 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1178 * configure.ac: Add GCC_BASE_VER.
1179 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1180 version from BASE-VER file.
1181 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1182 (gcc.o): Depend on $(BASEVER).
1183 * common.opt (dumpfullversion): New option.
1184 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1185 * doc/invoke.texi: Document -dumpfullversion.
1186 * doc/install.texi: Document --with-gcc-major-version-only.
1187 * configure: Regenerated.
1189 2017-01-17 Richard Biener <rguenther@suse.de>
1191 PR tree-optimization/71433
1192 * tree-vrp.c (register_new_assert_for): Merge same asserts
1193 on all incoming edges.
1194 (process_assert_insertions_for): Handle insertions at the
1197 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1199 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1200 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1202 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1205 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1208 2017-01-17 Alan Modra <amodra@gmail.com>
1211 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1212 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1213 symbolic stack limit when pic.
1215 2017-01-16 Martin Sebor <msebor@redhat.com>
1217 PR tree-optimization/78608
1218 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1220 2017-01-16 Jeff Law <law@redhat.com>
1223 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1224 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1225 for several include directories that may be relative to sysroot.
1226 * config/i386/x-mingw32 (gplus_includedir): Define.
1227 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1228 (native_system_includedir): Likewise.
1229 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1230 override if TARGET_SYSTEM_ROOT is defined.
1231 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1233 PR tree-optimization/79090
1234 PR tree-optimization/33562
1235 PR tree-optimization/61912
1236 PR tree-optimization/77485
1237 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1238 and computed trims into the dump file.
1240 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1242 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1244 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1247 * gimplify.c (gimplify_init_constructor): If want_value and
1248 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1252 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1253 sequence. Formatting fixes.
1254 (doloop_optimize): Formatting fixes.
1257 * gcc.c (debug_level_greater_than_spec_func): New function.
1258 (static_spec_functions): Add debug-level-gt spec function.
1259 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
1261 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1262 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1263 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
1264 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
1265 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
1266 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
1268 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
1270 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
1271 QImode fixups to general and mask registers only.
1273 2017-01-16 Carl Love <cel@us.ibm.com>
1275 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
1276 for built-in functions
1277 vector signed char vec_nabs (vector signed char)
1278 vector signed short vec_nabs (vector signed short)
1279 vector signed int vec_nabs (vector signed int)
1280 vector signed long long vec_nabs (vector signed long long)
1281 vector float vec_nabs (vector float)
1282 vector double vec_nabs (vector double)
1283 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
1285 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
1286 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
1287 * doc/extend.texi: Update the documentation file for the new built-in
1290 2017-01-16 Martin Sebor <msebor@redhat.com>
1292 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
1295 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1297 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
1298 UNSPEC_VSX__XXSPLTD to require special splat handling.
1300 2017-01-16 David Malcolm <dmalcolm@redhat.com>
1303 * system.h: Poison strndup.
1305 2017-01-16 Alan Modra <amodra@gmail.com>
1308 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
1311 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
1313 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
1315 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
1317 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
1318 call recog here. Assert that INSN_CODE (insn) is non-negative.
1320 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1323 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
1325 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
1326 in the currently scheduled RTL fragment.
1328 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1330 PR rtl-optimization/78751
1331 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
1334 2017-01-14 Jeff Law <law@redhat.com>
1336 PR tree-optimization/79090
1337 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
1338 variable length stores.
1339 (compute_trims): Delete dead assignment to *trim_tail.
1340 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
1343 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
1345 PR rtl-optimization/78626
1346 PR rtl-optimization/78727
1347 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
1348 of a block, and split such blocks after everything else is finished.
1350 2017-01-14 Alan Modra <amodra@gmail.com>
1353 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
1354 target legitimate_combined_insn.
1355 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
1356 (rs6000_legitimate_combined_insn): New function.
1357 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
1359 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
1360 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
1361 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
1363 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
1365 * doc/frontends.texi (G++ and GCC): Remove references to Java.
1367 2017-01-13 Jeff Law <law@redhat.com>
1369 PR tree-optimization/33562
1370 PR tree-optimization/61912
1371 PR tree-optimization/77485
1372 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
1374 (delete_dead_assignment): Likewise.
1375 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
1376 statement to delete_dead_call and delete_dead_assignment.
1378 2017-01-13 David Malcolm <dmalcolm@redhat.com>
1381 * substring-locations.c (format_warning_va): Strengthen case 1 so
1382 that both endpoints of the substring must be within the format
1383 range for just the substring to be printed.
1385 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
1387 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
1388 * config/i386/i386.c (ix86_target_string): Add missing options
1389 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
1390 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
1391 flags_other and ix86_target_other to flags2_other. Display unknown
1393 (ix86_valid_target_attribute_inner_p): Add missing options and
1394 reorder options by implied ISAs, as in ix86_target_string.
1396 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1398 * hash-table.h (hash_table::too_empty_p): New function.
1399 (hash_table::expand): Use it.
1400 (hash_table::traverse): Likewise.
1401 (hash_table::empty_slot): Use sizeof (value_type) instead of
1402 sizeof (PTR) to convert bytes to elements. Shrink the table
1403 if the current size is excessive for the current number of
1406 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1408 * ira-costs.c (record_reg_classes): Break from the inner loop
1409 early once alt_fail is known to be true. Update outer loop
1410 handling accordingly.
1412 2017-01-13 Jeff Law <law@redhat.com>
1414 * tree-ssa-dse.c (decrement_count): New function.
1415 (increment_start_addr, maybe_trim_memstar_call): Likewise.
1416 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
1417 when we know the partially dead statement is a mem* function.
1419 PR tree-optimization/61912
1420 PR tree-optimization/77485
1421 * tree-ssa-dse.c: Include expr.h.
