1 2014-06-09 Alan Lawrence <alan.lawrence@arm.com>
4 * config/arm/arm_neon.h (vtrn_s8, vtrn_s16, vtrn_u8, vtrn_u16, vtrn_p8,
5 vtrn_p16, vtrn_s32, vtrn_f32, vtrn_u32, vtrnq_s8, vtrnq_s16, vtrnq_s32,
6 vtrnq_f32, vtrnq_u8, vtrnq_u16, vtrnq_u32, vtrnq_p8, vtrnq_p16, vzip_s8,
7 vzip_s16, vzip_u8, vzip_u16, vzip_p8, vzip_p16, vzip_s32, vzip_f32,
8 vzip_u32, vzipq_s8, vzipq_s16, vzipq_s32, vzipq_f32, vzipq_u8,
9 vzipq_u16, vzipq_u32, vzipq_p8, vzipq_p16, vuzp_s8, vuzp_s16, vuzp_s32,
10 vuzp_f32, vuzp_u8, vuzp_u16, vuzp_u32, vuzp_p8, vuzp_p16, vuzpq_s8,
11 vuzpq_s16, vuzpq_s32, vuzpq_f32, vuzpq_u8, vuzpq_u16, vuzpq_u32,
12 vuzpq_p8, vuzpq_p16): Correct mask for bigendian.
14 2014-06-09 Jan Hubicka <hubicka@ucw.cz>
16 * tree-core.h (tree_decl_with_vis): Remove section_name.
18 2014-06-09 Kito Cheng <kito@0xlab.org>
20 * ira.c (ira): Don't call init_caller_save if LRA enabled
21 since LRA use its own infrastructure to handle that.
23 2014-06-07 Jan Hubicka <hubicka@ucw.cz>
25 * symtab.c (dump_symtab_base): Update dumping.
26 (symtab_make_decl_local): Clear only DECL_COMDAT.
27 * tree-vect-data-refs.c (Check that variable is static before
28 tampering with sections.
29 * cgraphclones.c (duplicate_thunk_for_node): Do not clear section name.
30 (cgraph_create_virtual_clone): Likewise.
31 * tree.c (decl_comdat_group, decl_comdat_group_id): Constify argument.
32 (decl_section_name, set_decl_section_name): New accessors.
33 (find_decls_types_r): Do not walk section name
34 * tree.h (DECL_SECTION_NAME): Implement using
36 (decl_comdat_group, decl_comdat_group_id): Constify.
37 (decl_section_name, set_decl_section_name): Update.
38 * varpool.c (varpool_finalize_named_section_flags): Use
40 * cgraph.c (cgraph_add_thunk): Reset node instead of rebuilding.
41 (cgraph_make_node_local_1): Clear section and comdat group.
42 * cgraph.h (set_comdat_group): Sanity check.
43 (get_section, set_section): New.
44 * ipa-comdats.c (ipa_comdats): Use get_section.
45 * ipa.c (ipa_discover_readonly_nonaddressable_var): Likewise.
46 * lto-streamer-out.c: Do not follow section names.
47 * c-family/c-common.c (handle_section_attribute):
49 * lto-cgraph.c (lto_output_node): Output section.
50 (lto_output_varpool_node): Likewise.
51 (read_comdat_group): Rename to ...
52 (read_identifier): ... this one.
53 (read_string_cst): New function.
54 (input_node, input_varpool_node): Input section names.
55 * tree-emutls.c (get_emutls_init_templ_addr): Update.
56 (new_emutls_decl): Update.
57 (secname_for_decl): Check section names only of static
59 * config/mep/mep.c (mep_unique_section): Use set_decl_section_name.
60 * config/i386/winnt.c (i386_pe_unique_section): Likewise.
61 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
62 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
63 * config/rs6000/rs6000.c (rs6000_xcoff_unique_section): Likewise.
64 * config/mcore/mcore.c (mcore_unique_section): Likewise.
65 * config/mips/mips.c (mips16_build_function_stub): Likewise.
66 * config/v850/v850.c (v850_insert_attributes): Likewise.
67 * config/h8300/h8300.c: (h8300_handle_eightbit_data_attribute):
69 (h8300_handle_tiny_data_attribute): Likewise.
70 * config/bfin/bfin.c (bfin_handle_l1_text_attribute): Likewise.
71 (bfin_handle_l2_attribute): Likewise.
73 2014-06-07 Jan Hubicka <hubicka@ucw.cz>
75 * symtab.c (symtab_nonoverwritable_alias): Copy VIRTUAL flag;
76 remove static initializer.
78 2014-06-07 Jan Hubicka <hubicka@ucw.cz>
80 * varasm.c (use_blocks_for_decl_p): Check symbol table instead of alias
82 (place_block_symbol): Recurse on aliases.
84 2014-06-07 Jan Hubicka <hubicka@ucw.cz>
86 * ipa-visibility.c: Include varasm.h
87 (can_replace_by_local_alias): Ceck decl_binds_to_current_def_p.
89 2014-06-07 Jan Hubicka <hubicka@ucw.cz>
91 * cgraphunit.c (assemble_thunks_and_aliases): Expand thunks before
94 2014-06-07 Steven Bosscher <steven@gcc.gnu.org>
96 * gcse.c (can_assign_to_reg_without_clobbers_p): Do not let pointers
97 from test_insn into GGC space escape via SET_SRC.
99 2014-06-07 Eric Botcazou <ebotcazou@adacore.com>
101 * tree-ssa-tail-merge.c (same_succ_hash): Hash the static chain of a
102 call statement, if any.
103 (gimple_equal_p) <GIMPLE_CALL>: Compare the static chain of the call
104 statements, if any. Tidy up.
106 2014-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
109 * config/rs6000/vsx.md (VSX_LE): Split VSX_D into 2 separate
110 iterators, VSX_D that handles 64-bit types, and VSX_LE that
111 handles swapping the two 64-bit double words on little endian
112 systems. Include V1TImode and optionally TImode in VSX_LE so that
113 these types are properly swapped. Change all of the insns and
114 splits that do the 64-bit swaps to use VSX_LE.
115 (vsx_le_perm_load_<mode>): Likewise.
116 (vsx_le_perm_store_<mode>): Likewise.
117 (splitters for little endian memory operations): Likewise.
118 (vsx_xxpermdi2_le_<mode>): Likewise.
119 (vsx_lxvd2x2_le_<mode>): Likewise.
120 (vsx_stxvd2x2_le_<mode>): Likewise.
122 2014-06-06 Uros Bizjak <ubizjak@gmail.com>
125 * config/i386/i386.md (*floatunssi<mode>2_i387_with_xmm): New
126 define_insn_and_split pattern, merged from *floatunssi<mode>2_1
127 and corresponding splitters. Zero extend general register
128 or memory input operand to XMM temporary. Enable for
129 TARGET_SSE2 and TARGET_INTER_UNIT_MOVES_TO_VEC only.
130 (floatunssi<mode>2): Update expander predicate.
132 2014-06-06 Vladimir Makarov <vmakarov@redhat.com>
134 PR rtl-optimization/61325
135 * lra-constraints.c (process_address_1): Check scale equal to one
136 to prevent transformation: base + scale * index => base + new_reg.
138 2014-06-06 Richard Biener <rguenther@suse.de>
140 PR tree-optimization/59299
141 * tree-ssa-sink.c (all_immediate_uses_same_place): Work on
143 (nearest_common_dominator_of_uses): Likewise.
144 (statement_sink_location): Adjust. Support sinking loads.
146 2014-06-06 Martin Jambor <mjambor@suse.cz>
148 * ipa-prop.c (get_place_in_agg_contents_list): New function.
149 (build_agg_jump_func_from_list): Likewise.
150 (determine_known_aggregate_parts): Renamed to
151 determine_locally_known_aggregate_parts. Moved some functionality
152 to the two functions above, removed bound checks.
154 2014-06-06 James Greenhalgh <james.greenhalgh@arm.com>
156 * config/aarch64/aarch64-protos.h (aarch64_expand_movmem): New.
157 * config/aarch64/aarch64.c (aarch64_move_pointer): New.
158 (aarch64_progress_pointer): Likewise.
159 (aarch64_copy_one_part_and_move_pointers): Likewise.
160 (aarch64_expand_movmen): Likewise.
161 * config/aarch64/aarch64.h (MOVE_RATIO): Set low.
162 * config/aarch64/aarch64.md (movmem<mode>): New.
164 2014-06-06 Bingfeng Mei <bmei@broadcom.com>
166 * targhooks.c (default_add_stmt_cost): Call target specific
167 hook instead of default one.
169 2014-06-06 Thomas Preud'homme <thomas.preudhomme@arm.com>
171 * ChangeLog (2014-05-23): Fix ChangeLog entry to refer to target
172 endianness instead of host endianness.
173 * tree-ssa-math-opts.c (find_bswap_or_nop_1): Likewise in dumps and
176 2014-06-06 Eric Botcazou <ebotcazou@adacore.com>
179 * function.c (instantiate_decls): Process the saved static chain.
180 (expand_function_start): If not optimizing, save the static chain
182 * tree-nested.c (convert_all_function_calls): Always create the static
183 chain for nested functions if not optimizing.
185 2014-06-06 Eric Botcazou <ebotcazou@adacore.com>
187 * tree-cfg.c (make_edges) <GIMPLE_RETURN>: Put a location on the edge.
189 2014-06-06 Richard Biener <rguenther@suse.de>
191 * cfgexpand.c (expand_gimple_cond): Remove check for current_loops.
192 (construct_init_block): Likewise.
193 (construct_exit_block): Likewise.
194 (pass_expand::execute): Likewise.
195 * graphite.c (graphite_transforms): Replace check for current_loops
196 with a check for > 1 loops.
197 (pass_graphite_transforms::execute): Adjust.
198 * ipa-split.c (split_function): Remove check for current_loops.
199 * omp-low.c (expand_parallel_call): Likewise.
200 (expand_omp_for_init_counts): Likewise.
201 (extract_omp_for_update_vars): Likewise.
202 (expand_omp_for_generic): Likewise.
203 (expand_omp_sections): Likewise.
204 (expand_omp_target): Likewise.
205 * tracer.c (tail_duplicate): Likewise.
206 (pass_tracer::execute): Likewise.
207 * trans-mem.c (expand_transaction): Likewise.
208 * tree-complex.c (expand_complex_div_wide): Likewise.
209 * tree-eh.c (lower_resx): Likewise.
210 (cleanup_empty_eh_merge_phis): Likewise.
211 * tree-predcom.c (run_tree_predictive_commoning): Replace check for
212 current_loops with a check for > 1 loops.
213 (pass_predcom::execute): Adjust.
214 * tree-scalar-evolution.c (scev_reset): Remove check for current_loops.
215 * tree-ssa-copy.c (copy_prop_visit_phi_node): Likewise.
216 * tree-ssa-dom.c (pass_phi_only_cprop::execute): Likewise.
217 * tree-ssa-tail-merge.c (tail_merge_optimize): Likewise.
218 * tree-ssa-threadupdate.c (thread_through_all_blocks): Likewise.
219 * tree-switch-conversion.c (process_switch): Likewise.
220 * tree-tailcall.c (tree_optimize_tail_calls_1): Likewise.
221 * tree-vrp.c (vrp_visit_phi_node): Likewise.
222 (execute_vrp): Likewise.
223 * ubsan.c (ubsan_expand_null_ifn): Likewise.
225 2014-06-06 Eric Botcazou <ebotcazou@adacore.com>
227 * rtl.h (insn_location): Declare.
228 * cfgcleanup.c (try_forward_edges): Compare the locus of locations
229 with UNKNOWN_LOCATION.
230 * emit-rtl.c (insn_location): New function.
231 * final.c (notice_source_line): Check that the instruction has a
232 location before retrieving it and use insn_location.
233 * modulo-sched.c (loop_single_full_bb_p): Likewise.
234 * print-rtl.c (print_rtx): Likewise.
236 2014-06-06 Richard Biener <rguenther@suse.de>
238 * passes.def: Move 2nd VRP pass before phi-only-cprop.
240 2014-06-06 Christian Bruel <christian.bruel@st.com>
242 PR tree-optimization/43934
243 * tree-ssa-loop-im.c (determine_max_movement): Add PHI def constant
246 2014-06-06 Richard Sandiford <rdsandiford@googlemail.com>
248 * ira-lives.c (single_reg_class): Add missing break. Explicitly
249 return NO_REGS for extra address and memory constraints. Handle
250 operands that match (or are equivalent to something that matches)
251 extra constant constraints. Ignore other non-register operands.
253 2014-06-06 Alan Modra <amodra@gmail.com>
256 * doc/tm.texi.in (INCOMING_REG_PARM_STACK_SPACE): Document.
257 * doc/tm.texi: Regenerate.
258 * function.c (INCOMING_REG_PARM_STACK_SPACE): Provide default.
259 Use throughout in place of REG_PARM_STACK_SPACE.
260 * config/rs6000/rs6000.c (rs6000_reg_parm_stack_space): Add
261 "incoming" param. Pass to rs6000_function_parms_need_stack.
262 (rs6000_function_parms_need_stack): Add "incoming" param, ignore
263 prototype_p when incoming. Use function decl when incoming
264 to handle K&R style functions.
265 * config/rs6000/rs6000.h (REG_PARM_STACK_SPACE): Adjust.
266 (INCOMING_REG_PARM_STACK_SPACE): Define.
268 2014-06-05 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
271 * cfgexpand.c (expand_debug_expr): Use address space of nested
272 TREE_TYPE for ADDR_EXPR and MEM_REF.
274 2014-06-05 Jeff Law <law@redhat.com>
276 PR tree-optimization/61289
277 * tree-ssa-threadedge.c (invalidate_equivalences): Remove SRC_MAP and
278 DST_MAP parameters. Invalidate by walking all the SSA_NAME_VALUES
279 looking for those which match LHS. All callers changed.
280 (record_temporary_equivalences_from_phis): Remove SRC_MAP and DST_MAP
281 parameters and code which manipulated them. All callers changed.
282 (record_temporary_equivalences_from_stmts_at_dest): Remove SRC_MAP
283 and DST_MAP parameters. Simplify invalidation code by just calling
284 invalidate_equivalences. All callers changed.
285 (thread_across_edge): Simplify now that we don't need to maintain
286 the map of equivalences to invalidate.
288 2014-06-05 Kai Tietz <ktietz@redhat.com>
289 Richard Henderson <rth@redhat.com>
292 * config/i386/predicates.md (memory_nox32_operand): Add memory_operand
293 checking for !TARGET_X32.
294 * config/i386/i386.md (UNSPEC_PEEPSIB): New unspec constant.
295 (sibcall_intern): New define_insn, plus required peepholes.
296 (sibcall_pop_intern): Likewise.
297 (sibcall_value_intern): Likewise.
298 (sibcall_value_pop_intern): Likewise.
300 2014-06-05 Ilya Enkovich <ilya.enkovich@intel.com>
302 * tree-inline.c (tree_function_versioning): Check DF info existence
305 2014-06-05 Marcus Shawcroft <marcus.shawcroft@arm.com>
307 * config/aarch64/aarch64.h (aarch64_frame): Add hard_fp_offset and
309 * config/aarch64/aarch64.c (aarch64_layout_frame): Initialize
310 aarch64_frame hard_fp_offset and frame_size.
311 (aarch64_expand_prologue): Use aarch64_frame hard_fp_offset and
312 frame_size; remove original_frame_size.
313 (aarch64_expand_epilogue, aarch64_final_eh_return_addr): Likewise.
314 (aarch64_initial_elimination_offset): Remove frame_size and
315 offset. Use aarch64_frame frame_size.
317 2014-06-05 Marcus Shawcroft <marcus.shawcroft@arm.com>
318 Jiong Wang <jiong.wang@arm.com>
319 Renlin <renlin.li@arm.com>
321 * config/aarch64/aarch64.c (aarch64_layout_frame): Correct
322 initialization of R30 offset. Update offset. Iterate core
323 regisers upto X30. Remove X29, X30 specific code.
325 2014-06-05 Marcus Shawcroft <marcus.shawcroft@arm.com>
326 Jiong Wang <jiong.wang@arm.com>
328 * config/aarch64/aarch64.c (SLOT_NOT_REQUIRED, SLOT_REQUIRED): Define.
329 (aarch64_layout_frame): Use SLOT_NOT_REQUIRED and SLOT_REQUIRED.
330 (aarch64_register_saved_on_entry): Adjust test.
332 2014-06-05 Marcus Shawcroft <marcus.shawcroft@arm.com>
334 * config/aarch64/aarch64.h (machine_function): Move
335 saved_varargs_size from here...
336 (aarch64_frame): ... to here.
338 * config/aarch64/aarch64.c (aarch64_expand_prologue)
339 (aarch64_expand_epilogue, aarch64_final_eh_return_addr)
340 (aarch64_initial_elimination_offset)
341 (aarch64_setup_incoming_varargs): Adjust location of
344 2014-06-05 Marcus Shawcroft <marcus.shawcroft@arm.com>
346 * config/aarch64/aarch64.c (aarch64_expand_prologue): Update stack
349 2014-06-05 Jaydeep Patil <Jaydeep.Patil@imgtec.com>
350 Prachi Godbole <Prachi.Godbole@imgtec.com>
352 * config/mips/mips-cpus.def: Add definition for p5600. Updated
353 mips32r5 entry to use PROCESSOR_P5600.
354 * config/mips/mips-tables.opt: Regenerate.
355 * config/mips/mips-protos.h (mips_fmadd_bypass): Add prototype.
356 * config/mips/mips.c (mips_fmadd_bypass): New function.
357 (mips_rtx_cost_data): Add costs for p5600.
358 (mips_issue_rate): Add support for p5600.
359 (mips_multipass_dfa_lookahead): Likewise.