1422 (maybe_trim_constructor_store): New function.
1423 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
1425 PR tree-optimization/33562
1426 PR tree-optimization/61912
1427 PR tree-optimization/77485
1428 * doc/invoke.texi: Document new dse-max-object-size param.
1429 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
1430 * tree-ssa-dse.c: Include params.h.
1431 (dse_store_status): New enum.
1432 (initialize_ao_ref_for_dse): New, partially extracted from
1434 (valid_ao_ref_for_dse, normalize_ref): New.
1435 (setup_live_bytes_from_ref, compute_trims): Likewise.
1436 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
1437 (maybe_trim_partially_dead_store): Likewise.
1438 (maybe_trim_complex_store): Likewise.
1439 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
1440 Track what bytes live from the original store. Return tri-state
1441 for dead, partially dead or live.
1442 (dse_dom_walker): Add constructor, destructor and new private members.
1443 (delete_dead_call, delete_dead_assignment): New extracted from
1445 (dse_optimize_stmt): Make a member of dse_dom_walker.
1446 Use initialize_ao_ref_for_dse.
1448 PR tree-optimization/33562
1449 PR tree-optimization/61912
1450 PR tree-optimization/77485
1451 * sbitmap.h (bitmap_count_bits): Prototype.
1452 (bitmap_clear_range, bitmap_set_range): Likewise.
1453 * sbitmap.c (bitmap_clear_range): New function.
1454 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
1456 2017-01-13 Martin Liska <mliska@suse.cz>
1459 * function.c (set_cfun): Add new argument force.
1460 * function.h (set_cfun): Likewise.
1461 * ipa-inline-transform.c (inline_call): Use the function when
1462 strict alising from is dropped for function we inline to.
1464 2017-01-13 Richard Biener <rguenther@suse.de>
1466 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
1467 for dumping GIMPLE INTEGER_CSTs.
1469 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1471 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
1472 to 201112L since C++17.
1474 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
1477 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
1478 if -fsanitize=kernel-address is present.
1480 2017-01-13 Richard Biener <rguenther@suse.de>
1482 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
1483 as _Literal ( type ) number in case usual suffixes do not
1484 preserve all information.
1486 2017-01-13 Richard Biener <rguenther@suse.de>
1488 PR tree-optimization/77283
1489 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
1490 and ssa-iterators.h.
1491 (is_feasible_trace): Implement a cost model based on joiner
1494 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1497 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
1498 char or short to __float128/_Float128 directly.
1500 2017-01-12 Martin Sebor <msebor@redhat.com>
1502 to -Wformat-overflow.
1503 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
1504 (min_bytes_remaining): Same.
1505 (get_string_length): Same.
1506 (format_string): Same.
1507 (format_directive): Same.
1509 (pass_sprintf_length::handle_gimple_call): Same.
1511 2017-01-12 Jakub Jelinek <jakub@redhat.com>
1513 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
1514 info.nowrite calls with no lhs that can't throw. Return bool
1515 whether gsi_remove has been called or not.
1516 (pass_sprintf_length::handle_gimple_call): Return bool whether
1517 try_substitute_return_value called gsi_remove. Formatting fix.
1518 (pass_sprintf_length::execute): Don't use gsi_remove if
1519 handle_gimple_call returned true.
1522 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
1523 be removed due to side-effects, don't remove following barrier nor
1524 turn the successor edge into fallthru edge.
1526 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1529 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
1530 element-reversing loads and stores as not swappable.
1532 2017-01-12 Nathan Sidwell <nathan@acm.org>
1533 Nicolai Stange <nicstange@gmail.com>
1535 * combine.c (try_combine): Don't ignore result of overlap checking
1536 loop. Combine overlap & asm check into single loop.
1538 2017-01-12 Richard Biener <rguenther@suse.de>
1540 * tree-pretty-print.c (dump_generic_node): Provide -gimple
1541 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
1543 2017-01-12 Richard Biener <rguenther@suse.de>
1545 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
1546 and TS_TARGET_OPTION directly derive from TS_BASE.
1547 * tree-core.h (tree_optimization_option): Derive from tree_base.
1548 (tree_target_option): Likewise.
1550 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
1552 * config/i386/i386.c (memory_address_length): Increase len
1553 only when rip_relative_addr_p returns false.
1555 2017-01-11 Julia Koval <julia.koval@intel.com>
1557 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
1558 (OPTION_MASK_ISA_SGX_SET): New.
1559 (ix86_handle_option): Handle OPT_msgx.
1560 * config.gcc: Added sgxintrin.h.
1561 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
1562 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
1563 * config/i386/i386.c (ix86_target_string): Add -msgx.
1565 (ix86_option_override_internal): Handle new options.
1566 (ix86_valid_target_attribute_inner_p): Add sgx.
1567 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
1568 * config/i386/i386.opt: Add msgx.
1569 * config/i386/sgxintrin.h: New file.
1570 * config/i386/x86intrin.h: Add sgxintrin.h.
1572 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1575 * fold-const.c (maybe_nonzero_address): Return 1 for function
1577 (tree_single_nonzero_warnv_p): Don't handle function local objects
1581 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
1584 2017-01-11 David Malcolm <dmalcolm@redhat.com>
1587 * opts.c: Include "spellcheck.h"
1588 (struct string_fragment): New struct.
1589 (struct edit_distance_traits<const string_fragment &>): New
1591 (get_closest_sanitizer_option): New function.
1592 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
1594 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1596 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
1598 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
1599 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
1600 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
1601 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
1602 for initial die_offset if dwarf_split_debug_info.
1603 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
1604 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
1606 (output_skeleton_debug_sections): Formatting fix. Use
1607 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
1608 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
1610 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
1612 * config/arm/cortex-a53.md: Add bypasses for
1614 (cortex_a53_r2f): Only use for transfers.
1615 (cortex_a53_f2r): Likewise.
1616 (cortex_a53_r2f_cvt): Add reservation for conversions.
1617 (cortex_a53_f2r_cvt): Likewise.
1619 2017-01-11 Tamar Christina <tamar.christina@arm.com>
1621 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
1622 to all inlined functions, change static to extern.
1624 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
1627 * config/arm/arm.c (legitimize_pic_address): Handle reference to
1629 (arm_assemble_integer): Likewise.
1631 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
1633 * config.gcc: Use new awk script to check CPU, FPU and architecture
1634 parameters for --with-... options.
1635 * config/arm/parsecpu.awk: New file
1636 * config/arm/arm-cpus.in: New file.
1637 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
1639 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
1641 * config/arm/t-arm: Update dependency rules.
1642 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
1643 of processing .def files.
1644 * config/arm/genopt.sh: Deleted.
1645 * config/arm/gentune.sh: Deleted.
1646 * config/arm/arm-cores.def: Deleted.
1647 * config/arm/arm-arches.def: Deleted.
1648 * config/arm/arm-fpus.def: Deleted.
1649 * config/arm/arm-tune.md: Regenerated.
1650 * config/arm/arm-tables.opt: Regenerated.
1651 * config/arm/arm-cpu.h: New generated file.
1652 * config/arm/arm-cpu-data.h: New generated file.
1653 * config/arm/arm-cpu-cdata.h: New generated file.
1655 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
1658 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
1660 (input_varpool_node): Unpack dynamically_initialized bit.
1662 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
1664 PR rtl-optimization/79032
1665 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
1666 the alignment of the adjusted memory reference against that of MODE,
1667 instead of the alignment of the original memory reference.
1669 2017-01-11 Martin Jambor <mjambor@suse.cz>
1671 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
1673 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
1674 decorated functions.
1676 2017-01-11 Richard Biener <rguenther@suse.de>
1678 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
1679 set range/nonnull info for PHI results. Do not set it on
1680 stmts marked for removal.
1682 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
1684 * expr.c (store_field): In the bitfield case, fetch the return value
1685 from the registers before applying a single big-endian adjustment.
1686 Always do a final load for a BLKmode value not larger than a word.
1688 2017-01-10 David Malcolm <dmalcolm@redhat.com>
1691 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
1692 that we correctly handle column numbers greater than
1693 LINE_MAP_MAX_COLUMN_NUMBER.
1695 2017-01-10 Martin Sebor <msebor@redhat.com>
1698 * gimple-ssa-sprintf.c (get_destination_size): Call
1699 {init,fini}object_sizes.
1700 * tree-object-size.c (addr_object_size): Adjust.
1701 (pass_through_call): Adjust.
1702 (pass_object_sizes::execute): Adjust.
1703 * tree-object-size.h (fini_object_sizes): Declare.
1705 2017-01-10 Martin Sebor <msebor@redhat.com>
1707 PR tree-optimization/78775
1708 * builtins.c (get_size_range): Move...
1709 * calls.c: ...to here.
1710 (alloc_max_size): Accept zero argument.
1711 (operand_signed_p): Remove.
1712 (maybe_warn_alloc_args_overflow): Call get_size_range.
1713 * calls.h (get_size_range): Declare.
1715 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
1717 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
1718 from TI's devices.csv file as of September 2016.
1719 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
1721 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
1723 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
1724 * doc/invoke.texi: Likewise.
1725 * doc/md.texi: Likewise.
1726 * doc/objc.texi: Likewise.
1728 2017-01-10 Joshua Conner <joshconner@google.com>
1730 * config/arm/fuchsia-elf.h: New file.
1731 * config/fuchsia.h: New file.
1732 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
1733 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
1735 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
1737 2016-01-10 Richard Biener <rguenther@suse.de>
1739 PR tree-optimization/79034
1740 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
1741 Propagate out degenerate PHIs in the joiner.
1743 2017-01-10 Martin Liska <mliska@suse.cz>
1745 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
1746 (sort_congruence_classes_by_decl_uid): Likewise.
1747 (sort_congruence_class_groups_by_decl_uid): Likewise.
1748 (sem_item_optimizer::merge_classes): Sort class, groups in these
1749 classes and members in the groups by DECL_UID of declarations.
1750 This would make merge operations stable.
1752 2017-01-10 Martin Liska <mliska@suse.cz>
1754 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
1755 usage of m_classes_vec.
1756 (sem_item_optimizer::~sem_item_optimizer): Likewise.
1757 (sem_item_optimizer::get_group_by_hash): Likewise.
1758 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
1759 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
1760 (sem_item_optimizer::verify_classes): Likewise.
1761 (sem_item_optimizer::process_cong_reduction): Likewise.
1762 (sem_item_optimizer::dump_cong_classes): Likewise.
1763 (sem_item_optimizer::merge_classes): Likewise.
1764 * ipa-icf.h (congruence_class_hash): Rename from
1765 congruence_class_group_hash. Remove declaration of m_classes_vec.
1767 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
1769 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
1770 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
1771 * config.gcc: Add avx512vpopcntdqintrin.h.
1772 * config/i386/avx512vpopcntdqintrin.h: New.
1773 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
1774 * config/i386/i386-builtin-types.def: Add new types.
1775 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
1776 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
1777 __builtin_ia32_vpopcountq_v8di_mask): New.
1778 * config/i386/i386-c.c (ix86_target_macros_internal): Define
1779 __AVX512VPOPCNTDQ__.
1780 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
1781 (PTA_AVX512VPOPCNTDQ): Define.
1782 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
1783 TARGET_AVX512VPOPCNTDQ_P): Define.
1784 * config/i386/i386.opt: Add mavx512vpopcntdq.