360 * config/mips/mips.h (TUNE_P5600): New define.
361 (TUNE_MACC_CHAINS): Add TUNE_P5600.
362 (MIPS_ISA_LEVEL_SPEC): Map -march=p5600 to -mips32r5.
363 * config/mips/mips.md: Include p5600.md.
364 (processor): Add p5600.
365 * config/mips/p5600.md: New file.
367 2014-06-05 Evgeny Stupachenko <evstupac@gmail.com>
369 * config/i386/sse.md (*ssse3_palignr<mode>_perm): New.
370 * config/i386/predicates.md (palignr_operand): New.
371 Indicates if permutation is suitable for palignr instruction.
373 2014-06-05 Yuri Rumyantsev <ysrumyan@gmail.com>
375 PR tree-optimization/61319
376 * tree-if-conv.c (is_cond_scalar_reduction): Add missed check that
377 stmt belongs to loop.
379 2014-06-05 Richard Biener <rguenther@suse.de>
381 * gimplify.c (create_tmp_from_val): Remove is_formal parameter
382 and set DECL_GIMPLE_REG_P unconditionally if appropriate.
383 (lookup_tmp_var): Adjust.
384 (prepare_gimple_addressable): Unset DECL_GIMPLE_REG_P here.
386 2014-06-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
388 * config/arm/arm.md (enabled): Disable opt_enabled attribute.
390 2014-06-05 Marek Polacek <polacek@redhat.com>
393 * doc/invoke.texi: Document -Wlogical-not-parentheses.
395 2014-06-04 Tom de Vries <tom@codesourcery.com>
397 * config/s390/s390.md ("addptrdi3", "addptrsi3"): Use INTVAL only on
400 2014-06-04 Marc Glisse <marc.glisse@inria.fr>
402 PR tree-optimization/61385
403 * tree-ssa-phiopt.c (value_replacement): Punt if there are PHI nodes.
405 2014-06-04 Bernd Schmidt <bernds@codesourcery.com>
407 * lto-wrapper.c (fatal, fatal_perror): Remove functions. All callers
408 changed to use fatal_error.
409 (main): Ensure lto_wrapper_cleanup is run atexit.
411 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
413 * lra-constraints.c (valid_address_p): Move earlier in file.
414 (address_eliminator): New structure.
415 (satisfies_memory_constraint_p): New function.
416 (satisfies_address_constraint_p): Likewise.
417 (process_alt_operands, process_address, curr_insn_transform): Use them.
419 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
421 * lra-int.h (lra_static_insn_data): Make operand_alternative a
423 (target_lra_int, default_target_lra_int, this_target_lra_int)
424 (op_alt_data): Delete.
425 * lra.h (lra_init): Delete.
426 * lra.c (default_target_lra_int, this_target_lra_int): Delete.
427 (init_insn_code_data_once): Remove op_alt_data handling.
428 (finish_insn_code_data_once): Likewise.
429 (init_op_alt_data): Delete.
430 (get_static_insn_data): Initialize operand_alternative to null.
431 (free_insn_recog_data): Cast operand_alternative before freeing it.
432 (setup_operand_alternative): Take the operand_alternative as
433 parameter and assume it isn't already cached in the static
435 (lra_set_insn_recog_data): Update accordingly.
437 * ira.c (ira_init): Don't call lra_init.
438 * target-globals.h (this_target_lra_int): Declare.
439 (target_globals): Remove lra_int.
440 (restore_target_globals): Update accordingly.
441 * target-globals.c: Don't include lra-int.h.
442 (default_target_globals, save_target_globals): Remove lra_int.
444 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
446 * recog.h (operand_alternative): Convert reg_class, reject,
447 matched and matches into bitfields.
448 (preprocess_constraints): New overload.
449 (preprocess_insn_constraints): New function.
450 (preprocess_constraints): Take the insn as parameter.
451 (recog_op_alt): Change into a pointer.
452 (target_recog): Add x_op_alt.
453 * recog.c (asm_op_alt): New variable.
454 (recog_op_alt): Change into a pointer.
455 (preprocess_constraints): New overload, replacing the old function
456 definition with one that doesn't use global state.
457 (preprocess_insn_constraints): New function.
458 (preprocess_constraints): Use them. Take the insn as parameter.
459 Use asm_op_alt for asms.
460 (recog_init): Free existing x_op_alt entries.
461 * ira-lives.c (check_and_make_def_conflict): Make operand_alternative
463 (make_early_clobber_and_input_conflicts): Likewise.
464 (process_bb_node_lives): Pass the insn to process_constraints.
465 * reg-stack.c (check_asm_stack_operands): Likewise.
466 (subst_asm_stack_regs): Likewise.
467 * regcprop.c (copyprop_hardreg_forward_1): Likewise.
468 * regrename.c (build_def_use): Likewise.
469 * sched-deps.c (sched_analyze_insn): Likewise.
470 * sel-sched.c (get_reg_class, implicit_clobber_conflict_p): Likewise.
471 * config/arm/arm.c (xscale_sched_adjust_cost): Likewise.
472 (note_invalid_constants): Likewise.
473 * config/i386/i386.c (ix86_legitimate_combined_insn): Likewise.
474 (ix86_legitimate_combined_insn): Make operand_alternative pointer
477 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
479 * recog.c (preprocess_constraints): Don't skip disabled alternatives.
480 * ira-lives.c (check_and_make_def_conflict): Check for disabled
482 (make_early_clobber_and_input_conflicts): Likewise.
483 * config/i386/i386.c (ix86_legitimate_combined_insn): Likewise.
485 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
487 * recog.h (alternative_class): New function.
488 (which_op_alt): Return a const recog_op_alt.
489 * reg-stack.c (check_asm_stack_operands): Update type accordingly.
490 (subst_asm_stack_regs): Likewise.
491 * config/arm/arm.c (note_invalid_constants): Likewise.
492 * regcprop.c (copyprop_hardreg_forward_1): Likewise. Don't modify
493 the operand_alternative; use alternative class instead.
494 * sel-sched.c (get_reg_class): Likewise.
495 * regrename.c (build_def_use): Likewise.
496 (hide_operands, restore_operands, record_out_operands): Update type
499 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
501 * recog.h (recog_op_alt): Convert to a flat array.
502 (which_op_alt): New function.
503 * recog.c (recog_op_alt): Convert to a flat array.
504 (preprocess_constraints): Update accordingly, grouping all
505 operands of the same alternative together, rather than the
507 * ira-lives.c (check_and_make_def_conflict): Likewise.
508 (make_early_clobber_and_input_conflicts): Likewise.
509 * config/i386/i386.c (ix86_legitimate_combined_insn): Likewise.
510 * reg-stack.c (check_asm_stack_operands): Use which_op_alt.
511 (subst_asm_stack_regs): Likewise.
512 * regcprop.c (copyprop_hardreg_forward_1): Likewise.
513 * regrename.c (hide_operands, record_out_operands): Likewise.
514 (build_def_use): Likewise.
515 * sel-sched.c (get_reg_class): Likewise.
516 * config/arm/arm.c (note_invalid_constants): Likewise.
518 2014-06-04 Jason Merrill <jason@redhat.com>
522 * gimplify.c (gimplify_arg): Non-static.
523 * gimplify.h: Declare it.
525 2014-06-04 Richard Biener <rguenther@suse.de>
527 * tree.h (may_be_aliased): Trust TREE_ADDRESSABLE from
528 TREE_PUBLIC and DECL_EXTERNAL decls.
530 2014-06-04 Matthew Fortune <matthew.fortune@imgtec.com>
532 * regcprop.c (copyprop_hardreg_forward_1): Account for
533 HARD_REGNO_CALL_PART_CLOBBERED.
535 2014-06-04 Richard Biener <rguenther@suse.de>
537 * configure.ac: Check whether the underlying type of int64_t
538 is long or long long.
539 * configure: Regenerate.
540 * config.in: Likewise.
541 * hwint.h (HOST_WIDE_INT): Match the underlying type of int64_t.
542 (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
544 2014-06-04 Richard Biener <rguenther@suse.de>
546 PR tree-optimization/60098
547 * tree-ssa-dse.c (dse_possible_dead_store_p): Walk until
549 (dse_optimize_stmt): Simplify, now that we found a kill
552 2014-06-04 Richard Biener <rguenther@suse.de>
554 * tree-ssa-alias.c (stmt_may_clobber_ref_p): Improve handling
555 of accesses with non-invariant address.
557 2014-06-04 Martin Liska <mliska@suse.cz>
559 * cgraph.h (cgraph_make_wrapper): New function introduced.
560 * cgraphunit.c (cgraph_make_wrapper): The function implementation.
561 * ipa-inline.h (inline_analyze_function): The function is global.
562 * ipa-inline-analysis.c (inline_analyze_function): Likewise.
564 2014-06-04 Martin Liska <mliska@suse.cz>
566 * tree.h (private_lookup_attribute_starting): New function.
567 (lookup_attribute_starting): Likewise.
568 * tree.c (private_lookup_attribute_starting): Likewise.
570 2014-06-04 Martin Liska <mliska@suse.cz>
572 * cgraph.h (expand_thunk): New argument added.
573 (address_taken_from_non_vtable_p): New global function.
574 * ipa-visibility.c (address_taken_from_non_vtable_p): Likewise.
575 * cgraphclones.c (duplicate_thunk_for_node): Argument added to call.
576 * cgraphunit.c (analyze_function): Likewise.
577 (assemble_thunks_and_aliases): Argument added to call.
578 (expand_thunk): New argument forces to produce GIMPLE thunk.
580 2014-06-04 Martin Liska <mliska@suse.cz>
582 * coverage.h (coverage_compute_cfg_checksum): Argument added.
583 * coverage.c (coverage_compute_cfg_checksum): Likewise.
584 * profile.c (branch_prob): Likewise.
586 2014-06-04 Martin Jambor <mjambor@suse.cz>
589 * ipa-pure-const.c (propagate_pure_const): Add unreachable default
590 handler for switch on an ipa_ref_use enum.
591 * ipa-reference.c (analyze_function): Likewise.
593 2014-06-04 Kai Tietz <ktietz@redhat.com>
595 * recog.c (peep2_attempt): Copy SIBLING_CALL_P flag
596 from old call-instruction.
598 2014-06-04 Bin Cheng <bin.cheng@arm.com>
600 * config/aarch64/aarch64.c (aarch64_classify_address)
601 (aarch64_legitimize_reload_address): Support full addressing modes
603 * config/aarch64/aarch64.md (mov<mode>, movmisalign<mode>)
604 (*aarch64_simd_mov<mode>, *aarch64_simd_mov<mode>): Relax predicates.
606 2014-06-03 Andrew Pinski <apinski@cavium.com>
608 * config/aarch64/aarch64.c (aarch64_if_then_else_costs): Allow non comparisons
611 2014-06-03 Andrew Pinski <apinski@cavium.com>
613 * config/aarch64/aarch64.c (aarch64_if_then_else_costs): New function.
614 (aarch64_rtx_costs): Use aarch64_if_then_else_costs.
616 2014-06-03 Kai Tietz <ktietz@redhat.com>
618 * config/i386/i386.c (ix86_function_value_regno_p): Disallow DX_REG
621 2014-06-03 Dehao Chen <dehao@google.com>
623 * tree-cfg.c (gimple_merge_blocks): Only reset count when BBs are in
626 2014-06-03 Marek Polacek <polacek@redhat.com>
629 * doc/invoke.texi: Document -Wswitch-bool.
630 * function.c (stack_protect_epilogue): Cast controlling expression of
632 * gengtype.c (walk_type): Generate switch expression with its
633 controlling expression cast to int.
635 2014-06-03 Vishnu K S <Vishnu.k_s@atmel.com>
637 * config/avr/avr-mcus.def: Add new avr25 devices attiny441, attiny828
639 * config/avr/avr-tables.opt: Regenerate.
640 * config/avr/t-multilib: Regenerate.
641 * doc/avr-mmcu.texi: Regenerate.
643 2014-06-03 Vishnu K S <vishnu.k_s@atmel.com>
644 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
646 * config/avr/avr-mcus.def (ata6616c): Add new avr25 device.
647 (ata6617c, ata664251): Add new avr35 devices.
648 (ata6612c): Add new avr4 device.
649 (ata6613c, ata6614q): Add new avr5 devices.
650 * config/avr/avr-tables.opt: Regenerate.
651 * config/avr/t-multilib: Regenerate.
652 * doc/avr-mmcu.texi: Regenerate.
654 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
656 * gcc/config/aarch64/aarch64-builtins.c
657 (aarch64_types_binop_ssu_qualifiers): New static data.
658 (TYPES_BINOP_SSU): Define.
659 * gcc/config/aarch64/aarch64-simd-builtins.def (suqadd, ushl, urshl,
660 urshr_n, ushll_n): Use appropriate unsigned qualifiers.
661 * gcc/config/aarch64/arm_neon.h (vrshl_u8, vrshl_u16, vrshl_u32,
662 vrshl_u64, vrshlq_u8, vrshlq_u16, vrshlq_u32, vrshlq_u64, vrshld_u64,
663 vrshr_n_u8, vrshr_n_u16, vrshr_n_u32, vrshr_n_u64, vrshrq_n_u8,
664 vrshrq_n_u16, vrshrq_n_u32, vrshrq_n_u64, vrshrd_n_u64, vshll_n_u8,
665 vshll_n_u16, vshll_n_u32, vuqadd_s8, vuqadd_s16, vuqadd_s32,
666 vuqadd_s64, vuqaddq_s8, vuqaddq_s16, vuqaddq_s32, vuqaddq_s64,
667 vuqaddb_s8, vuqaddh_s16, vuqadds_s32, vuqaddd_s64): Add signedness
668 suffix to builtin function name, remove cast.
669 (vshl_s8, vshl_s16, vshl_s32, vshl_s64, vshl_u8, vshl_u16, vshl_u32,
670 vshl_u64, vshlq_s8, vshlq_s16, vshlq_s32, vshlq_s64, vshlq_u8,
671 vshlq_u16, vshlq_u32, vshlq_u64, vshld_s64, vshld_u64): Remove cast.
673 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
675 * gcc/config/aarch64/aarch64-builtins.c
676 (aarch64_types_binop_uus_qualifiers,
677 aarch64_types_shift_to_unsigned_qualifiers,
678 aarch64_types_unsigned_shiftacc_qualifiers): Define.
679 * gcc/config/aarch64/aarch64-simd-builtins.def (uqshl, uqrshl, uqadd,
680 uqsub, usqadd, usra_n, ursra_n, uqshrn_n, uqrshrn_n, usri_n, usli_n,
681 sqshlu_n, uqshl_n): Update qualifiers.
682 * gcc/config/aarch64/arm_neon.h (vqadd_u8, vqadd_u16, vqadd_u32,
683 vqadd_u64, vqaddq_u8, vqaddq_u16, vqaddq_u32, vqaddq_u64, vqsub_u8,
684 vqsub_u16, vqsub_u32, vqsub_u64, vqsubq_u8, vqsubq_u16, vqsubq_u32,
685 vqsubq_u64, vqaddb_u8, vqaddh_u16, vqadds_u32, vqaddd_u64, vqrshl_u8,
686 vqrshl_u16, vqrshl_u32, vqrshl_u64, vqrshlq_u8, vqrshlq_u16,
687 vqrshlq_u32, vqrshlq_u64, vqrshlb_u8, vqrshlh_u16, vqrshls_u32,
688 vqrshld_u64, vqrshrn_n_u16, vqrshrn_n_u32, vqrshrn_n_u64,
689 vqrshrnh_n_u16, vqrshrns_n_u32, vqrshrnd_n_u64, vqshl_u8, vqshl_u16,
690 vqshl_u32, vqshl_u64, vqshlq_u8, vqshlq_u16, vqshlq_u32, vqshlq_u64,
691 vqshlb_u8, vqshlh_u16, vqshls_u32, vqshld_u64, vqshl_n_u8, vqshl_n_u16,
692 vqshl_n_u32, vqshl_n_u64, vqshlq_n_u8, vqshlq_n_u16, vqshlq_n_u32,
693 vqshlq_n_u64, vqshlb_n_u8, vqshlh_n_u16, vqshls_n_u32, vqshld_n_u64,
694 vqshlu_n_s8, vqshlu_n_s16, vqshlu_n_s32, vqshlu_n_s64, vqshluq_n_s8,
695 vqshluq_n_s16, vqshluq_n_s32, vqshluq_n_s64, vqshlub_n_s8,
696 vqshluh_n_s16, vqshlus_n_s32, vqshlud_n_s64, vqshrn_n_u16,
697 vqshrn_n_u32, vqshrn_n_u64, vqshrnh_n_u16, vqshrns_n_u32,
698 vqshrnd_n_u64, vqsubb_u8, vqsubh_u16, vqsubs_u32, vqsubd_u64,
699 vrsra_n_u8, vrsra_n_u16, vrsra_n_u32, vrsra_n_u64, vrsraq_n_u8,
700 vrsraq_n_u16, vrsraq_n_u32, vrsraq_n_u64, vrsrad_n_u64, vsli_n_u8,
701 vsli_n_u16, vsli_n_u32,vsli_n_u64, vsliq_n_u8, vsliq_n_u16,
702 vsliq_n_u32, vsliq_n_u64, vslid_n_u64, vsqadd_u8, vsqadd_u16,
703 vsqadd_u32, vsqadd_u64, vsqaddq_u8, vsqaddq_u16, vsqaddq_u32,
704 vsqaddq_u64, vsqaddb_u8, vsqaddh_u16, vsqadds_u32, vsqaddd_u64,
705 vsra_n_u8, vsra_n_u16, vsra_n_u32, vsra_n_u64, vsraq_n_u8,
706 vsraq_n_u16, vsraq_n_u32, vsraq_n_u64, vsrad_n_u64, vsri_n_u8,
707 vsri_n_u16, vsri_n_u32, vsri_n_u64, vsriq_n_u8, vsriq_n_u16,
708 vsriq_n_u32, vsriq_n_u64, vsrid_n_u64): Remove casts.
710 2014-06-03 Teresa Johnson <tejohnson@google.com>
712 * tree-sra.c (modify_function): Record caller nodes after rebuild.
714 2014-06-02 Jason Merrill <jason@redhat.com>
717 * varpool.c (ctor_for_folding): Handle uninitialized vtables.
719 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
721 * config/aarch64/aarch64.c (aarch64_evpc_ext): allow and handle
724 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
726 * config/aarch64/aarch64-simd.md (aarch64_rev<REVERSE:rev-op><mode>):
728 * config/aarch64/aarch64.c (aarch64_evpc_rev): New function.
729 (aarch64_expand_vec_perm_const_1): Add call to aarch64_evpc_rev.
730 * config/aarch64/iterators.md (REVERSE): New iterator.
731 (UNSPEC_REV64, UNSPEC_REV32, UNSPEC_REV16): New enum elements.
732 (rev_op): New int_attribute.
733 * config/aarch64/arm_neon.h (vrev16_p8, vrev16_s8, vrev16_u8,
734 vrev16q_p8, vrev16q_s8, vrev16q_u8, vrev32_p8, vrev32_p16, vrev32_s8,
735 vrev32_s16, vrev32_u8, vrev32_u16, vrev32q_p8, vrev32q_p16, vrev32q_s8,
736 vrev32q_s16, vrev32q_u8, vrev32q_u16, vrev64_f32, vrev64_p8,
737 vrev64_p16, vrev64_s8, vrev64_s16, vrev64_s32, vrev64_u8, vrev64_u16,
738 vrev64_u32, vrev64q_f32, vrev64q_p8, vrev64q_p16, vrev64q_s8,
739 vrev64q_s16, vrev64q_s32, vrev64q_u8, vrev64q_u16, vrev64q_u32):
740 Replace temporary __asm__ with __builtin_shuffle.
742 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
744 * config/mips/mips-cpus.def: Add mips32r3, mips32r5, mips64r3 and
746 * config/mips/mips-tables.opt: Regenerate.
747 * config/mips/mips.c (mips_compute_frame_info): Changed if statement
748 to use mips_isa_rev rather than ISA_MIPS32R2.
749 * config/mips/mips.h (ISA_MIPS32R3): New define.
750 (ISA_MIPS32R5): New define.
751 (ISA_MIPS64R3): New define.
752 (ISA_MIPS64R5): New define.
753 (TARGET_CPU_CPP_BUILTINS): Added support for ISA_MIPS32R3,
754 ISA_MIPS32R5, ISA_MIPS64R3 and ISA_MIPS64R5.
755 (MIPS_ISA_LEVEL_SPEC): Added support for mips32r3, mips32r5, mips64r3
757 (MIPS_ISA_SYNCI_SPEC): Likewise.
758 (ISA_HAS_64BIT_REGS): Added ISA_MIPS64R3 and ISA_MIPS64R5.
759 (LINK_SPEC): Added mips32r3 and mips32r5.
760 * config/mips/t-isa3264 (MULTILIB_MATCHES): Map mips32r3 and mips32r5
761 to mips32r2; and mips64r3 and mips64r5 to mips64r2.
762 * config/mips/t-mti-elf (MULTILIB_MATCHES): Likewise.
763 * config/mips/t-mti-linux (MULTILIB_MATCHES): Likewise.
764 * config/mips/t-sde (MULTILIB_MATCHES): Likewise.
765 * config/mips/t-sdemtk (MULTILIB_MATCHES): New define.
766 * doc/invoke.texi: Document mips32r3, mips32r5, mips64r3 and mips64r5.
768 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
770 * doc/invoke.texi: Document -mxpa and -mno-xpa MIPS command line
772 * config/mips/mips.opt (mxpa): New option.
773 * config/mips/mips.h (ASM_SPEC): Pass mxpa and mno-xpa to the
776 2014-06-03 Martin Jambor <mjambor@suse.cz>
779 * ipa-cp.c (cgraph_edge_brings_value_p): Handle edges leading to
782 2014-06-03 Thomas Preud'homme <thomas.preudhomme@arm.com>
784 PR tree-optimization/61328
785 * tree-ssa-math-opts.c (init_symbolic_number): Extract symbolic number
786 initialization from find_bswap_or_nop_1.
787 (find_bswap_or_nop_1): Test return value of find_bswap_or_nop_1 stored
788 in source_expr2 before using the size value the function sets. Also
789 make use of init_symbolic_number () in both the old place and
790 find_bswap_or_nop_load () to avoid reading uninitialized memory when
791 doing recursion in the GIMPLE_BINARY_RHS case.
793 2014-06-03 Richard Biener <rguenther@suse.de>
795 PR tree-optimization/61383
796 * tree-ssa-ifcombine.c (bb_no_side_effects_p): Make sure
799 2014-06-03 Richard Sandiford <rdsandiford@googlemail.com>
801 * defaults.h (USE_MD_CONSTRAINTS, EXTRA_MEMORY_CONSTRAINT)
802 (EXTRA_ADDRESS_CONSTRAINT, DEFAULT_CONSTRAINT_LEN, CONSTRAINT_LEN)
803 (CONST_OK_FOR_CONSTRAINT_P, CONST_DOUBLE_OK_FOR_LETTER_P)
804 (REG_CLASS_FROM_CONSTRAINT, EXTRA_CONSTRAINT_STR): Delete definitions
806 (REG_CLASS_FROM_LETTER, CONST_OK_FOR_LETTER_P)
807 (CONST_DOUBLE_OK_FOR_LETTER_P, EXTRA_CONSTRAINT): Move poising to...
808 * system.h: ...here and make it unconditional.
809 * target.def (conditional_register_usage): Mention
810 define_register_constraint instead of old-style constraint macros.
811 * doc/tm.texi.in: Remove documentation for old-style constraint macros.
812 * doc/tm.texi: Regenerate.
813 * genoutput.c: Remove USE_MD_CONSTRAINTS conditions and all code
814 protected by !USE_MD_CONSTRAINTS.
815 * config/frv/frv.md: Remove quote from old version of documentation.
816 * config/frv/frv.c (frv_conditional_register_usage): Likewise.
817 * config/m32r/m32r.c (easy_di_const, easy_df_const): Avoid mentioning
818 CONST_DOUBLE_OK_FOR_LETTER.
819 * config/sh/constraints.md: Likewise EXTRA_CONSTRAINT.
821 2014-06-02 Andrew Pinski <apinski@cavium.com>
823 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER):
824 /lib/ld-linux32-aarch64.so.1 is used for ILP32.
825 (LINUX_TARGET_LINK_SPEC): Update linker script for ILP32.
826 file whose name depends on -mabi= and -mbig-endian.
827 * config/aarch64/t-aarch64-linux (MULTILIB_OSDIRNAMES):
828 Handle LP64 better and handle ilp32 too.
829 (MULTILIB_OPTIONS): Delete.
830 (MULTILIB_DIRNAMES): Delete.
832 2014-06-02 Andrew MacLeod <amacleod@redhat.com>
834 * expr.h: Remove prototypes of functions defined in builtins.c.
835 * tree.h: (build_call_expr_*, build_string_literal): Add prototypes.
836 Remove prototypes of functions defined in builtins.c.
837 * builtins.h: Update prototype list to include all exported functions.
838 * builtins.c: (default_libc_has_function, gnu_libc_has_function,
839 no_c99_libc_has_function): Move to targhooks.c
840 (build_string_literal, build_call_expr_loc_array,
841 build_call_expr_loc_vec, build_call_expr_loc, build_call_expr): Move
843 (expand_builtin_object_size, fold_builtin_object_size): Make static.
844 * targhooks.c (default_libc_has_function, gnu_libc_has_function,
845 no_c99_libc_has_function): Relocate from builtins.c.
846 * tree.c: Include builtins.h.
847 (build_call_expr_loc_array, build_call_expr_loc_vec,
848 build_call_expr_loc, build_call_expr, build_string_literal): Relocate
850 * fold-const.h (fold_fma): Move prototype to builtins.h.
851 * realmpfr.h (do_mpc_arg2): Move prototype to builtins.h.
852 * asan.c: Include builtins.h.
853 * cfgexpand.c: Likewise.
854 * convert.c: Likewise.
855 * emit-rtl.c: Likewise.
856 * except.c: Likewise.
858 * fold-const.c: Likewise.
859 * gimple-fold.c: Likewise.
860 * gimple-ssa-strength-reduction.c: Likewise.
861 * gimplify.c: Likewise.
862 * ipa-inline.c: Likewise.
863 * ipa-prop.c: Likewise.
864 * lto-streamer-out.c: Likewise.
866 * tree-inline.c: Likewise.
867 * tree-object-size.c: Likewise.
868 * tree-sra.c: Likewise.
869 * tree-ssa-ccp.c: Likewise.
870 * tree-ssa-forwprop.c: Likewise.
871 * tree-ssa-loop-ivcanon.c: Likewise.
872 * tree-ssa-loop-ivopts.c: Likewise.
873 * tree-ssa-math-opts.c: Likewise.
874 * tree-ssa-reassoc.c: Likewise.
875 * tree-ssa-threadedge.c: Likewise.
876 * tree-streamer-in.c: Likewise.
877 * tree-vect-data-refs.c: Likewise.
878 * tree-vect-patterns.c: Likewise.
879 * tree-vect-stmts.c: Likewise.
880 * config/aarch64/aarch64.c: Likewise.
881 * config/alpha/alpha.c: Likewise.
882 * config/arc/arc.c: Likewise.
883 * config/arm/arm.c: Likewise.
884 * config/avr/avr.c: Likewise.
885 * config/bfin/bfin.c: Likewise.
886 * config/c6x/c6x.c: Likewise.
887 * config/cr16/cr16.c: Likewise.
888 * config/cris/cris.c: Likewise.
889 * config/epiphany/epiphany.c: Likewise.
890 * config/fr30/fr30.c: Likewise.
891 * config/frv/frv.c: Likewise.
892 * config/h8300/h8300.c: Likewise.
893 * config/i386/i386.c: Likewise.
894 * config/i386/winnt.c: Likewise.
895 * config/ia64/ia64.c: Likewise.
896 * config/iq2000/iq2000.c: Likewise.
897 * config/lm32/lm32.c: Likewise.
898 * config/m32c/m32c.c: Likewise.
899 * config/m32r/m32r.c: Likewise.
900 * config/m68k/m68k.c: Likewise.
901 * config/mcore/mcore.c: Likewise.
902 * config/mep/mep.c: Likewise.
903 * config/microblaze/microblaze.c: Likewise.
904 * config/mips/mips.c: Likewise.
905 * config/mmix/mmix.c: Likewise.
906 * config/mn10300/mn10300.c: Likewise.
907 * config/moxie/moxie.c: Likewise.
908 * config/msp430/msp430.c: Likewise.
909 * config/nds32/nds32.c: Likewise.
910 * config/pa/pa.c: Likewise.
911 * config/pdp11/pdp11.c: Likewise.
912 * config/picochip/picochip.c: Likewise.
913 * config/rl78/rl78.c: Likewise.
914 * config/rs6000/rs6000.c: Likewise.
915 * config/rx/rx.c: Likewise.
916 * config/s390/s390.c: Likewise.
917 * config/score/score.c: Likewise.
918 * config/sh/sh.c: Likewise.
919 * config/sparc/sparc.c: Likewise.
920 * config/spu/spu.c: Likewise.
921 * config/stormy16/stormy16.c: Likewise.
922 * config/tilegx/tilegx.c: Likewise.
923 * config/tilepro/tilepro.c: Likewise.
924 * config/v850/v850.c: Likewise.
925 * config/vax/vax.c: Likewise.
926 * config/xtensa/xtensa.c: Likewise.
928 2014-06-02 Jeff Law <law@redhat.com>
930 PR rtl-optimization/61094
931 * ree.c (combine_reaching_defs): Do not reextend an insn if it
932 was marked as do_no_reextend. If a copy is needed to eliminate
933 an extension, then mark it as do_not_reextend.
935 2014-06-02 Marcus Shawcroft <marcus.shawcroft@arm.com>
937 * config/aarch64/aarch64.md (set_fpcr): Drop ISB after FPCR write.
939 2014-06-02 Richard Henderson <rth@redhat.com>
942 * config/alpha/alpha.c (print_operand_address): Allow symbolic
943 addresses inside asms. Use output_operand_lossage instead of
946 2014-06-02 Uros Bizjak <ubizjak@gmail.com>
949 * config/i386/i386.c (ix86_expand_vec_perm) [case V32QImode]: Use
950 GEN_INT (-128) instead of GEN_INT (128) to set MSB of QImode constant.
952 2014-06-02 Tom de Vries <tom@codesourcery.com>
954 * config/aarch64/aarch64.c (aarch64_float_const_representable_p): Handle
955 case that x has VOIDmode.
957 2014-06-02 Bernd Schmidt <bernds@codesourcery.com>
959 * varasm.c (copy_constant): Delete function.
960 (build_constant_desc): Don't call it.
962 2014-06-02 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
965 * config/arm/arm.h (TARGET_SUPPORTS_WIDE_INT): Define.
966 * config/arm/arm.md (mov64 splitter): Replace const_double_operand
967 with immediate_operand.
969 2014-06-02 Andreas Schwab <schwab@suse.de>
972 (ia64_first_cycle_multipass_dfa_lookahead_guard): Check
973 pending_data_specs first.
975 2014-06-02 Richard Biener <rguenther@suse.de>
977 PR tree-optimization/61378
978 * tree-ssa-sccvn.c (vn_reference_lookup_3): Initialize
981 2014-06-01 Uros Bizjak <ubizjak@gmail.com>
983 * config/i386/constraints.md (Bw): Rename from 'w'.
984 (Bz): Rename from 'z'.
985 * config/i386/i386.md: Change 'w' to 'Bw' and 'z' to 'Bz' globally.
987 2014-06-01 Kai Tietz <ktietz@redhat.com>
990 * config/i386/constrains.md (define_constrain): New 'Bs' constraint.
991 * config/i386/i386.md (sibcall_insn_operand): Use Bs
992 instead of m constraint.
994 2014-05-31 Andreas Schwab <schwab@linux-m68k.org>
996 * config/m68k/m68k.md (beq0_di, bne0_di): Make the "o" constraint
997 a separate alternative where the scratch operand 2 is marked as
1000 2014-05-31 Kugan Vivekanandarajah <kuganv@linaro.org>
1002 * config/arm/arm.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New define.
1003 (arm_builtins) : Add ARM_BUILTIN_GET_FPSCR and ARM_BUILTIN_SET_FPSCR.
1004 (bdesc_2arg) : Add description for builtins __builtins_arm_set_fpscr
1005 and __builtins_arm_get_fpscr.
1006 (arm_init_builtins) : Initialize builtins __builtins_arm_set_fpscr and
1007 __builtins_arm_get_fpscr.
1008 (arm_expand_builtin) : Expand builtins __builtins_arm_set_fpscr and
1009 __builtins_arm_ldfpscr.
1010 (arm_atomic_assign_expand_fenv): New function.
1011 * config/arm/vfp.md (set_fpscr): New pattern.
1012 (get_fpscr) : Likewise.
1013 * config/arm/unspecs.md (unspecv): Add VUNSPEC_GET_FPSCR and
1015 * doc/extend.texi (AARCH64 Built-in Functions) : Document
1016 __builtins_arm_set_fpscr, __builtins_arm_get_fpscr.
1018 2014-05-30 Jakub Jelinek <jakub@redhat.com>
1020 * asan.c (report_error_func): Add SLOW_P argument, use
1021 BUILT_IN_ASAN_*_N if set.
1022 (build_check_stmt): Likewise.
1023 (instrument_derefs): If T has insufficient alignment,
1024 force same handling as for odd sizes.
1026 * sanitizer.def (BUILT_IN_ASAN_REPORT_LOAD_N,
1027 BUILT_IN_ASAN_REPORT_STORE_N): New.
1028 * asan.c (struct asan_mem_ref): Change access_size type to
1030 (asan_mem_ref_init, asan_mem_ref_new, get_mem_refs_of_builtin_call,
1031 update_mem_ref_hash_table): Likewise.
1032 (asan_mem_ref_hasher::hash): Hash in a HWI.
1033 (report_error_func): Change size_in_bytes argument to HWI.
1034 Use *_N builtins if size_in_bytes is larger than 16 or not power of
1036 (build_shadow_mem_access): New function.
1037 (build_check_stmt): Use it. Change size_in_bytes argument to HWI.
1038 Handle size_in_bytes not power of two or larger than 16.
1039 (instrument_derefs): Don't give up if size_in_bytes is not
1040 power of two or is larger than 16.
1042 2014-05-30 Kai Tietz <ktietz@redhat.com>
1045 * config/i386/i386.c (x86_output_mi_thunk): Add memory case
1046 for sibling-tail-calls.
1047 * config/i386/i386.md (sibcall_insn_operand): Add memory-constrain
1049 * config/i386/predicates.md (sibcall_memory_operand): New predicate.
1050 (sibcall_insn_operand): Add check for sibcall_memory_operand.
1052 2014-05-30 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
1054 * config/avr/avr-mcus.def: Change ATA6289 ISA to AVR4
1055 * config/avr/avr-tables.opt: Regenerate.
1056 * config/avr/t-multilib: Regenerate.
1057 * doc/avr-mmcu.texi: Regenerate.