1785 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
1786 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
1788 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
1791 * predict.def (PRED_CALL): Set to 67.
1793 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
1795 * expr.c (store_field): In the bitfield case, if the value comes from
1796 a function call and is of an aggregate type returned in registers, do
1797 not modify the field mode; extract the value in all cases if the mode
1798 is BLKmode and the size is not larger than a word.
1800 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
1803 * config/i386/cpuid.h: Fix undefined behavior.
1805 2017-01-04 Jeff Law <law@redhat.com>
1807 PR tree-optimization/79007
1808 PR tree-optimization/67955
1809 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
1810 conservative for pt.null when flag_non_call_exceptions is on.
1812 2017-01-09 Jakub Jelinek <jakub@redhat.com>
1814 PR translation/79019
1815 PR translation/79020
1816 * params.def (PARAM_INLINE_MIN_SPEEDUP,
1817 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
1818 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
1820 * config/avr/avr.opt (maccumulate-args): Likewise.
1821 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
1822 * common.opt (freport-bug): Likewise.
1823 * cif-code.def (CIF_FINAL_ERROR): Likewise.
1824 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
1825 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
1826 translatable string.
1827 * config/i386/i386.c (function_value_32): Likewise.
1828 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
1829 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
1831 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
1832 * common/config/msp430/msp430-common.c (msp430_handle_option):
1834 * symtab.c (symtab_node::verify_base): Likewise.
1835 * opts.c (set_debug_level): Likewise.
1836 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
1837 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
1838 missing whitespace to translatable strings.
1839 * config/avr/avr.md (bswapsi2): Fix typo in comment.
1840 * config/sh/superh.h: Likewise.
1841 * config/i386/xopintrin.h: Likewise.
1842 * config/i386/znver1.md: Likewise.
1843 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
1844 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
1845 * double-int.h (struct double_int): Likewise.
1846 * double-int.c (div_and_round_double): Likewise.
1847 * wide-int.cc: Likewise.
1848 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1849 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
1850 * cfgcleanup.c (crossjumps_occured): Renamed to ...
1851 (crossjumps_occurred): ... this.
1852 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
1855 PR tree-optimization/78899
1856 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
1857 returning bool return struct loop *, NULL for failure and the new
1859 (versionable_outer_loop_p): Don't version outer loop if it has
1860 dont_vectorized bit set.
1861 (tree_if_conversion): When versioning outer loop, ensure
1862 tree_if_conversion is performed also on the inner loop of the
1863 non-vectorizable outer loop copy.
1864 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
1865 LOOP_VECTORIZED in inner loop of the scalar outer loop and
1866 prevent vectorization of it.
1867 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
1868 the outer loop vectorization of the non-scalar version is attempted
1869 before vectorization of the inner loop in scalar version. If
1870 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
1871 vectorization of its inner loop.
1872 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
1873 has 2 inner loops, rename also on edges from bb whose single pred
1874 is outer_loop->header. Fix typo in function comment.
1876 2017-01-09 Martin Sebor <msebor@redhat.com>
1879 * asan.c (asan_emit_stack_protection): Increase local buffer size
1880 to avoid snprintf truncation warning.
1882 2017-01-09 Andrew Pinski <apinski@cavium.com>
1884 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
1885 to reference thunderx2t99 for the tuning structure
1886 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
1888 (thunderx2t99_extra_costs): This.
1889 * config/aarch64/aarch64-tune.md: Regenerate.
1890 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
1891 (vulcan_addrcost_table): This.
1892 (vulcan_regmove_cost): Rename to ...
1893 (thunderx2t99_regmove_cost): This.
1894 (vulcan_vector_cost): Rename to ...
1895 (thunderx2t99_vector_cost): this.
1896 (vulcan_branch_cost): Rename to ...
1897 (thunderx2t99_branch_cost): This.
1898 (vulcan_tunings): Rename to ...
1899 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
1900 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
1902 2017-01-09 Martin Jambor <mjambor@suse.cz>
1906 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
1907 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
1908 (propagate_vr_accross_jump_function): Use the above function for all
1909 value range computations for pass-through jump functions and type
1910 converasion from explicit value range values.
1911 (ipcp_propagate_stage): Do not attempt to deduce types of formal
1912 parameters from TYPE_ARG_TYPES.
1913 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
1914 (ipa_write_node_info): Stream type of the actual argument.
1915 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
1917 2017-01-09 Martin Liska <mliska@suse.cz>
1920 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
1921 (lookup_compiler): Do not show error message with have_E.
1923 2017-01-09 Jakub Jelinek <jakub@redhat.com>
1925 PR tree-optimization/78938
1926 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
1927 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
1928 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
1929 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
1932 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1934 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
1937 2017-01-09 Richard Biener <rguenther@suse.de>
1939 PR tree-optimization/78997
1940 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
1941 name condition properly.
1943 2017-01-09 Richard Biener <rguenther@suse.de>
1946 * dwarf2out.c (is_cxx): New overload with context.
1947 (is_naming_typedef_decl): Use it.
1949 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
1951 * invoke.texi (Option Summary): Correct spacing in option lists
1952 and add line breaks to fix over-long lines.
1954 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
1958 * extend.texi (Common Variable Attributes): Add xref to GCC
1959 Internals manual to explain mode attribute keywords.
1961 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
1964 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
1965 and Preprocessor Options.
1966 (Options for Linking): Document -pthread here....
1967 (RS/6000 and PowerPC Options): ...not here.
1968 (Solaris 2 Options): ...or here.
1969 * doc/cppopts.texi: Document -pthread.
1971 2017-01-08 Martin Sebor <msebor@redhat.com>
1974 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
1975 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
1976 New member functions.
1977 (format_directive): Used them.