1059 2014-05-30 Ian Lance Taylor <iant@google.com>
1061 * config/i386/xmmintrin.h (_mm_pause): Move out of scope of pragma
1064 2014-05-30 Tom de Vries <tom@codesourcery.com>
1066 * config/i386/i386.c (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS):
1069 2014-05-30 Tom de Vries <tom@codesourcery.com>
1071 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
1072 * lra.c (initialize_lra_reg_info_element): Add init of
1073 actual_call_used_reg_set field.
1074 (lra): Call lra_create_live_ranges before lra_inheritance for
1076 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
1078 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
1079 instead of call_used_reg_set for -fuse-caller-save.
1080 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
1082 2014-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1084 * config/arm/thumb2.md (*thumb2_movhi_insn): Set type of movw
1086 * config/arm/vfp.md (*thumb2_movsi_vfp): Likewise.
1088 2014-05-30 Richard Sandiford <rdsandiford@googlemail.com>
1090 * ira.c (ira_get_dup_out_num): Check for output operands at
1091 the start of the loop. Handle cases where an included alternative
1092 follows an excluded one.
1094 2014-05-29 Mike Stump <mikestump@comcast.net>
1097 * collect2.c (maybe_run_lto_and_relink): Be sure to always run
1098 post ld passes when lto is used.
1100 2014-05-29 Vladimir Makarov <vmakarov@redhat.com>
1102 PR rtl-optimization/61325
1103 * lra-constraints.c (process_address): Rename to process_address_1.
1104 (process_address): New function.
1106 2014-05-29 Alan Lawrence <alan.lawrence@arm.com>
1108 * config/aarch64/aarch64-builtins.c (aarch64_types_binopv_qualifiers,
1109 TYPES_BINOPV): New static data.
1110 * config/aarch64/aarch64-simd-builtins.def (im_lane_bound):
1112 * config/aarch64/aarch64-simd.md (aarch64_ext,
1113 aarch64_im_lane_boundsi): New patterns.
1114 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_const_1): Match
1116 (aarch64_evpc_ext): New function.
1118 * config/aarch64/iterators.md (UNSPEC_EXT): New enum element.
1120 * config/aarch64/arm_neon.h (vext_f32, vext_f64, vext_p8, vext_p16,
1121 vext_s8, vext_s16, vext_s32, vext_s64, vext_u8, vext_u16, vext_u32,
1122 vext_u64, vextq_f32, vextq_f64, vextq_p8, vextq_p16, vextq_s8,
1123 vextq_s16, vextq_s32, vextq_s64, vextq_u8, vextq_u16, vextq_u32,
1124 vextq_u64): Replace __asm with __builtin_shuffle and im_lane_boundsi.
1126 2014-05-29 Tom de Vries <tom@codesourcery.com>
1128 * rtl.h (BLOCK_SYMBOL_CHECK): Use SYMBOL_REF_FLAGS.
1130 2014-05-29 Richard Earnshaw <rearnsha@arm.com>
1131 Richard Sandiford <rdsandiford@googlemail.com>
1133 * arm/iterators.md (shiftable_ops): New code iterator.
1134 (t2_binop0, arith_shift_insn): New code attributes.
1135 * arm/predicates.md (shift_nomul_operator): New predicate.
1136 * arm/arm.md (insn_enabled): Delete.
1137 (enabled): Remove insn_enabled test.
1138 (*arith_shiftsi): Delete. Replace with ...
1139 (*<arith_shift_insn>_multsi): ... new pattern.
1140 (*<arith_shift_insn>_shiftsi): ... new pattern.
1141 * config/arm/arm.c (arm_print_operand): Handle operand format 'b'.
1143 2014-05-29 Radovan Obradovic <robradovic@mips.com>
1144 Tom de Vries <tom@codesourcery.com>
1146 * config/mips/mips.h (POST_CALL_TMP_REG): Define.
1147 * config/mips/mips.c (mips_emit_call_insn): Add POST_CALL_TMP_REG
1149 (mips_split_call): Use POST_CALL_TMP_REG.
1150 (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): Redefine to true.
1152 2014-05-29 Tom de Vries <tom@codesourcery.com>
1154 * final.c (collect_fn_hard_reg_usage): Guard variable declaration
1155 with #ifdef STACK_REGS.
1157 2014-05-28 Jan Hubicka <hubicka@ucw.cz>
1159 * varasm.c (get_variable_section): Walk aliases.
1160 (place_block_symbol): Walk aliases.
1162 2014-05-28 Tom de Vries <tom@codesourcery.com>
1165 2014-05-28 Tom de Vries <tom@codesourcery.com>
1167 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
1168 * lra.c (initialize_lra_reg_info_element): Add init of
1169 actual_call_used_reg_set field.
1170 (lra): Call lra_create_live_ranges before lra_inheritance for
1172 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
1174 * lra-constraints.c (need_for_call_save_p): Use
1175 actual_call_used_reg_set instead of call_used_reg_set for
1177 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
1179 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
1181 * doc/md.texi: Document that the % constraint character must
1182 be at the beginning of the string.
1183 * genoutput.c (validate_insn_alternatives): Check that '=',
1184 '+' and '%' only appear at the beginning of a constraint.
1185 * ira.c (commutative_constraint_p): Delete.
1186 (ira_get_dup_out_num): Expect the '%' commutativity marker to be
1187 at the start of the string.
1188 * config/alpha/alpha.md (*movmemdi_1, *clrmemdi_1): Remove
1190 * config/arm/neon.md (bicdi3_neon): Likewise.
1191 * config/iq2000/iq2000.md (addsi3_internal, subsi3_internal, sgt_si)
1192 (slt_si, sltu_si): Likewise.
1193 * config/vax/vax.md (sbcdi3): Likewise.
1194 * config/h8300/h8300.md (*cmpstz): Remove duplicate '+'.
1195 * config/arc/arc.md (mulsi_600, mulsidi_600, umulsidi_600)
1196 (mul64): Move '%' to beginning of constraint.
1197 * config/arm/arm.md (*xordi3_insn): Likewise.
1198 * config/nds32/nds32.md (add<mode>3, mulsi3, andsi3, iorsi3)
1201 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
1203 * doc/md.texi: Document the restrictions on the "enabled" attribute.
1205 2014-05-28 Jason Merrill <jason@redhat.com>
1208 * cgraph.h (symtab_node::get_comdat_group_id): New.
1209 * cgraphunit.c (analyze_functions): Call it.
1210 * symtab.c (dump_symtab_node): Likewise.
1211 * tree.c (decl_comdat_group_id): New.
1212 * tree.h: Declare it.
1213 * lto-streamer-out.c (write_symbol): Use it.
1214 * trans-mem.c (ipa_tm_create_version_alias): Likewise.
1216 2014-05-28 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
1218 PR bootstrap/PR61146
1219 * wide-int.cc: Do not include longlong.h when compiling with clang.
1221 2014-05-28 Richard Biener <rguenther@suse.de>
1223 * tree-ssa-propagate.c (add_control_edge): Print less vertical space.
1224 * tree-vrp.c (extract_range_from_ssa_name): Also copy VR_UNDEFINED.
1225 (vrp_visit_assignment_or_call): Print less vertical space.
1226 (vrp_visit_stmt): Likewise.
1227 (vrp_visit_phi_node): Likewise. For a PHI argument with
1228 VR_VARYING range consider recording it as copy.
1230 2014-05-28 Richard Biener <rguenther@suse.de>
1233 2014-05-28 Richard Biener <rguenther@suse.de>
1235 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
1237 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
1239 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
1240 sufficiently aligned and an offset is used at the same time.
1241 (expand_expr_real_1): Likewise.
1243 2014-05-28 Richard Biener <rguenther@suse.de>
1246 * fold-const.c (fold_comparison): When folding
1247 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
1248 the sign of the remaining constant operand stays the same.
1250 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
1252 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1253 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
1254 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
1256 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES is true.
1257 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
1258 (m32bit-doubles) Likewise.
1259 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
1260 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
1263 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1265 * configure.ac ($gcc_cv_ld_clearcap): New test.
1266 * configure: Regenerate.
1267 * config.in: Regenerate.
1268 * config/sol2.opt (mclear-hwcap): New option.
1269 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
1270 * config/sol2-clearcap.map: Moved here from
1271 testsuite/gcc.target/i386/clearcap.map.
1272 * config/sol2-clearcapv2.map: Move here from
1273 gcc.target/i386/clearcapv2.map.
1274 * config/t-sol2 (install): Depend on install-clearcap-map.
1275 (install-clearcap-map): New target.
1276 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
1279 2014-05-28 Richard Biener <rguenther@suse.de>
1281 * hwint.h (*_HALF_WIDE_INT*): Move to ...
1282 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
1283 ... here and remove the rest.
1284 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
1286 2014-05-28 Richard Biener <rguenther@suse.de>
1288 PR tree-optimization/61335
1289 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
1290 new range fails, drop to varying.
1292 2014-05-28 Tom de Vries <tom@codesourcery.com>
1294 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
1295 * lra.c (initialize_lra_reg_info_element): Add init of
1296 actual_call_used_reg_set field.
1297 (lra): Call lra_create_live_ranges before lra_inheritance for
1299 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
1301 * lra-constraints.c (need_for_call_save_p): Use
1302 actual_call_used_reg_set instead of call_used_reg_set for
1304 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
1306 2014-05-28 Radovan Obradovic <robradovic@mips.com>
1307 Tom de Vries <tom@codesourcery.com>
1309 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save
1311 (@item -fuse-caller-save): New item.
1313 2014-05-28 Radovan Obradovic <robradovic@mips.com>
1314 Tom de Vries <tom@codesourcery.com>
1316 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
1317 OPT_fuse_caller_save.
1319 2014-05-28 Radovan Obradovic <robradovic@mips.com>
1320 Tom de Vries <tom@codesourcery.com>
1322 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
1323 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
1324 get_call_reg_set_usage.
1325 * resource.c (mark_set_resources, mark_target_live_regs): Use
1326 get_call_reg_set_usage.
1327 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
1329 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
1330 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
1331 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
1332 * ira-build.c (ira_create_allocno): Init
1333 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
1334 (create_cap_allocno, propagate_allocno_info)
1335 (propagate_some_info_from_allocno)
1336 (copy_info_to_removed_store_destinations): Handle
1337 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
1338 * ira-costs.c (ira_tune_allocno_costs): Use
1339 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
1341 2014-05-28 Radovan Obradovic <robradovic@mips.com>
1342 Tom de Vries <tom@codesourcery.com>
1344 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
1345 and function_used_regs_valid fields.
1346 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
1347 find_all_hard_reg_sets.
1348 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
1349 (get_call_reg_set_usage): New function.
1350 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
1351 * regs.h (get_call_reg_set_usage): Declare.
1353 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
1356 * config/dbx.h (License): Add Runtime Library Exception.
1357 * config/newlib-stdint.h (License): Same.
1358 * config/rtems.h (License): Same
1359 * config/initfini-array.h (License): Same
1360 * config/v850/v850.h (License): Same.
1361 * config/v850/v850-opts.h (License): Same
1362 * config/v850/rtems.h (License): Same.
1364 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
1367 * doc/extend.texi (Local Labels): Note that label differences are
1368 not supported for AVR.
1370 2014-05-27 Andrew Pinski <apinski@cavium.com>
1372 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
1373 Use <w> for the register in assembly template.
1374 (stack_protect_test): Use the mode of operands[0] for the result.
1375 (stack_protect_test_<mode>): Use <w> for the register
1376 in assembly template.
1378 2014-05-27 DJ Delorie <dj@redhat.com>
1380 * config/rx/rx.c (add_vector_labels): New.
1381 (rx_output_function_prologue): Call it.
1382 (rx_handle_func_attribute): Don't require empty arguments.
1383 (rx_handle_vector_attribute): New.
1384 (rx_attribute_table): Add "vector" attribute.
1385 * doc/extend.texi (interrupt, vector): Document new/changed
1386 RX-specific attributes.
1388 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
1390 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
1392 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
1393 predicate to detect a negative quotient.
1395 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
1397 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
1398 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
1399 Add X - Y CMP 0 to X CMP Y transformation.
1400 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
1402 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
1404 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
1407 2014-05-27 Steve Ellcey <sellcey@mips.com>
1409 * config/mips/mips.c: Add include of cgraph.h.
1411 2014-05-27 Richard Biener <rguenther@suse.de>
1413 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
1415 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
1418 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
1419 * config/arm/arm-cores.def (License): Same.
1420 * config/arm/arm-opts.h (License): Same.
1421 * config/arm/aout.h (License): Same.
1422 * config/arm/bpabi.h (License): Same.
1423 * config/arm/elf.h (License): Same.
1424 * config/arm/linux-elf.h (License): Same.
1425 * config/arm/linux-gas.h (License): Same.
1426 * config/arm/netbsd-elf.h (License): Same.
1427 * config/arm/uclinux-eabi.h (License): Same.
1428 * config/arm/uclinux-elf.h (License): Same.
1429 * config/arm/vxworks.h (License): Same.
1431 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1433 * config/arm/neon.md (neon_bswap<mode>): New pattern.
1434 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
1435 (arm_init_neon_builtins): Handle NEON_BSWAP.
1436 Define required type nodes.
1437 (arm_expand_neon_builtin): Handle NEON_BSWAP.
1438 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
1439 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
1440 * config/arm/iterators.md (VDQHSD): New mode iterator.
1442 2014-05-27 Richard Biener <rguenther@suse.de>
1444 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
1445 Try using literal operands when comparing value-ranges failed.
1447 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
1449 * ira.c (commutative_operand): Adjust for change to recog_data.
1450 [Missing from previous commit.]
1452 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
1454 * system.h (TEST_BIT): New macro.
1455 * recog.h (alternative_mask): New type.
1456 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
1457 (recog_data_d): Replace alternative_enabled_p array with
1458 enabled_alternatives.
1459 (target_recog): New structure.
1460 (default_target_recog, this_target_recog): Declare.
1461 (get_enabled_alternatives, recog_init): Likewise.
1462 * recog.c (default_target_recog, this_target_recog): New variables.
1463 (get_enabled_alternatives): New function.
1464 (extract_insn): Use it.
1465 (recog_init): New function.
1466 (preprocess_constraints, constrain_operands): Adjust for change to
1468 * postreload.c (reload_cse_simplify_operands): Likewise.
1469 * reload.c (find_reloads): Likewise.
1470 * ira-costs.c (record_reg_classes): Likewise.
1471 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
1472 all alternatives after a disabled one would be skipped.
1473 (ira_implicitly_set_insn_hard_regs): Likewise.
1474 * ira.c (ira_setup_alts): Adjust for change to recog_data.
1475 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
1476 with enabled_alternatives.
1477 * lra.c (free_insn_recog_data): Update accordingly.
1478 (lra_update_insn_recog_data): Likewise.
1479 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
1480 * lra-constraints.c (process_alt_operands): Likewise. Handle
1481 only_alternative as part of the enabled mask.
1482 * target-globals.h (this_target_recog): Declare.
1483 (target_globals): Add a recog field.
1484 (restore_target_globals): Restore this_target_recog.
1485 * target-globals.c: Include recog.h.
1486 (default_target_globals): Initialize recog field.
1487 (save_target_globals): Likewise.
1488 * reginfo.c (reinit_regs): Call recog_init.
1489 * toplev.c (backend_init_target): Likewise.
1491 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
1493 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
1494 rather than any named insn's code.
1496 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
1499 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
1500 * config/arm/arm-cores.def (License): Same.
1502 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
1504 * tree.h (decl_comdat_group): Declare.
1505 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
1506 * tree.c (decl_comdat_group): Here.
1508 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
1510 PR rtl-optimization/61222
1511 * combine.c (simplify_shift_const_1): When moving a PLUS outside
1512 the shift, truncate the PLUS operand to the result mode.
1514 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
1517 * config/i386/i386.c (ix86_rtx_costs)
1518 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
1521 2014-05-26 Martin Jambor <mjambor@suse.cz>
1523 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
1526 2014-05-26 Richard Biener <rguenther@suse.de>
1528 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
1529 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
1530 Provide specializations.
1531 (wi::int_traits <HOST_WIDE_INT>,
1532 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
1534 2014-05-26 Alan Modra <amodra@gmail.com>
1537 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
1538 params and return a bool. Remove dead code. Update comment.
1539 Assert we have a const_int source. Remove bogus code from
1540 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
1541 handling of constants > 2G and reg_equal note, from..
1542 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
1543 return value. Update comment. If we can, use a new pseudo
1544 for intermediate calculations.
1545 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
1547 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
1548 call to rs6000_emit_set_const in splitter.
1549 (movdi_internal64+2, +3): Likewise.
1551 2014-05-26 Richard Biener <rguenther@suse.de>
1553 * system.h: Define __STDC_FORMAT_MACROS before
1554 including inttypes.h.
1555 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
1556 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
1557 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
1558 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
1559 HOST_WIDEST_INT_C): Remove.
1560 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
1561 if C99 inttypes.h is not available.
1562 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
1563 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
1564 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
1565 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
1566 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
1567 (struct output_info): Likewise.
1568 (print_statistics): Adjust.
1569 (dump_bitmap_statistics): Likewise.
1570 * bt-load.c (migrate_btr_defs): Print with PRId64.
1571 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
1572 (MAX_SAFE_MULTIPLIER): Adjust.
1573 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
1574 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
1575 dump_cgraph_node): Likewise.
1576 * final.c (dump_basic_block_info): Likewise.
1577 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
1578 * gcov.c (format_gcov): Likewise.
1579 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
1581 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
1582 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
1583 (inline_small_functions, dump_overall_stats, dump_inline_stats):
1584 Use PRId64 for dumping.
1585 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
1586 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
1587 (add_allocno_hard_regs): Adjust.
1588 * loop-doloop.c (doloop_modify): Print using PRId64.
1589 * loop-iv.c (inverse): Compute in uint64_t.
1590 (determine_max_iter, iv_number_of_iterations): Likewise.
1591 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
1593 * lto-streamer-out.c (write_symbol): Use uint64_t.
1594 * mcf.c (CAP_INFINITY): Use int64_t maximum.
1595 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
1596 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
1597 * modulo-sched.c (const_iteration_count): Use int64_t.
1598 (sms_schedule): Dump using PRId64.
1599 * predict.c (dump_prediction): Likewise.
1600 * pretty-print.h (pp_widest_integer): Remove.
1601 * profile.c (get_working_sets, is_edge_inconsistent,
1602 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
1603 * tree-pretty-print.c (pp_double_int): Remove case handling
1604 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
1605 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
1607 (pass_optimize_bswap::execute): Remove restriction on hosts.
1608 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
1609 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
1610 * tree.c (widest_int_cst_value): Remove.
1611 * tree.h (widest_int_cst_value): Likewise.
1612 * value-prof.c (dump_histogram_value): Print using PRId64.
1613 * gengtype.c (main): Also inject int64_t.
1614 * ggc-page.c (struct max_alignment): Use int64_t.
1615 * alloc-pool.c (struct allocation_object_def): Likewise.
1616 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
1618 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
1619 * doc/tm.texi: Regenerated.
1620 * gengtype-lex.l (IWORD): Handle [u]int64_t.
1621 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
1622 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
1623 mmix_output_register_setting): Use [u]int64_t in prototypes.
1624 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
1625 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
1626 mmix_output_octa, mmix_output_shifted_value): Adjust.
1627 (mmix_intval): Adjust. Remove unreachable case.
1628 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
1630 2014-05-26 Richard Biener <rguenther@suse.de>
1632 * configure.ac: Drop __int64 type check. Insist that we
1633 found uint64_t and int64_t.
1634 * hwint.h (HOST_BITS_PER___INT64): Remove.
1635 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
1636 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
1637 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
1638 (HOST_WIDEST_FAST_INT): Remove __int64 case.
1639 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
1640 for dst_q_src_df_rms_cdt.
1641 * configure: Regenerate.
1642 * config.in: Likewise.
1644 2014-05-26 Michael Tautschnig <mt@debian.org>
1647 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
1648 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
1650 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1652 PR rtl-optimization/61278
1653 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
1655 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1657 PR rtl-optimization/61220
1658 Part of PR rtl-optimization/61225
1659 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
1660 insn; skip split_edge for a block with only one successor.
1662 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1664 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
1667 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1669 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
1670 (update_vtable_references): New function.
1671 (function_and_variable_visibility): Rewrite also vtable initializers.
1672 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
1674 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1676 * ggc.h (ggc_grow): New function.
1677 * ggc-none.c (ggc_grow): New function.
1678 * ggc-page.c (ggc_grow): Likewise.
1680 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1682 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
1683 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
1684 comdat_can_be_unshared_p, cgraph_externally_visible_p,
1685 varpool_externally_visible_p, can_replace_by_local_alias,
1686 update_visibility_by_resolution_info, function_and_variable_visibility,
1687 pass_data_ipa_function_and_variable_visibility,
1688 make_pass_ipa_function_and_variable_visibility,
1689 whole_program_function_and_variable_visibility,
1690 pass_data_ipa_whole_program_visibility,
1691 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
1692 * cgraph.h (cgraph_local_node_p): Declare.
1693 * ipa-visibility.c: New file.
1694 * Makefile.in (OBJS): Add ipa-visiblity.o
1696 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1698 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
1699 that var decl is available.
1701 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1703 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
1704 symtab_node pointer.
1705 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
1706 (find_decls_types_r): Do not walk COMDAT_GROUP.
1707 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
1708 * varasm.c (make_decl_one_only): Use set_comdat_group;
1709 create node if needed.
1710 * ipa-inline-transform.c (save_inline_function_body): Update
1711 way we decl->symtab mapping.
1712 * symtab.c (symtab_hash, hash_node, eq_node
1713 symtab_insert_node_to_hashtable): Remove.
1714 (symtab_register_node): Update.
1715 (symtab_unregister_node): Update.
1716 (symtab_get_node): Reimplement as inline function.
1717 (symtab_add_to_same_comdat_group): Update.
1718 (symtab_dissolve_same_comdat_group_list): Update.
1719 (dump_symtab_base): Update.
1720 (verify_symtab_base): Update.
1721 (symtab_make_decl_local): Update.
1722 (fixup_same_cpp_alias_visibility): Update.
1723 (symtab_nonoverwritable_alias): Update.
1724 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
1725 * ipa.c (update_visibility_by_resolution_info): UPdate.
1726 * bb-reorder.c: Include cgraph.h
1727 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
1729 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
1730 * cgraph.c (cgraph_get_create_node): Update.
1731 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
1733 (symtab_get_node): Make inline.
1734 (symtab_insert_node_to_hashtable): Remove.
1735 (symtab_can_be_discarded): Update.
1736 (decl_comdat_group): New function.
1737 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
1739 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
1741 (read_comdat_group): New function.
1742 (input_node, input_varpool_node): Use it.
1743 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
1745 * mips.c (mips_start_unique_function): Likewise.
1746 (ix86_code_end): Likewise.
1747 (rs6000_code_end): Likweise.
1748 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
1750 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1752 * gengtype-state.c (fatal_reading_state): Bring offline.
1753 * optabs.c (widening_optab_handler): Bring offline.
1754 * optabs.h (widening_optab_handler): Likewise.
1755 * final.c (get_attr_length_1): Likewise.
1757 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1759 * sched-int.h (sd_iterator_cond): Manually tail recurse.
1761 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1763 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
1764 (ppc440-compare): Include shift with dot.
1765 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
1766 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
1768 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
1770 (e6500_sfx2): Include it.
1771 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
1772 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
1773 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
1774 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
1775 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
1776 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
1777 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
1778 *lshiftrt_internal1le, *lshiftrt_internal1be,
1779 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
1780 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
1781 *rotldi3_internal10le, *rotldi3_internal10be,
1782 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
1783 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
1784 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
1785 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
1786 define_insns): Use type "shift" in the appropriate alternatives.
1788 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1790 * config/rs6000/rs6000.md (type): Add "logical". Delete
1792 (dot): Adjust comment.
1793 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
1794 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
1795 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
1796 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
1797 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
1798 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
1799 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
1800 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1802 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1803 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1804 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1805 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1806 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1807 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1808 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1809 * config/rs6000/8540.md (ppc8540_su): Adjust.
1810 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1811 cell-cmp-microcoded): Adjust.
1812 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1813 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1814 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1815 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1816 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1817 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1818 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1819 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1820 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1822 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1823 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
1824 Adjust. Adjust comment.
1825 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1826 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1828 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1830 * config/rs6000/rs6000.md (type): Add "add".
1831 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
1832 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
1833 define_insns): Use it.
1834 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1836 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1837 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1838 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1839 * config/rs6000/601.md (ppc601-integer): Adjust.
1840 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1841 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1842 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1843 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1844 * config/rs6000/8540.md (ppc8540_su): Adjust.
1845 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1846 cell-cmp-microcoded): Adjust.
1847 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1848 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1849 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1850 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1851 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1852 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1853 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1854 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1855 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1857 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1858 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
1859 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1860 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1862 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1864 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
1865 "delayed_compare", "var_delayed_compare".
1866 (var_shift): New attribute.
1867 (cell_micro): Adjust.
1868 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
1869 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
1870 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
1871 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
1872 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
1873 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
1874 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
1875 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
1876 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
1877 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
1878 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
1879 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
1880 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
1881 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
1882 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
1883 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
1884 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
1885 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
1886 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
1887 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
1888 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
1889 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
1890 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
1891 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1892 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1894 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1895 * config/rs6000/440.md (ppc440-integer): Adjust.
1896 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1897 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
1898 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1899 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1900 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1901 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1902 * config/rs6000/8540.md (ppc8540_su): Adjust.
1903 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1904 cell-cmp-microcoded): Adjust.
1905 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
1906 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1907 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
1908 e500mc64_delayed): Adjust.
1909 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
1910 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
1911 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1912 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
1913 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
1914 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
1915 power6-delayed-compare, power6-var-delayed-compare): Adjust.
1916 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
1917 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
1919 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1920 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1922 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1924 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
1925 (bits): New mode_attr.
1926 (idiv_ldiv): Delete mode_attr.
1927 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
1928 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1929 rs6000_adjust_priority, is_nonpipeline_insn,
1930 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1932 * config/rs6000/40x.md (ppc403-idiv): Adjust.
1933 * config/rs6000/440.md (ppc440-idiv): Adjust.
1934 * config/rs6000/476.md (ppc476-idiv): Adjust.
1935 * config/rs6000/601.md (ppc601-idiv): Adjust.
1936 * config/rs6000/603.md (ppc603-idiv): Adjust.
1937 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
1938 ppc620-ldiv): Adjust.
1939 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
1940 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
1941 * config/rs6000/8540.md (ppc8540_divide): Adjust.
1942 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
1943 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
1944 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
1945 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
1946 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
1947 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
1948 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
1949 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
1950 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
1951 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
1952 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
1953 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
1954 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
1955 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
1956 * config/rs6000/titan.md (titan_fxu_div): Adjust.
1958 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1960 * config/rs6000/rs6000.md (type): Delete "insert_word",
1961 "insert_dword". Add "insert".
1962 (size): Update comment.
1963 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1964 insn_must_be_first_in_group): Adjust.
1965 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
1966 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
1967 *insvsi_internal6, insvdi_internal): Adjust.
1969 * config/rs6000/40x.md (ppc403-integer): Adjust.
1970 * config/rs6000/440.md (ppc440-integer): Adjust.
1971 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
1972 * config/rs6000/601.md (ppc601-integer): Adjust.
1973 * config/rs6000/603.md (ppc603-integer): Adjust.
1974 * config/rs6000/6xx.md (ppc604-integer): Adjust.
1975 * config/rs6000/7450.md (ppc7450-integer): Adjust.
1976 * config/rs6000/7xx.md (ppc750-integer): Adjust.
1977 * config/rs6000/8540.md (ppc8540_su): Adjust.
1978 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
1979 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
1980 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1981 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
1982 * config/rs6000/e5500.md (e5500_sfx): Adjust.
1983 * config/rs6000/e6500.md (e6500_sfx): Adjust.
1984 * config/rs6000/mpc.md (mpccore-integer): Adjust.
1985 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
1986 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
1987 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
1988 * config/rs6000/power7.md (power7-integer): Adjust.
1989 * config/rs6000/power8.md (power8-1cyc): Adjust.
1990 * config/rs6000/rs64.md (rs64a-integer): Adjust.
1991 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1993 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1995 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
1996 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
1997 (size): New attribute.
1998 (dot): New attribute.
1999 (cell_micro): Adjust.
2000 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
2001 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
2002 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
2003 umuldi3_highpart): Adjust.
2004 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
2005 rs6000_adjust_priority, is_nonpipeline_insn,
2006 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
2008 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
2009 ppc405-imul3): Adjust.
2010 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
2011 * config/rs6000/476.md (ppc476-imul): Adjust.
2012 * config/rs6000/601.md (ppc601-imul): Adjust.
2013 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
2014 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
2015 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
2016 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
2017 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
2019 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
2020 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
2021 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
2023 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
2024 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
2025 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
2026 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
2027 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
2028 * config/rs6000/mpc.md (mpccore-imul): Adjust.
2029 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
2030 power4-lmul, power4-imul, power4-imul3): Adjust.
2031 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
2032 power5-lmul, power5-imul, power5-imul3): Adjust.
2033 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
2034 power6-lmul, power6-imul, power6-imul3): Adjust.
2035 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
2036 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
2038 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
2039 rs64a-lmul): Adjust.
2040 * config/rs6000/titan.md (titan_imul): Adjust.
2042 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
2044 * config/rs6000/rs6000.md (type): Add new value "halfmul".
2045 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
2046 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
2047 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
2048 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
2049 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
2050 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
2051 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
2052 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
2053 * config/rs6000/titan.md: Delete nonsensical comment.
2054 (titan_imul): Add type imul3.
2055 (titan_mulhw): Remove type imul3; add type halfmul.
2057 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
2059 * config/rs6000/rs6000.md (type): Reorder, reformat.
2061 2014-05-23 Martin Jambor <mjambor@suse.cz>
2063 PR tree-optimization/53787
2064 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
2065 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
2066 analysis_done, update all uses.
2067 * ipa-prop.c: Include domwalk.h
2068 (param_analysis_info): Removed.
2069 (param_aa_status): New type.
2070 (ipa_bb_info): Likewise.
2071 (func_body_info): Likewise.
2072 (ipa_get_bb_info): New function.
2073 (aa_overwalked): Likewise.
2074 (find_dominating_aa_status): Likewise.
2075 (parm_bb_aa_status_for_bb): Likewise.
2076 (parm_preserved_before_stmt_p): Changed to use new param AA info.
2077 (load_from_unmodified_param): Accept func_body_info as a parameter
2078 instead of parms_ainfo.
2079 (parm_ref_data_preserved_p): Changed to use new param AA info.
2080 (parm_ref_data_pass_through_p): Likewise.
2081 (ipa_load_from_parm_agg_1): Likewise. Update callers.
2082 (compute_complex_assign_jump_func): Changed to use new param AA info.
2083 (compute_complex_ancestor_jump_func): Likewise.
2084 (ipa_compute_jump_functions_for_edge): Likewise.
2085 (ipa_compute_jump_functions): Removed.
2086 (ipa_compute_jump_functions_for_bb): New function.
2087 (ipa_analyze_indirect_call_uses): Likewise, moved variable
2089 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
2090 and info, moved variable declarations down.
2091 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
2093 (ipa_analyze_stmt_uses): Likewise.
2094 (ipa_analyze_params_uses): Removed.
2095 (ipa_analyze_params_uses_in_bb): New function.
2096 (ipa_analyze_controlled_uses): Likewise.
2097 (free_ipa_bb_info): Likewise.
2098 (analysis_dom_walker): New class.
2099 (ipa_analyze_node): Handle node-specific forbidden analysis,
2100 initialize and free func_body_info, use dominator walker.
2101 (ipcp_modif_dom_walker): New class.
2102 (ipcp_transform_function): Create and free func_body_info, use
2103 ipcp_modif_dom_walker, moved a lot of functionality there.
2105 2014-05-23 Marek Polacek <polacek@redhat.com>
2106 Jakub Jelinek <jakub@redhat.com>
2108 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
2109 * gcc.c (sanitize_spec_function): Likewise.
2110 * convert.c (convert_to_integer): Include "ubsan.h". Add
2111 floating-point to integer instrumentation.
2112 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
2113 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
2114 SANITIZE_NONDEFAULT.
2115 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
2116 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
2117 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
2118 * ubsan.c: Include "realmpfr.h" and "dfp.h".
2119 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
2120 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
2121 float/double/long double.
2122 (ubsan_instrument_float_cast): New function.
2123 * ubsan.h (ubsan_instrument_float_cast): Declare.
2125 2014-05-23 Jiong Wang <jiong.wang@arm.com>
2127 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
2129 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
2130 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
2131 Adjust for tailcalling through registers.
2132 * config/aarch64/aarch64.h (enum reg_class): New caller save
2134 (REG_CLASS_NAMES): Likewise.
2135 (REG_CLASS_CONTENTS): Likewise.
2136 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
2137 Allow tailcalling without decls.
2139 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
2141 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
2142 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
2144 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
2145 gsi, and variables v_* to v*.
2147 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
2149 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
2151 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
2153 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
2154 * omp-low.c: Update accordingly.
2156 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
2157 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
2158 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
2159 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
2160 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
2161 GF_OMP_TARGET_KIND_UPDATE.
2163 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
2164 Explicitly enumerate the expected region types.
2166 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
2169 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
2170 documentation; the old documentation didn't clearly state the
2171 constraints on the contents of the pointed-to storage.
2173 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
2175 Fix bootstrap error on ia64
2176 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
2177 Return default value.
2179 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
2181 PR tree-optimization/54733
2182 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
2184 (find_bswap_or_nop_load): New.
2185 (find_bswap_1): Renamed to ...
2186 (find_bswap_or_nop_1): This. Also add support for memory source.
2187 (find_bswap): Renamed to ...
2188 (find_bswap_or_nop): This. Also add support for memory source and
2189 detection of bitwise operations equivalent to load in target
2191 (execute_optimize_bswap): Likewise. Also move its leading comment back
2192 in place and split statement transformation into ...
2193 (bswap_replace): This.
2195 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
2197 PR rtl-optimization/61215
2198 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
2199 simplify_gen_subreg until final substitution.
2201 2014-05-23 Alan Modra <amodra@gmail.com>
2204 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
2205 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
2206 Use "Y" constraint rather than "m".
2208 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
2210 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
2212 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
2213 New function declaration.
2214 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
2215 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
2216 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
2217 (aarch64_init_builtins) : Initialize builtins
2218 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
2219 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
2220 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
2221 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
2222 and __builtins_aarch64_set_fpsr.
2223 (aarch64_atomic_assign_expand_fenv): New function.
2224 * config/aarch64/aarch64.md (set_fpcr): New pattern.
2225 (get_fpcr) : Likewise.
2226 (set_fpsr) : Likewise.
2227 (get_fpsr) : Likewise.
2228 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
2229 and UNSPECV_SET_FPSR.
2230 * doc/extend.texi (AARCH64 Built-in Functions) : Document
2231 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
2232 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
2234 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
2236 PR rtl-optimization/60969
2237 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
2238 constraints. Set up mem cost for NO_REGS case.
2240 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
2242 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
2244 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
2246 * config/darwin.c: Include "lto-section-names.h".
2247 (LTO_SEGMENT_NAME): Don't define.
2248 * config/i386/winnt.c: Include "lto-section-names.h".
2249 * lto-streamer.c: Include "lto-section-names.h".
2250 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
2251 * lto-wrapper.c: Include "lto-section-names.h".