1979 (pass_sprintf_length::handle_gimple_call): Same.
1980 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
1981 to avoid truncation for any argument.
1982 (extract_affine_mul): Same.
1983 * tree.c (get_file_function_name): Same.
1985 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
1988 * predict.def (PRED_INDIR_CALL): Set to 86.
1990 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
1992 PR preprocessor/54124
1993 * doc/cppopts.texi: Reformat -d subtable to list the full name
1994 of the options. Add cross-reference to the docs for the general
1995 compiler -d options.
1996 * doc/invoke.texi (Developer Options): Add cross-reference to the
1997 preprocessor-specific -d option documentation.
1999 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2001 PR preprocessor/13498
2002 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2003 redudant material, and reflect new command-line options.
2004 (System Headers): Likewise.
2006 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2008 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2009 -isystem, and -idirafter. Copy-edit.
2010 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2011 default for -ftrack-macro-expansion. Delete obsolete and
2012 badly-formatted implementation details about -fdebug-cpp output.
2013 * doc/cppwarnopts.texi: Copy-edit.
2015 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2018 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2019 that the transition from a max line width >= 1<<10 to narrower
2020 lines works correctly.
2022 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2024 * doc/options.texi (PerFunction): New.
2025 * opt-functions.awk (switch_flags): Map both Optimization and
2026 PerFunction to CL_OPTIMIZATION.
2027 * opth-gen.awk: Test for PerFunction flag along with
2029 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2030 it only when the latter is present. Skip those that don't in
2031 the hash function generator.
2032 * common.opt (fvar-tracking): Mark as PerFunction instead of
2034 (fvar-tracking-assignments): Likewise.
2035 (fvar-tracking-assignments-toggle): Likewise.
2036 (fvar-tracking-uninit): Likewise.
2038 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2040 PR translation/79018
2041 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2044 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2047 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2048 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2049 TARGET_LONG_JUMP_TABLE_OFFSETS.
2050 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2051 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2052 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2053 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2054 * config/m68k/m68k.md (tablejump expander): Likewise.
2055 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2056 TARGET_LONG_JUMP_TABLE_OFFSETS.
2057 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2058 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2060 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2061 David Holsgrove <david.holsgrove@xilinx.com>
2063 * common/config/microblaze/microblaze-common.c
2064 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2065 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2067 * config/microblaze/microblaze.c (microblaze_must_save_register)
2068 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2070 (microblaze_eh_return): New function.
2071 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2072 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2073 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2074 * config/microblaze/microblaze.md (eh_return): New pattern.
2076 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2078 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2079 GCC_DIAGNOSTIC_STRINGIFY): Define.
2081 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2083 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2085 * config/arm/arm.md (<mcrr>): New.
2087 * config/arm/arm.c (arm_arch5te): New.
2088 (arm_option_override): Set arm_arch5te.
2089 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2091 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2092 (arm_mcrr_qualifiers): ... this. New.
2093 (MRRC_QUALIFIERS): Define to...
2094 (arm_mrrc_qualifiers): ... this. New.
2095 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2097 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2098 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2099 (MRRCI, mrrc, MRRC): New.
2100 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2101 VUNSPEC_MRRC2): New.
2103 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2105 * config/arm/arm.md (<mcr>): New.
2107 * config/arm/arm.c (arm_coproc_builtin_available): Add
2108 support for mcr, mrc, mcr2 and mrc2.
2109 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2110 (arm_mcr_qualifiers): ... this. New.
2111 (MRC_QUALIFIERS): Define to ...
2112 (arm_mrc_qualifiers): ... this. New.
2113 (MCR_QUALIFIERS): Define to ...
2114 (arm_mcr_qualifiers): ... this. New.
2115 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2117 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2118 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2119 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2122 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2124 * config/arm/arm.md (*ldc): New.
2128 * config/arm/arm.c (arm_coproc_builtin_available): Add
2129 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2130 (arm_coproc_ldc_stc_legitimate_address): New.
2131 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2132 'qualifier_const_pointer'.
2133 (LDC_QUALIFIERS): Define to...
2134 (arm_ldc_qualifiers): ... this. New.
2135 (STC_QUALIFIERS): Define to...
2136 (arm_stc_qualifiers): ... this. New.
2137 * config/arm/arm-protos.h
2138 (arm_coproc_ldc_stc_legitimate_address): New.
2139 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2140 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2141 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2142 stc2, stcl, stc2l): New.
2143 * config/arm/constraints.md (Uz): New.
2144 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2145 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2146 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2147 VUNSPEC_STC2L): New.
2149 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2151 * config/arm/arm.md (<cdp>): New.
2152 * config/arm/arm.c (neon_const_bounds): Rename this ...
2153 (arm_const_bounds): ... this.
2154 (arm_coproc_builtin_available): New.
2155 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2156 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2157 (CDP_QUALIFIERS): Define to...
2158 (arm_cdp_qualifiers): ... this. New.
2160 (arm_expand_builtin_args): Add case for 6 arguments.
2161 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2162 (arm_const_bounds): ... this.
2163 (arm_coproc_builtin_available): New.
2164 * config/arm/arm_acle.h (__arm_cdp): New.
2166 * config/arm/arm_acle_builtins.def (cdp): New.
2168 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2169 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2171 * config/arm/types.md (coproc): New.
2172 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2173 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2174 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2175 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2177 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2179 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2180 (UBINOP_QUALIFIERS): New.
2182 (acle_builtin_data): New. Change comment.
2183 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2184 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2185 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2186 arm_acle_builtins.def.
2187 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2188 (arm_init_acle_builtins): New.
2189 (CRC32_BUILTIN): Remove.
2190 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2191 crc32cb, crc32ch and crc32cw.