2252 (LTO_SECTION_NAME_PREFIX): Don't define.
2253 * lto-section-names.h: New file.
2254 * cgraphunit.c: Include "lto-section-names.h".
2256 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
2258 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
2260 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
2263 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
2265 2014-05-22 Nick Clifton <nickc@redhat.com>
2267 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
2269 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
2271 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
2272 -> (T)A transformation to integer types.
2274 2014-05-22 Teresa Johnson <tejohnson@google.com>
2276 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
2277 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
2278 (gcov_rewrite): Use gcov_nonruntime_assert.
2280 (gcov_write_words): Ditto.
2281 (gcov_write_length): Ditto.
2282 (gcov_read_words): Use gcov_nonruntime_assert, and remove
2283 gcc_assert from IN_LIBGCOV code.
2284 (gcov_read_summary): Use gcov_error to flag profile corruption.
2285 (gcov_sync): Use gcov_nonruntime_assert.
2286 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
2287 (gcov_histo_index): Use gcov_nonruntime_assert.
2288 (static void gcov_histogram_merge): Ditto.
2289 (compute_working_sets): Ditto.
2290 * gcov-io.h (gcov_nonruntime_assert): Define.
2291 (gcov_error): Define for !IN_LIBGCOV
2293 2014-05-22 Richard Biener <rguenther@suse.de>
2295 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2296 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
2297 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
2298 and deallocation site.
2299 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2300 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
2301 passing through the incoming points-to set.
2302 (handle_lhs_call): Use flags argument instead of recomputing it.
2303 (find_func_aliases_for_call): Call handle_lhs_call with proper
2306 2014-05-22 Jakub Jelinek <jakub@redhat.com>
2308 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
2309 all padding bits in REAL_VALUE_TYPE are cleared.
2311 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
2313 Cleanup and improve multipass_dfa_lookahead_guard
2314 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
2315 (core2i7_first_cycle_multipass_begin,)
2316 (core2i7_first_cycle_multipass_issue,)
2317 (core2i7_first_cycle_multipass_backtrack): Update signature.
2318 * config/ia64/ia64.c
2319 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
2320 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
2321 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
2323 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
2324 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
2326 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
2328 * doc/tm.texi: Regenerate.
2330 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
2331 * haifa-sched.c (ready_try): Make signed to allow negative values.
2332 (rebug_ready_list_1): Update.
2333 (choose_ready): Simplify.
2334 (sched_extend_ready_list): Update.
2336 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
2338 Remove IA64 speculation tweaking flags
2339 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
2340 speculation tuning flags.
2341 (msched-prefer-non-data-spec-insns,)
2342 (msched-prefer-non-control-spec-insns): Obsolete options.
2343 * haifa-sched.c (choose_ready): Remove handling of
2344 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
2345 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
2346 and PREFER_NON_DATA_SPEC.
2347 * sel-sched.c (process_spec_exprs): Remove handling of
2348 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
2350 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
2352 Improve scheduling debug output
2353 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
2354 (advance_one_cycle): Update.
2355 (schedule_insn, queue_to_ready): Add debug printouts.
2356 (debug_ready_list_1): New static function.
2357 (debug_ready_list): Update.
2358 (max_issue): Add debug printouts.
2359 (dump_insn_stream): New static function.
2360 (schedule_block): Use it. Also better indent printouts.
2362 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
2364 Fix sched_insn debug counter
2365 * haifa-sched.c (schedule_insn): Update.
2366 (struct haifa_saved_data): Add nonscheduled_insns_begin.
2367 (save_backtrack_point, restore_backtrack_point): Update.
2368 (first_nonscheduled_insn): New static function.
2369 (queue_to_ready, choose_ready): Use it.
2370 (schedule_block): Init nonscheduled_insns_begin.
2371 (sched_emit_insn): Update.
2374 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
2376 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
2378 (aarch64_secondary_reload) : LikeWise.
2379 (aarch64_class_max_nregs) : Remove CORE_REGS.
2380 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
2381 (REG_CLASS_NAMES) : Likewise.
2382 (REG_CLASS_CONTENTS) : LikeWise.
2383 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
2385 2014-05-21 Guozhi Wei <carrot@google.com>
2388 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
2390 (vqdmulhq_n_s16): Likewise.
2392 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
2394 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
2396 2014-05-21 Marek Polacek <polacek@redhat.com>
2399 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
2401 2014-05-21 Martin Jambor <mjambor@suse.cz>
2403 * doc/invoke.texi (Optimize Options): Document parameters
2404 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
2405 ipa-cp-array-index-hint-bonus.
2407 2014-05-21 Mark Wielaard <mjw@redhat.com>
2410 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
2411 version >= 3 or not strict DWARF.
2412 * langhooks.h (struct lang_hooks_for_types): Add
2413 enum_underlying_base_type.
2414 * langhooks.c (lhd_enum_underlying_base_type): New function.
2415 * gcc/langhooks.h (struct lang_hooks_for_types): Add
2416 enum_underlying_base_type.
2417 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
2418 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
2419 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
2421 2014-05-21 Richard Biener <rguenther@suse.de>
2423 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
2425 2014-05-21 John Marino <gnugcc@marino.st>
2427 * config.gcc (*-*-dragonfly*): New target.
2428 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
2429 * configure: Regenerate.
2430 * config/dragonfly-stdint.h: New.
2431 * config/dragonfly.h: New.
2432 * config/dragonfly.opt: New.
2433 * config/i386/dragonfly.h: New.
2434 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
2436 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2438 * tree.def (VOID_CST): New.
2439 * tree-core.h (TI_VOID): New.
2440 * tree.h (void_node): New.
2441 * tree.c (tree_node_structure_for_code, tree_code_size)
2442 (iterative_hash_expr): Handle VOID_CST.
2443 (build_common_tree_nodes): Initialize void_node.
2445 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
2447 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
2449 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
2451 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
2454 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
2455 flag_reorder_blocks_and_partition.
2456 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
2458 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
2461 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
2463 (*addc_r_t): Add new insn_and_split.
2465 2014-05-21 Jakub Jelinek <jakub@redhat.com>
2468 * omp-low.c (handle_simd_reference): New function.
2469 (lower_rec_input_clauses): Use it. Defer adding reference
2470 initialization even for reduction without placeholder if in simd,
2471 handle it properly later on.
2473 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2475 PR tree-optimization/60899
2476 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
2477 assume all static symbols will have definition wile parsing and
2478 check the do have definition later in compilation; check that
2479 variable referring symbol will be output before concluding that
2480 reference is safe; be conservative for referring local statics;
2481 be more precise about when comdat is output in other partition.
2483 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2486 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
2488 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
2489 (ipa_inline): Loop inline_to_all_callers until no more aliases
2492 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2494 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
2495 set writeonly flag only for vars actually written to.
2497 2014-05-20 Dehao Chen <dehao@google.com>
2499 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
2500 and callee count to get clone count.
2501 * tree-inline.c (expand_call_inline): Use callee count instead of bb
2504 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
2506 PR rtl-optimization/61243
2507 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
2509 2014-05-20 Xinliang David Li <davidxl@google.com>
2511 * cgraphunit.c (walk_polymorphic_call_targets): Add
2512 dbgcnt and fopt-info support.
2513 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
2514 * ipa-devirt.c (ipa_devirt): Ditto.
2515 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
2516 * ipa.c (walk_polymorphic_call_targets): Ditto.
2517 * gimple-fold.c (fold_gimple_assign): Ditto.
2518 (gimple_fold_call): Ditto.
2519 * dbgcnt.def: New counter.
2521 2014-05-20 DJ Delorie <dj@redhat.com>
2523 * config/msp430/msp430.md (split): Don't allow subregs when
2524 splitting SImode adds.
2525 (andneghi): Fix subtraction logic.
2526 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
2528 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2530 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
2532 * except.c (switch_to_exception_section, resolve_unique_section,
2533 get_named_text_section, default_function_rodata_section,
2534 align_variable, get_block_for_decl, default_section_type_flags):
2535 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
2536 * symtab.c (symtab_add_to_same_comdat_group,
2537 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
2538 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
2540 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
2541 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
2542 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
2543 (c6x_function_in_section_p): Likewise.
2544 * config/darwin.c (machopic_select_section): Likewise.
2545 * config/arm/arm.c (arm_function_in_section_p): Likewise.
2546 * config/mips/mips.c (mips_function_rodata_section): Likewise.
2547 * config/mep/mep.c (mep_select_section): LIkewise.
2548 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
2550 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
2552 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
2553 EH region of calls to pure functions that can throw an exception.
2554 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
2555 (copy_reference_ops_from_call): Also copy the EH region of the call if
2556 it can throw an exception.
2558 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2560 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
2561 nested VEC_SELECTs that are inverses of each other.
2563 2014-05-20 Richard Biener <rguenther@suse.de>
2565 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
2566 (extract_and_process_scc_for_name): not here.
2567 (cond_dom_walker::before_dom_children): Only process
2568 stmts that end the BB in interesting ways.
2569 (run_scc_vn): Mark param uses as visited.
2571 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2573 * config/arm/arm.md (arith_shiftsi): Do not predicate for
2576 2014-05-20 Nick Clifton <nickc@redhat.com>
2578 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
2579 (msp430_gimplify_va_arg_expr): New function.
2580 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
2582 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
2583 operand 0 in order to prevent confusion about the number of
2586 2014-05-20 Richard Biener <rguenther@suse.de>
2588 PR tree-optimization/61221
2589 * tree-ssa-pre.c (el_to_update): Remove.
2590 (eliminate_dom_walker::before_dom_children): Handle released
2591 VDEFs by value-numbering them to the associated VUSE. Update
2592 stmt immediately for substituted call address.
2593 (eliminate): Remove delayed stmt updating code.
2594 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
2595 possibly late re-numbered vuses.
2596 (vn_reference_lookup_2): Adjust.
2597 (vn_reference_lookup_pieces): Likewise.
2598 (vn_reference_lookup): Likewise.
2600 2014-05-20 Richard Biener <rguenther@suse.de>
2602 * config.gcc: Remove need_64bit_hwint.
2603 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
2604 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
2606 * config.in: Regenerate.
2607 * configure: Likewise.
2609 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
2611 * doc/extend.texi: Create Label Attributes section,
2612 move all label attributes into it and reference it.
2614 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
2616 * arm.c (thumb1_reorg): When scanning backwards skip anything
2617 that's not a proper insn.
2619 2014-05-19 Richard Biener <rguenther@suse.de>
2621 PR tree-optimization/61221
2622 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
2623 Do nothing for unreachable blocks.
2624 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
2625 Improve unreachability detection.
2627 2014-05-19 Richard Biener <rguenther@suse.de>
2629 PR tree-optimization/61209
2630 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
2632 2014-05-19 Nick Clifton <nickc@redhat.com>
2634 * except.c (init_eh): Fix computation of builtin setjmp buffer
2635 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
2637 2014-05-19 Richard Biener <rguenther@suse.de>
2639 PR tree-optimization/61184
2640 * tree-vrp.c (is_negative_overflow_infinity): Use
2641 TREE_OVERFLOW_P and do that check first.
2642 (is_positive_overflow_infinity): Likewise.
2643 (is_overflow_infinity): Likewise.
2644 (vrp_operand_equal_p): Properly treat operands with
2645 differing overflow as not equal.
2647 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
2649 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
2650 shift simplification where it was intended.
2652 2014-05-19 Christian Bruel <christian.bruel@st.com>
2655 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
2657 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
2660 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
2663 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
2665 * reg-notes.def (CROSSING_JUMP): Likewise.
2666 * rtl.h (rtx_def): Update comment for jump flag.
2667 (CROSSING_JUMP_P): Define.
2668 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
2669 of a REG_CROSSING_JUMP note.
2670 * cfghooks.c (tidy_fallthru_edges): Likewise.
2671 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
2672 * emit-rtl.c (try_split): Likewise.
2673 * haifa-sched.c (sched_create_recovery_edges): Likewise.
2674 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
2675 * jump.c (redirect_jump_2): Likewise.
2676 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
2677 (relax_delay_slots): Likewise.
2678 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
2679 (bbit_di): Likewise.
2680 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
2681 * config/sh/sh.md (jump_compact): Likewise.
2682 * bb-reorder.c (rotate_loop): Likewise.
2683 (pass_duplicate_computed_gotos::execute): Likewise.
2684 (add_reg_crossing_jump_notes): Rename to...
2685 (update_crossing_jump_flags): ...this.
2686 (pass_partition_blocks::execute): Update accordingly.
2688 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
2690 * tree.h: Remove extraneous template <>.
2692 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2694 * ipa.c (symtab_remove_unreachable_nodes): Remove
2695 symbol from comdat group if its body was eliminated.
2696 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
2697 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
2698 (symtab_unregister_node): ... this one.
2699 (verify_symtab_base): More strict checking of comdats.
2700 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
2702 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2704 * tree-pass.h (make_pass_ipa_comdats): New pass.
2705 * timevar.def (TV_IPA_COMDATS): New timevar.
2706 * passes.def (pass_ipa_comdats): Add.
2707 * Makefile.in (OBJS): Add ipa-comdats.o
2708 * ipa-comdats.c: New file.
2710 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2712 * ipa.c (update_visibility_by_resolution_info): New function.
2713 (function_and_variable_visibility): Use it.
2715 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2717 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
2719 (FOR_EACH_DEFINED_SYMBOL): New macro.
2720 (varpool_first_static_initializer, varpool_next_static_initializer,
2721 varpool_first_defined_variable, varpool_next_defined_variable):
2723 (symtab_in_same_comdat_p): Correctly deal with inline functions.
2725 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2727 * ggc-page.c (ggc_handle_finalizers): Add comment.
2729 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2731 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
2732 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
2733 (ggc_internal_cleared_alloc): Likewise.
2734 * ggc-page.c (finalizer): New class.
2735 (vec_finalizer): Likewise.
2736 (globals::finalizers): New member.
2737 (globals::vec_finalizers): Likewise.
2738 (ggc_internal_alloc): Record the finalizer if any for the block being
2740 (ggc_handle_finalizers): New function.
2741 (ggc_collect): Call ggc_handle_finalizers.
2742 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
2744 (ggc_internal_cleared_alloc): Likewise.
2745 (finalize): New function.
2746 (need_finalization_p): Likewise.
2747 (ggc_alloc): Install the type's destructor as the finalizer if it
2749 (ggc_cleared_alloc): Likewise.
2750 (ggc_vec_alloc): Likewise.
2751 (ggc_cleared_vec_alloc): Likewise.
2753 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2755 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
2757 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2759 * alias.c (record_alias_subset): Adjust.
2760 * bitmap.c (bitmap_element_allocate): Likewise.
2761 (bitmap_gc_alloc_stat): Likewise.
2762 * cfg.c (init_flow): Likewise.
2763 (alloc_block): Likewise.
2764 (unchecked_make_edge): Likewise.
2765 * cfgloop.c (alloc_loop): Likewise.
2766 (flow_loops_find): Likewise.
2767 (rescan_loop_exit): Likewise.
2768 * cfgrtl.c (init_rtl_bb_info): Likewise.
2769 * cgraph.c (insert_new_cgraph_node_version): Likewise.
2770 (cgraph_allocate_node): Likewise.
2771 (cgraph_create_edge_1): Likewise.
2772 (cgraph_allocate_init_indirect_info): Likewise.
2773 * cgraphclones.c (cgraph_clone_edge): Likewise.
2774 * cgraphunit.c (add_asm_node): Likewise.
2775 (init_lowered_empty_function): Likewise.
2776 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
2777 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
2778 (alpha_use_linkage): Likewise.
2779 * config/arc/arc.c (arc_init_machine_status): Likewise.
2780 * config/arm/arm.c (arm_init_machine_status): Likewise.
2781 * config/avr/avr.c (avr_init_machine_status): Likewise.
2782 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
2783 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
2784 * config/cris/cris.c (cris_init_machine_status): Likewise.
2785 * config/darwin.c (machopic_indirection_name): Likewise.
2786 (darwin_build_constant_cfstring): Likewise.
2787 (darwin_enter_string_into_cfstring_table): Likewise.
2788 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
2789 * config/frv/frv.c (frv_init_machine_status): Likewise.
2790 * config/i386/i386.c (get_dllimport_decl): Likewise.
2791 (ix86_init_machine_status): Likewise.
2792 (assign_386_stack_local): Likewise.
2793 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
2794 (i386_pe_maybe_record_exported_symbol): Likewise.
2795 (i386_pe_record_stub): Likewise.
2796 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
2797 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
2798 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
2799 (m32c_note_pragma_address): Likewise.
2800 * config/mep/mep.c (mep_init_machine_status): Likewise.
2801 (mep_note_pragma_flag): Likewise.
2802 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
2803 (mips16_local_alias): Likewise.
2804 (mips_init_machine_status): Likewise.
2805 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
2806 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
2807 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
2808 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
2809 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
2810 * config/pa/pa.c (pa_init_machine_status): Likewise.
2811 (pa_get_deferred_plabel): Likewise.
2812 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
2813 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
2814 (rs6000_init_machine_status): Likewise.
2815 (output_toc): Likewise.
2816 * config/s390/s390.c (s390_init_machine_status): Likewise.
2817 * config/score/score.c (score_output_external): Likewise.
2818 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
2819 * config/spu/spu.c (spu_init_machine_status): Likewise.
2820 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
2821 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
2822 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
2823 * coverage.c (coverage_end_function): Likewise.
2824 * dbxout.c (dbxout_init): Likewise.
2825 * doc/gty.texi: Don't mention variable_size attribute.
2826 * dwarf2cfi.c (new_cfi): Adjust.
2827 (new_cfi_row): Likewise.
2828 (copy_cfi_row): Likewise.
2829 (create_cie_data): Likewise.