2192 (arm_init_crc32_builtins): Remove.
2193 (arm_init_builtins): Use arm_init_acle_builtins rather
2194 than arm_init_crc32_builtins.
2195 (arm_expand_acle_builtin): New.
2196 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2197 * config/arm/arm_acle_builtins.def: New.
2199 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2201 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2202 (arm_builtin_datum): ... this.
2203 (arm_init_neon_builtin): Rename to ...
2204 (arm_init_builtin): ... this. Add a new parameters PREFIX
2205 and USE_SIG_IN_NAME.
2206 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2207 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2208 'arm_builtin_datum'.
2209 (arm_init_vfp_builtins): Likewise.
2210 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2211 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2212 (arm_expand_neon_args): Rename to ...
2213 (arm_expand_builtin_args): ... this. Rename builtin_arg
2214 enum values and differentiate between ARG_BUILTIN_MEMORY
2215 and ARG_BUILTIN_NEON_MEMORY.
2216 (arm_expand_neon_builtin_1): Rename to ...
2217 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2218 values, arm_expand_builtin_args and add bool parameter NEON.
2219 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2220 (arm_expand_vfp_builtin): Likewise.
2221 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2223 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2226 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2227 * predict.c (tree_estimate_probability_bb): Reverse direction of
2228 polymorphic call predictor.
2230 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2232 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2233 (determine_pass_name_match): ...this new function and...
2234 (should_skip_pass_p): ...this new function.
2236 2017-01-06 Nathan Sidwell <nathan@acm.org>
2238 * ipa-visibility.c (function_and_variable_visibility): Reformat
2239 comments and long lines. Remove extrneous if.
2240 * symtab.c (symtab_node::make_decl_local): Fix code format.
2241 (symtab_node::set_section_for_node): Fix comment typo.
2243 2017-01-06 Martin Liska <mliska@suse.cz>
2246 * lra-constraints.c: Rename invariant to lra_invariant.
2247 * predict.c (set_even_probabilities): Initialize e to NULL.
2249 2017-01-05 Martin Sebor <msebor@redhat.com>
2251 PR tree-optimization/78910
2252 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2253 (format_integer): Correct off-by-one error in the handling
2254 of precision with negative numbers in signed conversions..
2256 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
2258 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
2260 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2262 PR tree-optimization/71016
2263 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
2264 factor_out_conditional_conversion. Formatting fix.
2265 (factor_out_conditional_conversion): Add cond_stmt argument.
2266 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
2267 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
2270 2017-01-05 David Malcolm <dmalcolm@redhat.com>
2272 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
2273 read-rtl-function.o, and selftest-rtl.o.
2274 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
2275 (selftest::aarch64_test_loading_full_dump): New function.
2276 (selftest::aarch64_run_selftests): New function.
2277 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
2278 selftest::aarch64_run_selftests.
2279 * config/i386/i386.c
2280 (selftest::ix86_test_loading_dump_fragment_1): New function.
2281 (selftest::ix86_test_loading_call_insn): New function.
2282 (selftest::ix86_test_loading_full_dump): New function.
2283 (selftest::ix86_test_loading_unspec): New function.
2284 (selftest::ix86_run_selftests): Call the new functions.
2285 * emit-rtl.c (maybe_set_max_label_num): New function.
2286 * emit-rtl.h (maybe_set_max_label_num): New decl.
2287 * function.c (instantiate_decls): Guard call to
2288 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
2289 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
2291 * gensupport.c (gen_reader::gen_reader): Pass "false"
2292 for new "compact" param of rtx_reader.
2293 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
2294 rather than an empty string for NULL strings.
2295 * read-md.c: Potentially include config.h rather than bconfig.h.
2296 Wrap include of errors.h with #ifdef GENERATOR_FILE.
2297 (have_error): New global, copied from errors.c.
2298 (md_reader::read_name): Rename to...
2299 (md_reader::read_name_1): ...this, adding "out_loc" param,
2300 and converting "missing name or number" to returning false, rather
2302 (md_reader::read_name): Reimplement in terms of read_name_1.
2303 (md_reader::read_name_or_nil): New function.
2304 (md_reader::read_string): Handle "(nil)" by returning NULL.
2305 (md_reader::md_reader): Add new param "compact".
2306 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
2307 (md_reader::read_file): New method.
2308 * read-md.h (md_reader::md_reader): Add new param "compact".
2309 (md_reader::read_file): New method.
2310 (md_reader::is_compact): New accessor.
2311 (md_reader::read_name): Convert return type from void to file_location.
2312 (md_reader::read_name_or_nil): New decl.
2313 (md_reader::read_name_1): New decl.
2314 (md_reader::m_compact): New field.
2315 (noop_reader::noop_reader): Pass "false" for new "compact" param
2317 (rtx_reader::rtx_reader): Add new "compact" param.
2318 (rtx_reader::read_rtx_operand): Make virtual and convert return
2319 type from void to rtx.
2320 (rtx_reader::read_until): New decl.
2321 (rtx_reader::handle_any_trailing_information): New virtual function.
2322 (rtx_reader::postprocess): New virtual function.
2323 (rtx_reader::finalize_string): New virtual function.
2324 (rtx_reader::m_in_call_function_usage): New field.
2325 (rtx_reader::m_reuse_rtx_by_id): New field.
2326 * read-rtl-function.c: New file.
2327 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
2328 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
2329 (selftest::verify_three_block_rtl_cfg): New decl.
2330 * read-rtl-function.h: New file.
2331 * read-rtl.c: Potentially include config.h rather than bconfig.h.
2332 For host, include function.h, memmodel.h, and emit-rtl.h.
2333 (one_time_initialization): New function.
2334 (struct compact_insn_name): New struct.
2335 (compact_insn_names): New array.
2336 (find_code): Handle insn codes in compact dumps.
2337 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
2338 (bind_subst_iter_and_attr): Likewise.
2339 (add_condition_to_string): Likewise.
2340 (add_condition_to_rtx): Likewise.
2341 (apply_attribute_uses): Likewise.
2342 (add_current_iterators): Likewise.
2343 (apply_iterators): Likewise.
2344 (initialize_iterators): Guard usage of apply_subst_iterator with
2345 #ifdef GENERATOR_FILE.
2346 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
2347 (md_reader::read_mapping): Likewise.
2348 (add_define_attr_for_define_subst): Likewise.
2349 (add_define_subst_attr): Likewise.
2350 (read_subst_mapping): Likewise.
2351 (check_code_iterator): Likewise.
2352 (rtx_reader::read_rtx): Likewise. Move one-time initialization
2354 (one_time_initialization): New function.
2355 (rtx_reader::read_until): New method.
2356 (read_flags): New function.
2357 (parse_reg_note_name): New function.
2358 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
2359 Handle reuse_rtx ids.
2360 Wrap iterator lookup within #ifdef GENERATOR_FILE.
2361 Add parsing support for RTL dumps, mirroring the special-cases in
2362 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
2363 values, and calling handle_any_trailing_information.
2364 (rtx_reader::read_rtx_operand): Convert return type from void
2365 to rtx, returning return_rtx. Handle case 'e'. Call
2366 finalize_string on XSTR and XTMPL fields.
2367 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
2368 "(nil)" values were omitted. Call the postprocess vfunc on the
2370 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
2371 class ctor. Initialize m_in_call_function_usage. Call
2372 one_time_initialization.
2373 * rtl-tests.c (selftest::test_uncond_jump): Call
2374 set_new_first_and_last_insn.
2375 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
2376 * selftest-rtl.c: New file.
2377 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
2378 (selftest::get_insn_by_uid): New decl.
2379 * selftest-run-tests.c (selftest::run_tests): Call
2380 read_rtl_function_c_tests.
2381 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
2382 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
2385 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
2387 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
2388 operands in a special way. Assert that pos+len <= mode precision.
2390 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2392 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
2393 3 argument Alias with unlimited for the negative form.
2394 (fno-vect-cost-model): Removed.
2396 2017-01-05 Martin Liska <mliska@suse.cz>
2398 * hsa-gen.c (gen_hsa_divmod): New function.
2399 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
2401 2017-01-05 Martin Liska <mliska@suse.cz>
2404 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
2407 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2409 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
2410 small constant length operands.
2412 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2414 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
2415 between loop iterations.
2417 2017-01-05 Martin Liska <mliska@suse.cz>
2420 * gimplify.c (gimplify_decl_expr): Compare to
2421 asan_poisoned_variables instread of checking flags.
2422 (gimplify_target_expr): Likewise.
2423 (gimplify_expr): Likewise.
2424 (gimplify_function_tree): Conditionally initialize
2425 asan_poisoned_variables.
2427 2017-01-04 Jeff Law <law@redhat.com>
2429 PR tree-optimizatin/78812
2430 * rtl.h (contains_mem_rtx_p): Prototype.
2431 * ifcvt.c (containts_mem_rtx_p): Move from here to...
2432 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
2433 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
2434 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
2435 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
2437 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2439 * input.c (assert_char_at_range): Default-initialize actual_range.
2441 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2443 * df-scan.c (df_ref_create_structure): Make regno unsigned,
2444 to match the caller.
2446 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2448 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
2449 insns after final jump in test to emit dummy move.
2451 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2453 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
2454 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
2456 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2458 * multiple_target.c (create_dispatcher_calls): Init e_next.
2459 * tree-ssa-loop-split.c (split_loop): Init border.
2460 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
2463 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
2468 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
2469 (altivec_register_operand): Do not return true if the operand
2470 contains a SUBREG mixing SImode and SFmode.
2471 (vsx_register_operand): Likewise.
2472 (vsx_reg_sfsubreg_ok): New predicate.
2473 (vfloat_operand): Do not return true if the operand contains a
2474 SUBREG mixing SImode and SFmode.
2475 (vint_operand): Likewise.
2476 (vlogical_operand): Likewise.
2477 (gpc_reg_operand): Likewise.
2478 (int_reg_operand): Likewise.
2479 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
2480 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
2481 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
2483 (rs6000_emit_move_si_sf_subreg): New helper function.
2484 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
2485 fixup SUBREGs involving SImode and SFmode.
2486 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
2487 numbers for the new peephole2 optimization.
2488 (peephole2 for SFmode unions): New peephole2 to optimize cases in
2489 the GLIBC math library that do AND/IOR/XOR operations on single
2490 precision floating point.
2491 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
2492 target macros to say whether we need to avoid SUBREGs mixing
2494 (TARGET_ALLOW_SF_SUBREG): Likewise.
2495 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
2496 (UNSPEC_SI_FROM_SF): Likewise.
2497 (iorxor): Change spacing.
2498 (and_ior_xor): New iterator for AND, IOR, and XOR.
2499 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
2500 (movdi_from_sf_zero_ext): Likewise.
2501 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
2502 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
2503 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
2504 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
2505 (fms<mode>4): Likewise.
2506 (fnma<mode>4): Likewise.
2507 (fnms<mode>4): Likewise.
2508 (nfma<mode>4): Likewise.
2509 (nfms<mode>4): Likewise.
2511 2017-01-04 Marek Polacek <polacek@redhat.com>
2514 * doc/invoke.texi: Document -Wpointer-compare.
2516 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2518 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
2521 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
2522 descriptions for -gdwarf-5 and emit them as uleb128 instead of
2525 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
2528 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
2529 documentation of the powerpc_popcntb_ok attribute.