2830 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
2831 (new_loc_descr): Likewise.
2832 (find_AT_string_in_table): Likewise.
2833 (add_addr_table_entry): Likewise.
2834 (new_die): Likewise.
2835 (add_var_loc_to_decl): Likewise.
2836 (clone_die): Likewise.
2837 (clone_as_declaration): Likewise.
2838 (break_out_comdat_types): Likewise.
2839 (new_loc_list): Likewise.
2840 (add_loc_descr_to_each): Likewise.
2841 (add_location_or_const_value_attribute): Likewise.
2842 (add_linkage_name): Likewise.
2843 (lookup_filename): Likewise.
2844 (dwarf2out_var_location): Likewise.
2845 (new_line_info_table): Likewise.
2846 (dwarf2out_init): Likewise.
2847 (mem_loc_descriptor): Likewise.
2848 (loc_descriptor): Likewise.
2849 (add_const_value_attribute): Likewise.
2850 (tree_add_const_value_attribute): Likewise.
2851 (comp_dir_string): Likewise.
2852 (dwarf2out_vms_debug_main_pointer): Likewise.
2853 (string_cst_pool_decl): Likewise.
2854 * emit-rtl.c (set_mem_attrs): Likewise.
2855 (get_reg_attrs): Likewise.
2856 (start_sequence): Likewise.
2857 (init_emit): Likewise.
2858 (init_emit_regs): Likewise.
2859 * except.c (init_eh_for_function): Likewise.
2860 (gen_eh_region): Likewise.
2861 (gen_eh_region_catch): Likewise.
2862 (gen_eh_landing_pad): Likewise.
2863 (add_call_site): Likewise.
2864 * function.c (add_frame_space): Likewise.
2865 (insert_temp_slot_address): Likewise.
2866 (assign_stack_temp_for_type): Likewise.
2867 (get_hard_reg_initial_val): Likewise.
2868 (allocate_struct_function): Likewise.
2869 (prepare_function_start): Likewise.
2870 (types_used_by_var_decl_insert): Likewise.
2871 * gengtype.c (variable_size_p): Remove function.
2872 (enum alloc_quantity): Remove enum.
2873 (write_typed_alloc_def): Remove function.
2874 (write_typed_struct_alloc_def): Likewise.
2875 (write_typed_typedef_alloc_def): Likewise.
2876 (write_typed_alloc_defns): Likewise.
2878 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
2879 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
2880 * ggc.h (ggc_alloc): new function.
2881 (ggc_cleared_alloc): Likewise.
2882 (ggc_vec_alloc): Template on type of vector element, and remove
2883 element size argument.
2884 (ggc_cleared_vec_alloc): Likewise.
2885 * gimple.c (gimple_build_omp_for): Adjust.
2886 (gimple_copy): Likewise.
2887 * ipa-cp.c (get_replacement_map): Likewise.
2888 (find_aggregate_values_for_callers_subset): Likewise.
2889 (known_aggs_to_agg_replacement_list): Likewise.
2890 * ipa-devirt.c (get_odr_type): Likewise.
2891 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
2892 (read_agg_replacement_chain): Likewise.
2893 * loop-iv.c (get_simple_loop_desc): Likewise.
2894 * lto-cgraph.c (input_node_opt_summary): Likewise.
2895 * lto-section-in.c (lto_new_in_decl_state): Likewise.
2896 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
2897 (input_eh_region): Likewise.
2898 (input_eh_lp): Likewise.
2899 (input_cfg): Likewise.
2900 * optabs.c (set_optab_libfunc): Likewise.
2901 (init_tree_optimization_optabs): Likewise.
2902 (set_conv_libfunc): Likewise.
2903 * passes.c (do_per_function_toporder): Likewise.
2904 * rtl.h: Don't use variable_size gty attribute.
2905 * sese.c (if_region_set_false_region): Adjust.
2906 * stringpool.c (gt_pch_save_stringpool): Likewise.
2907 * target-globals.c (save_target_globals): Likewise.
2908 * toplev.c (general_init): Likewise.
2909 * trans-mem.c (record_tm_replacement): Likewise.
2910 (split_bb_make_tm_edge): Likewise.
2911 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2912 * tree-data-ref.h (lambda_vector_new): Likewise.
2913 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
2914 * tree-iterator.c (tsi_link_before): Likewise.
2915 (tsi_link_after): Likewise.
2916 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
2917 * tree-ssa-loop-niter.c (record_estimate): Likewise.
2918 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
2919 * tree-ssa-operands.h: Don't use variable_size gty attribute.
2920 * tree-ssa.c (init_tree_ssa): Adjust.
2921 * tree-ssanames.c (set_range_info): Likewise.
2922 (get_ptr_info): Likewise.
2923 (duplicate_ssa_name_ptr_info): Likewise.
2924 (duplicate_ssa_name_range_info): Likewise.
2925 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
2926 (unpack_ts_fixed_cst_value_fields): Likewise.
2927 * tree.c (build_fixed): Likewise.
2928 (build_real): Likewise.
2929 (build_string): Likewise.
2930 (decl_priority_info): Likewise.
2931 (decl_debug_expr_insert): Likewise.
2932 (decl_value_expr_insert): Likewise.
2933 (decl_debug_args_insert): Likewise.
2934 (type_hash_add): Likewise.
2935 (build_omp_clause): Likewise.
2936 * ubsan.c (decl_for_type_insert): Likewise.
2937 * varasm.c (get_unnamed_section): Likewise.
2938 (get_noswitch_section): Likewise.
2939 (get_section): Likewise.
2940 (get_block_for_section): Likewise.
2941 (create_block_symbol): Likewise.
2942 (build_constant_desc): Likewise.
2943 (create_constant_pool): Likewise.
2944 (force_const_mem): Likewise.
2945 (record_tm_clone_pair): Likewise.
2946 * varpool.c (varpool_create_empty_node): Likewise.
2948 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2950 * dwarf2out.c (tree_add_const_value_attribute): Call
2951 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
2952 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
2953 instead of ggc_internal_<x>alloc_stat.
2954 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
2955 (ggc_realloc): Likewise.
2956 * ggc-none.c (ggc_internal_alloc): Likewise.
2957 (ggc_internal_cleared_alloc): Likewise.
2958 * ggc-page.c: Likewise.
2959 * ggc.h (ggc_internal_alloc_stat): Likewise.
2960 (ggc_internal_alloc): Remove macro.
2961 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
2962 (ggc_internal_cleared_alloc): Remove macro.
2963 (GGC_RESIZEVEC): Adjust.
2964 (ggc_resizevar): Remove macro.
2965 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
2966 (ggc_internal_cleared_vec_alloc_stat): Likewise.
2967 (ggc_internal_vec_cleared_alloc): Remove macro.
2968 (ggc_alloc_atomic_stat): Drop _stat suffix.
2969 (ggc_alloc_atomic): Remove macro.
2970 (ggc_alloc_cleared_atomic): Remove macro.
2971 (ggc_alloc_string_stat): Drop _stat suffix.
2972 (ggc_alloc_string): Remove macro.
2973 (ggc_alloc_rtx_def_stat): Adjust.
2974 (ggc_alloc_tree_node_stat): Likewise.
2975 (ggc_alloc_cleared_tree_node_stat): Likewise.
2976 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
2977 (ggc_alloc_cleared_simd_clone_stat): Likewise.
2978 * gimple.c (gimple_build_omp_for): Likewise.
2979 (gimple_copy): Likewise.
2980 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
2981 * toplev.c (realloc_for_line_map): Adjust.
2982 * tree-data-ref.h (lambda_vector_new): Likewise.
2983 * tree-phinodes.c (allocate_phi_node): Likewise.
2984 * tree.c (grow_tree_vec_stat): Likewise.
2985 * vec.h (va_gc::reserve): Adjust.
2987 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
2989 * config/microblaze/microblaze.c (break_handler): New Declaration.
2990 (microblaze_break_function_p,microblaze_is_break_handler): New.
2991 (compute_frame_size): Use microblaze_break_function_p.
2992 Add the test of break_handler.
2993 (microblaze_function_prologue) : Add the test of variable
2994 break_handler. Check the fnname by BREAK_HANDLER_NAME.
2995 (microblaze_function_epilogue) : Add the test of break_handler.
2996 (microblaze_globalize_label) : Add the test of break_handler.
2997 Check the name by BREAK_HANDLER_NAME.
2999 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
3001 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
3002 microblaze_is_break_handler test.
3003 (call_internal1,call_value_intern): Use microblaze_break_function_p.
3004 Use SYMBOL_REF_DECL.
3006 * config/microblaze/microblaze-protos.h
3007 (microblaze_break_function_p,microblaze_is_break_handler):
3010 * doc/extend.texi (MicroBlaze break_handler Functions): Document
3011 new MicroBlaze break_handler functions.
3013 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
3015 * doc/extend.texi (Size of an asm): Move node text according
3016 to its @menu entry position.
3018 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
3020 PR tree-optimization/61140
3021 PR tree-optimization/61150
3022 PR tree-optimization/61197
3023 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
3025 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
3027 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
3029 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
3031 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
3032 __SIZEOF_INT128__ is defined.
3034 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
3036 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
3037 (rs6000_delegitimize_address): Use it.
3039 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
3041 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
3042 inplace argument. Store the new address in the original MEM when true.
3043 * emit-rtl.c (change_address_1): Likewise.
3044 (adjust_address_1, adjust_automodify_address_1, offset_address):
3046 * rtl.h (plus_constant): Add an inplace argument.
3047 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
3048 when true. Avoid generating (plus X (const_int 0)).
3049 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
3050 in-place. Pass true to plus_constant.
3051 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
3053 2014-05-16 Dehao Chen <dehao@google.com>
3055 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
3057 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
3060 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
3062 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
3064 2014-05-16 Dehao Chen <dehao@google.com>
3066 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
3067 optimize_function_for_size_p.
3068 * regs.h (REG_FREQ_FROM_BB): Likewise.
3070 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
3073 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
3074 negt_reg_operand cases.
3075 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
3077 * config/sh/predicates.md (cbranch_treg_value): Simplify.
3079 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
3081 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
3084 2014-05-16 David Malcolm <dmalcolm@redhat.com>
3087 2014-04-29 David Malcolm <dmalcolm@redhat.com>
3089 * tree-cfg.c (dump_function_to_file): Dump the return type of
3090 functions, in a line to itself before the function body, mimicking
3091 the layout of a C function.
3093 2014-05-16 Dehao Chen <dehao@google.com>
3095 * cfghooks.c (make_forwarder_block): Use direct computation to
3096 get fall-through edge's count and frequency.
3098 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
3100 * config/arc/arc.c (arc_init): Fix typo in error message.
3101 * config/i386/i386.c (ix86_expand_builtin): Likewise.
3102 (split_stack_prologue_scratch_regno): Likewise.
3103 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
3104 word from error message.
3106 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
3108 * ira-costs.c: Fix typo in comment.
3110 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
3112 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
3114 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
3116 * varpool.c (dump_varpool_node): Dump write-only flag.
3117 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
3119 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
3120 write-only variables.
3121 * ipa.c (process_references): New function.
3122 (set_readonly_bit): New function.
3123 (set_writeonly_bit): New function.
3124 (clear_addressable_bit): New function.
3125 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
3126 fix handling of aliases.
3127 * cgraph.h (struct varpool_node): Add writeonly flag.
3129 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
3131 PR rtl-optimization/60969
3132 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
3133 Calculate costs for this case.
3135 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
3137 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
3138 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
3140 2014-05-16 Richard Biener <rguenther@suse.de>
3142 PR tree-optimization/61194
3143 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
3144 bool patterns ending in a COND_EXPR.
3146 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3148 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
3150 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3152 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
3153 where we were unable to cost an RTX.
3155 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3157 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
3160 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3161 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3163 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
3165 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3166 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3168 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
3169 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
3171 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3172 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3174 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
3177 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3178 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3180 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
3183 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3184 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3186 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
3187 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
3189 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3190 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3192 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
3195 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3196 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3198 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
3199 ZERO_EXTEND and SIGN_EXTEND better.
3201 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3202 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3204 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
3207 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3208 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3210 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
3211 costs when costing loads and stores to memory.
3213 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3214 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
3216 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
3219 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3221 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
3223 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3224 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3226 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
3228 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
3230 (aarch64_rtx_mult_cost): New.
3231 (aarch64_rtx_costs): Use it, refactor as appropriate.
3233 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3234 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
3236 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
3237 emit instructions, return number of instructions which would
3239 (aarch64_add_constant): Update call to aarch64_build_constant.
3240 (aarch64_output_mi_thunk): Likewise.
3241 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
3244 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3246 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
3247 (TARGET_RTX_COSTS): Call it.
3249 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3251 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
3252 (cortexa57_vector_cost): Likewise.
3253 (cortexa57_tunings): Use them.
3255 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3257 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
3258 (cpu_addrcost_table): Use it.
3259 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
3260 (aarch64_address_cost): Rewrite using aarch64_classify_address,
3263 2014-05-16 Richard Biener <rguenther@suse.de>
3265 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
3266 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
3267 (visit_phi): Ignore edges marked as not executable.
3268 (class cond_dom_walker): New.
3269 (cond_dom_walker::before_dom_children): Value-number
3270 control statements and mark successor edges as not
3271 executable if possible.
3272 (run_scc_vn): First walk all control statements in
3273 dominator order, marking edges as not executable.
3274 * tree-inline.c (copy_edges_for_bb): Be not confused
3275 about random edge flags.
3277 2014-05-16 Richard Biener <rguenther@suse.de>
3279 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
3281 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
3284 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
3285 (__TM_simple_begin): Use it.
3286 (__TM_begin): Likewise.
3288 2014-05-15 Martin Jambor <mjambor@suse.cz>
3291 * ipa-prop.c (update_indirect_edges_after_inlining): Check
3292 type_preserved flag when the indirect edge is polymorphic.
3294 2014-05-15 Martin Jambor <mjambor@suse.cz>
3296 PR tree-optimization/61090
3297 * tree-sra.c (sra_modify_expr): Pass the current gsi to
3298 build_ref_for_model.
3300 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3302 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
3303 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
3305 2014-05-15 Jakub Jelinek <jakub@redhat.com>
3307 PR tree-optimization/61158
3308 * fold-const.c (fold_binary_loc): If X is zero-extended and
3309 shiftc >= prec, make sure zerobits is all ones instead of
3310 invoking undefined behavior.
3312 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3314 * regcprop.h: New file.
3315 * regcprop.c (skip_debug_insn_p): New decl.
3316 (replace_oldest_value_reg): Check skip_debug_insn_p.
3317 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
3318 * shrink-wrap.c: Include regcprop.h.
3319 (prepare_shrink_wrap): Call
3320 copyprop_hardreg_forward_bb_without_debug_insn.
3322 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3324 * shrink-wrap.h: Update comment.
3325 * shrink-wrap.c: Update comment.
3326 (next_block_for_reg): Rename to live_edge_for_reg.
3327 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
3328 (move_insn_for_shrink_wrap): Split live_edge.
3329 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
3331 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
3333 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
3335 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
3336 * config/sparc/sparc.md (fptype_ut699): New attribute.
3337 (in_branch_delay): Return false if -mfix-ut699 is specified and
3338 fptype_ut699 is set to single.
3339 (truncdfsf2): Add fptype_ut699 attribute.
3340 (fix_truncdfsi2): Likewise.
3341 (floatsisf2): Change fptype attribute.
3342 (fix_truncsfsi2): Likewise.
3343 (negtf2_notv9): Delete.
3344 (negtf2_v9): Likewise.
3345 (negtf2_hq): New instruction.
3346 (negtf2): New instruction and splitter.
3347 (negdf2_notv9): Rewrite.
3348 (abstf2_notv9): Delete.
3349 (abstf2_hq_v9): Likewise.
3350 (abstf2_v9): Likewise.
3351 (abstf2_hq): New instruction.
3352 (abstf2): New instruction and splitter.
3353 (absdf2_notv9): Rewrite.
3355 2014-05-14 Cary Coutant <ccoutant@google.com>
3358 * opts.c (common_handle_option): Don't special-case "-g".
3359 (set_debug_level): Default to at least level 2 with "-g".
3361 2014-05-14 DJ Delorie <dj@redhat.com>
3363 * config/msp430/msp430.c (msp430_builtin): Add
3364 MSP430_BUILTIN_DELAY_CYCLES.
3365 (msp430_init_builtins): Register void __delay_cycles(long long).
3366 (msp430_builtin_decl): Add it.
3367 (cg_magic_constant): New.
3368 (msp430_expand_delay_cycles): New.
3369 (msp430_expand_builtin): Call it.
3370 (msp430_print_operand_raw): Change integer printing from "int" to
3372 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
3373 (delay_cycles_start): New.
3374 (delay_cycles_end): New.
3375 (delay_cycles_32): New.
3376 (delay_cycles_32x): New.
3377 (delay_cycles_16): New.
3378 (delay_cycles_16x): New.
3379 (delay_cycles_2): New.
3380 (delay_cycles_1): New.
3381 * doc/extend.texi: Document __delay_cycles().
3383 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
3385 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
3386 length attribute computation.
3388 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
3391 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
3393 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
3396 * config/sparc/sparc.md: Fix types of low and high in DI constant
3397 splitter. Use gen_int_mode in some other splitters.
3399 2014-05-14 Martin Jambor <mjambor@suse.cz>
3402 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
3404 2014-05-14 James Norris <jnorris@codesourcery.com>
3406 * omp-low.c (expand_parallel_call): Remove shadow variable.
3407 (expand_omp_taskreg): Likewise.
3409 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
3411 * common/config/i386/i386-common.c
3412 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
3413 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
3414 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
3415 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
3416 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
3417 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
3418 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
3419 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
3420 xsavecintrin.h, xsavesintrin.h.
3421 (x86_64-*-*): Ditto.