2530 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
2531 code to issue warning messages if a requested CPU configuration is
2532 not supported by the binary (assembler and loader) toolchain.
2533 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
2534 made to define a built-in function that has been disabled.
2535 (paired_init_builtins): Add assertion to prevent ICE if attempt is
2536 made to define a built-in function that has been disabled.
2537 (altivec_init_builtins): Add comment explaining why definition
2538 of the DST built-in functions is not preceded by an assertion
2539 check. Add assertions to prevent ICE if attempts are made to
2540 define an altivec predicate or an abs* built-in function that has
2542 (htm_init_builtins): Add comment explaining why definition of the
2543 htm built-in functions is not preceded by an assertion check.
2545 2017-01-04 Jeff Law <law@redhat.com>
2547 PR tree-optimizatin/67955
2548 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
2549 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
2550 the points-to solution does not include pt_null. Use DECL_PT_UID
2553 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
2555 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
2556 Use gen_int_mode instead of gen_lopwart for const_int operands.
2558 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2560 PR tree-optimization/71563
2561 * match.pd: Simplify X << Y into X if Y is known to be 0 or
2562 out of range value - has low bits known to be zero.
2564 2017-01-04 Alan Modra <amodra@gmail.com>
2566 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
2567 * configure: Regenerate.
2568 * config.in: Regenerate.
2570 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2573 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
2574 a substring of the message, but strcmp with the whole message. Ifdef
2575 ENABLE_NLS, translate the message first using dgettext.
2577 2017-01-03 Jeff Law <law@redhat.com>
2579 PR tree-optimizatin/78856
2580 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
2581 (mark_threaded_blocks): Remove code to truncate thread paths that
2582 cross multiple loop headers. Instead invalidate the cached loop
2583 iteration information and handle case of a thread path walking
2584 into an irreducible region.
2586 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2589 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
2590 assertions. Add support for doing the signbit if the IEEE 128-bit
2591 floating point value is in a GPR.
2592 * config/rs6000/rs6000.md (Fsignbit): Delete.
2593 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
2594 Update the length attribute if the value is in a GPR.
2595 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
2596 the sign or zero extension instruction, since the value is always 0/1.
2597 (signbit<mode>2_dm2): Delete using <Fsignbit>.
2600 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
2601 extracting SImode to a GPR register so that we can generate a
2602 store, limit the vector to be in a traditional Altivec register
2603 for the vextuwrx instruction.
2605 2017-01-03 Ian Lance Taylor <iant@google.com>
2607 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
2609 2017-01-03 Martin Sebor <msebor@redhat.com>
2611 PR tree-optimization/78696
2612 * gimple-ssa-sprintf.c (format_floating): Correct handling of
2613 precision. Use MPFR for %f for greater fidelity. Correct handling
2615 (pass_sprintf_length::compute_format_length): Set width and precision
2616 specified by asrerisk to void_node for vararg functions.
2617 (try_substitute_return_value): Adjust dump output.
2619 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
2621 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
2623 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
2625 * doc/invoke.texi (SPARC options): Document -mlra as the default.
2626 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
2627 -mlra/-mno-lra was passed to the compiler.
2629 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
2631 PR rtl-optimization/65618
2632 * emit-rtl.c (try_split): Move initialization of "before" and
2633 "after" to just before the call to emit_insn_after_setloc.
2635 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
2637 * doc/md.texi (Standard Names): Remove reference to Java frontend.
2639 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
2641 * dwarf2out.c (gen_enumeration_type_die): When
2642 -gno-strict-dwarf, add a DW_AT_encoding attribute.
2644 2017-01-03 Jakub Jelinek <jakub@redhat.com>
2646 PR tree-optimization/78965
2647 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
2648 Change first argument from const call_info & to call_info &. For %n
2649 set info.nowrite to false.
2652 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
2653 possibly throwing calls.
2655 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
2656 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
2657 and fns handling, rather than in a separate case SSA_NAME.
2659 2017-01-02 Jeff Law <law@redhat.com>
2661 * config/darwin-driver.c (darwin_driver_init): Const-correctness
2662 fixes for first_period and second_period variables.
2664 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
2667 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
2668 (*insvqi_1): New insn pattern.
2669 (*insvqi_1_mem_rex64): Ditto.
2671 (*insvqi_3): Rename from *insvqi.
2673 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
2675 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
2677 * doc/cfg.texi (Edges): Remove reference to Java.
2678 (Maintaining the CFG): Ditto.
2680 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2683 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
2684 transparent aliases.
2686 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2689 * predict.def (PRED_CALL): Update hitrate.
2690 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
2691 * predict.c (tree_estimate_probability_bb): Split CALL predictor
2692 into direct/indirect/polymorphic variants.
2694 2017-01-01 Jakub Jelinek <jakub@redhat.com>
2696 Update copyright years.
2698 * gcc.c (process_command): Update copyright notice dates.
2699 * gcov-dump.c (print_version): Ditto.
2700 * gcov.c (print_version): Ditto.
2701 * gcov-tool.c (print_version): Ditto.
2702 * gengtype.c (create_file): Ditto.
2703 * doc/cpp.texi: Bump @copying's copyright year.
2704 * doc/cppinternals.texi: Ditto.
2705 * doc/gcc.texi: Ditto.
2706 * doc/gccint.texi: Ditto.
2707 * doc/gcov.texi: Ditto.
2708 * doc/install.texi: Ditto.
2709 * doc/invoke.texi: Ditto.
2711 Copyright (C) 2017 Free Software Foundation, Inc.
2713 Copying and distribution of this file, with or without modification,
2714 are permitted in any medium without royalty provided the copyright
2715 notice and this notice are preserved.