3422 * config/i386/clflushoptintrin.h: New.
3423 * config/i386/xsavecintrin.h: Ditto.
3424 * config/i386/xsavesintrin.h: Ditto.
3425 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
3426 (bit_XSAVES): Ditto.
3427 (bit_XSAVES): Ditto.
3428 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
3429 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
3431 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
3432 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
3433 OPTION_MASK_ISA_XSAVES.
3434 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
3436 (PTA_CLFLUSHOPT) Define.
3437 (PTA_XSAVEC): Ditto.
3438 (PTA_XSAVES): Ditto.
3439 (ix86_option_override_internal): Handle new options.
3440 (ix86_valid_target_attribute_inner_p): Ditto.
3441 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
3442 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
3443 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
3444 (bdesc_special_args): Add __builtin_ia32_xsaves,
3445 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
3446 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
3447 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
3448 (ix86_expand_builtin): Handle new builtins.
3449 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
3450 (TARGET_CLFLUSHOPT_P): Ditto.
3451 (TARGET_XSAVEC): Ditto.
3452 (TARGET_XSAVEC_P): Ditto.
3453 (TARGET_XSAVES): Ditto.
3454 (TARGET_XSAVES_P): Ditto.
3455 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
3456 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
3457 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
3459 (ANY_XRSTOR64): Ditto.
3461 (xrstor): Change into <xrstor>.
3462 (xrstor_rex64): Change into <xrstor>_rex64.
3463 (xrstor64): Change into <xrstor>64
3465 * config/i386/i386.opt (mclflushopt): New.
3468 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
3470 * doc/invoke.texi: Document new options.
3472 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
3474 PR rtl-optimization/60866
3475 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
3476 Default it to -1. Pass it down to init_simplejump_data.
3477 (init_simplejump_data): New parameter old_seqno. Pass it down
3478 to get_seqno_for_a_jump.
3479 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
3480 initializing new jump seqno as a last resort. Add comment.
3481 (sel_redirect_edge_and_branch): Save old seqno of the conditional
3482 jump and pass it down to sel_init_new_insn.
3483 (sel_redirect_edge_and_branch_force): Likewise.
3485 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
3487 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
3488 shifted values to avoid build warning.
3490 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
3492 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
3493 * cfgrtl.c (rtl_merge_blocks): Fix comment.
3494 (cfg_layout_merge_blocks): Likewise.
3495 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
3497 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
3499 PR rtl-optimization/60901
3500 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
3501 bb predecessor belongs to the same scheduling region. Adjust comment.
3503 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
3505 * doc/sourcebuild.texi: (dfp_hw): Document.
3506 (p8vector_hw): Likewise.
3507 (powerpc_eabi_ok): Likewise.
3508 (powerpc_elfv2): Likewise.
3509 (powerpc_htm_ok): Likewise.
3510 (ppc_recip_hw): Likewise.
3513 2014-05-13 Cary Coutant <ccoutant@google.com>
3515 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
3517 2014-05-13 David Malcolm <dmalcolm@redhat.com>
3519 * gengtype-parse.c (require3): Eliminate in favor of...
3521 (require_template_declaration): Update to support optional single *
3524 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
3525 (create_user_defined_type): Handle a single level of explicit
3526 pointerness within template arguments.
3527 (struct write_types_data): Add field "kind".
3528 (filter_type_name): Handle "*" character.
3529 (write_user_func_for_structure_ptr): Require a write_types_data
3530 rather than just a prefix string, so that we can look up the kind
3531 of the wtd and use it as an index into wrote_user_func_for_ptr,
3532 ensuring that such functions are written at most once. Support
3533 subclasses by invoking the marking function of the ultimate base class.
3534 (write_user_func_for_structure_body): Require a write_types_data
3535 rather than just a prefix string, so that we can pass this to
3536 write_user_func_for_structure_ptr.
3537 (write_func_for_structure): Likewise.
3538 (ggc_wtd): Add initializer of new "kind" field.
3539 (pch_wtd): Likewise.
3541 * gengtype.h (enum write_types_kinds): New.
3542 (struct type): Add field wrote_user_func_for_ptr to the "s"
3545 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
3547 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
3548 instead of const_binop.
3549 (fold_binary_loc): Likewise.
3551 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
3553 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
3554 calculation to match get_ref_base_and_extent.
3556 2014-05-13 Catherine Moore <clm@codesourcery.com>
3557 Sandra Loosemore <sandra@codesourcery.com>
3559 * configure.ac: Fix assembly for explicit JALR relocation check.
3560 * configure: Regenerate.
3562 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3564 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
3565 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
3566 Remove associated type declarations and initialisations.
3567 (arm_expand_neon_builtin): Likewise.
3568 (neon_emit_pair_result_insn): Delete.
3569 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
3570 * config/arm/neon.md (neon_vtrn<mode>): Delete.
3571 (neon_vzip<mode>): Likewise.
3572 (neon_vuzp<mode>): Likewise.
3574 2014-05-13 Richard Biener <rguenther@suse.de>
3577 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
3578 it needs revisiting whether the call still may be tail-called.
3580 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3582 * rtl.def (SYMBOL_REF): Remove middle "0" field.
3583 * rtl.h (block_symbol): Reduce number of fields to 2.
3584 (rtx_def): Add u2.symbol_ref_flags.
3585 (SYMBOL_REF_FLAGS): Use it.
3586 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
3587 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
3588 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
3589 Lower index of SYMBOL_REF_DATA.
3590 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
3591 Print SYMBOL_REF_FLAGS at the same time.
3592 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
3594 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3596 * rtl.def (VAR_LOCATION): Remove "i" field.
3597 * rtl.h (rtx_def): Add u2.var_location_status.
3598 (PAT_VAR_LOCATION_STATUS): Use it.
3599 (gen_rtx_VAR_LOCATION): Declare.
3600 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
3601 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
3602 * var-tracking.c (emit_note_insn_var_location): Remove casts.
3604 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3606 * rtl.def (scratch): Fix outdated comment and remove "0" field.
3607 * gengtype.c (adjust_field_rtx_def): Update accordingly.
3609 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3611 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
3612 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
3613 * rtl.h (rtx_def): Add insn_uid to u2 field.
3614 (RTX_FLAG_CHECK8): Delete in favor of...
3615 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
3616 (INSN_DELETED_P): Update accordingly.
3617 (INSN_UID): Use u2.insn_uid.
3618 (INSN_CHAIN_CODE_P): Define.
3619 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
3620 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
3621 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
3622 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
3623 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
3624 indices accordingly.
3625 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
3626 Update indices for insn-chain rtxes.
3627 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
3628 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
3629 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
3630 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
3631 * combine.c (try_combine): Likewise.
3632 * ira.c (setup_prohibited_mode_move_regs): Likewise.
3634 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3636 * rtl.def (REG): Remove middle field.
3637 * rtl.h (rtx_def): Add orignal_regno to u2.
3638 (ORIGINAL_REGNO): Use it instead of field 1.
3639 (REG_ATTRS): Lower field index accordingly.
3640 * gengtype.c (adjust_field_rtx_def): Remove handling of
3641 ORIGINAL_REGNO. Move REG_ATTRS index down.
3642 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
3643 code that prints the REGNO.
3645 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3647 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
3650 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3652 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
3654 2014-05-13 Bin Cheng <bin.cheng@arm.com>
3656 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
3657 (alloc_iv): Lower base expressions containing ADDR_EXPR.
3659 2014-05-13 Ian Bolton <ian.bolton@arm.com>
3661 * config/aarch64/aarch64-protos.h
3662 (aarch64_hard_regno_caller_save_mode): New prototype.
3663 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
3665 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
3667 2014-05-13 Christian Bruel <christian.bruel@st.com>
3669 * target.def (mode_switching): New hook vector.
3670 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
3671 (mode_exit, modepriority_to_mode): Likewise.
3672 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
3673 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3674 * target.h: Include tm.h and hard-reg-set.h.
3675 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
3676 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
3677 * doc/tm.texi Regenerate.
3678 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
3679 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3680 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
3681 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
3682 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
3683 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3684 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
3685 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
3686 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
3687 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
3688 (ix86_emit_mode_set): Hookify.
3689 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
3691 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3692 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
3693 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
3694 (epiphany_mode_priority_to_mode): Remove declaration.
3695 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
3696 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
3697 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
3699 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
3700 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
3701 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
3703 2014-05-13 Jakub Jelinek <jakub@redhat.com>
3706 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
3707 is const0_rtx, return immediately. Don't test count == 0 when
3710 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3712 * Makefile.in: add shrink-wrap.o.
3713 * config/i386/i386.c: include "shrink-wrap.h"
3714 * function.c: Likewise.
3715 (requires_stack_frame_p, next_block_for_reg,
3716 move_insn_for_shrink_wrap, prepare_shrink_wrap,
3717 dup_block_and_redirect): Move to shrink-wrap.c
3718 (thread_prologue_and_epilogue_insns): Extract three code segments
3719 as functions in shrink-wrap.c
3720 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
3722 * shrink-wrap.c: New file.
3723 * shrink-wrap.h: New file.
3725 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3727 * doc/extend.texi: Reflect current numbers of pragmas. Remove
3728 reference to Solaris.
3730 2014-05-12 Mike Stump <mikestump@comcast.net>
3733 * genattrtab.c (filename): Add.
3734 (convert_set_attr_alternative): Improve error message.
3735 (check_defs): Restore read_md_filename for error messages.
3736 (gen_insn): Save filename.
3738 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
3740 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
3741 -fno-local-ivars and -fivar-visibility.
3742 * c-family/c.opt: Make -Wshadow also implicitly enable
3745 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3747 * doc/tm.texi: Remove reference to deleted macro.
3748 * doc/tm.texi.in: Likewise.
3750 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
3753 * config/avr/avr.c (avr_out_store_psi): Use correct constant
3756 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
3759 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
3760 * config/arm/aout.h (License): Same.
3761 * config/arm/bpabi.h (License): Same.
3762 * config/arm/elf.h (License): Same.
3763 * config/arm/linux-elf.h (License): Same.
3764 * config/arm/linux-gas.h (License): Same.
3765 * config/arm/netbsd-elf.h (License): Same.
3766 * config/arm/uclinux-eabi.h (License): Same.
3767 * config/arm/uclinux-elf.h (License): Same.
3768 * config/arm/vxworks.h (License): Same.
3770 2014-05-11 Jakub Jelinek <jakub@redhat.com>
3772 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
3773 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
3774 number of operands to 3.
3775 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
3776 * tree-nested.c (convert_nonlocal_omp_clauses,
3777 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
3778 * gimplify.c (gimplify_scan_omp_clauses): Handle
3779 OMP_CLAUSE_LINEAR_STMT.
3780 * omp-low.c (lower_rec_input_clauses): Fix typo.
3781 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
3782 cast between Fortran boolean_type_node and C _Bool if
3785 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
3787 PR tree-optimization/61136
3788 * wide-int.h (multiple_of_p): Define a version that doesn't return
3790 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
3791 integer_zerop/const_binop pair.
3792 (multiple_of_p): Likewise, converting both operands to widest_int
3795 2014-05-09 Teresa Johnson <tejohnson@google.com>
3797 * cgraphunit.c (analyze_functions): Use correct dump file.
3799 2014-05-09 Florian Weimer <fweimer@redhat.com>
3801 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
3803 (stack_protect_return_slot_p): New function.
3804 (expand_used_vars): Call stack_protect_decl_p and
3805 stack_protect_return_slot_p for -fstack-protector-strong.
3807 2014-05-09 David Wohlferd <LimeGreenSocks@yahoo.com>
3808 Andrew Haley <aph@redhat.com>
3809 Richard Sandiford <rdsandiford@googlemail.com>
3811 * doc/extend.texi: Rewrite inline asm page / re-org asm-related
3814 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
3817 * fold-const.c (fold_binary_loc): Changed width of mask.
3819 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3821 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
3822 unsigned int initializers for regno_in, regno_out.
3824 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3827 * config/avr/avr.md (cc): Add new attribute set_vzn.
3828 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
3829 Set cc insn attribute to set_vzn instead of set_zn for alternatives
3830 with INC, DEC or NEG.
3831 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
3832 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
3833 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
3835 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3838 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3840 * wide-int.cc (UTItype): Define.
3841 (UDWtype): Define for appropriate W_TYPE_SIZE.
3843 2014-05-09 Richard Biener <rguenther@suse.de>
3845 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
3846 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
3847 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
3848 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
3849 ssa_propagate): Adjust.
3851 2014-05-08 Jeff Law <law@redhat.com>
3853 PR tree-optimization/61009
3854 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
3855 tri-state rather than a boolean. When a block is too big to
3856 thread through, inform caller via negative return value.
3857 (thread_across_edge): If a block was too big for normal threading,
3858 then it's too big for a joiner too, so remove temporary equivalences
3859 and return immediately.
3861 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
3862 Matthias Klose <doko@ubuntu.com>
3865 * optc-gen.awk: Fix option handling for -Wunused-parameter.
3867 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3870 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
3872 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3875 * config/alpha/alpha.c: Include gimple-iterator.h.
3876 (alpha_gimple_fold_builtin): New function. Move
3877 ALPHA_BUILTIN_UMULH folding from ...
3878 (alpha_fold_builtin): ... here.
3879 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
3881 2014-05-08 Wei Mi <wmi@google.com>
3884 * config/i386/i386.c (ix86_compute_frame_layout): Update
3885 preferred_stack_boundary for call, expanded from tls descriptor.
3886 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
3887 to depend on SP register.
3888 (*tls_local_dynamic_base_32_gnu): Ditto.
3889 (*tls_local_dynamic_32_once): Ditto.
3890 (tls_global_dynamic_64_<mode>): Set
3891 ix86_tls_descriptor_calls_expanded_in_cfun.
3892 (tls_local_dynamic_base_64_<mode>): Ditto.
3893 (tls_global_dynamic_32): Set
3894 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
3895 to depend on SP register.
3896 (tls_local_dynamic_base_32): Ditto.
3898 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3900 * config/arm/arm_neon.h: Update comment.
3901 * config/arm/neon-docgen.ml: Delete.
3902 * config/arm/neon-gen.ml: Delete.
3903 * doc/arm-neon-intrinsics.texi: Update comment.
3905 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3907 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
3909 (vand, vorr, veor, vorn, vbic): Remove.
3910 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
3912 (neon_vsub_unspec): Likewise.
3913 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
3915 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3917 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
3918 (vadd_s16): Likewise.
3919 (vadd_s32): Likewise.
3920 (vadd_f32): Likewise.
3921 (vadd_u8): Likewise.
3922 (vadd_u16): Likewise.
3923 (vadd_u32): Likewise.
3924 (vadd_s64): Likewise.
3925 (vadd_u64): Likewise.
3926 (vaddq_s8): Likewise.
3927 (vaddq_s16): Likewise.
3928 (vaddq_s32): Likewise.
3929 (vaddq_s64): Likewise.
3930 (vaddq_f32): Likewise.
3931 (vaddq_u8): Likewise.
3932 (vaddq_u16): Likewise.
3933 (vaddq_u32): Likewise.
3934 (vaddq_u64): Likewise.
3935 (vmul_s8): Likewise.
3936 (vmul_s16): Likewise.
3937 (vmul_s32): Likewise.
3938 (vmul_f32): Likewise.
3939 (vmul_u8): Likewise.
3940 (vmul_u16): Likewise.
3941 (vmul_u32): Likewise.
3942 (vmul_p8): Likewise.
3943 (vmulq_s8): Likewise.
3944 (vmulq_s16): Likewise.
3945 (vmulq_s32): Likewise.
3946 (vmulq_f32): Likewise.
3947 (vmulq_u8): Likewise.
3948 (vmulq_u16): Likewise.
3949 (vmulq_u32): Likewise.
3950 (vsub_s8): Likewise.
3951 (vsub_s16): Likewise.
3952 (vsub_s32): Likewise.
3953 (vsub_f32): Likewise.
3954 (vsub_u8): Likewise.
3955 (vsub_u16): Likewise.
3956 (vsub_u32): Likewise.
3957 (vsub_s64): Likewise.
3958 (vsub_u64): Likewise.
3959 (vsubq_s8): Likewise.
3960 (vsubq_s16): Likewise.
3961 (vsubq_s32): Likewise.
3962 (vsubq_s64): Likewise.
3963 (vsubq_f32): Likewise.
3964 (vsubq_u8): Likewise.
3965 (vsubq_u16): Likewise.
3966 (vsubq_u32): Likewise.
3967 (vsubq_u64): Likewise.
3968 (vand_s8): Likewise.
3969 (vand_s16): Likewise.
3970 (vand_s32): Likewise.
3971 (vand_u8): Likewise.
3972 (vand_u16): Likewise.
3973 (vand_u32): Likewise.
3974 (vand_s64): Likewise.
3975 (vand_u64): Likewise.
3976 (vandq_s8): Likewise.
3977 (vandq_s16): Likewise.
3978 (vandq_s32): Likewise.
3979 (vandq_s64): Likewise.
3980 (vandq_u8): Likewise.
3981 (vandq_u16): Likewise.
3982 (vandq_u32): Likewise.
3983 (vandq_u64): Likewise.
3984 (vorr_s8): Likewise.
3985 (vorr_s16): Likewise.
3986 (vorr_s32): Likewise.
3987 (vorr_u8): Likewise.
3988 (vorr_u16): Likewise.
3989 (vorr_u32): Likewise.
3990 (vorr_s64): Likewise.
3991 (vorr_u64): Likewise.
3992 (vorrq_s8): Likewise.
3993 (vorrq_s16): Likewise.
3994 (vorrq_s32): Likewise.
3995 (vorrq_s64): Likewise.
3996 (vorrq_u8): Likewise.
3997 (vorrq_u16): Likewise.
3998 (vorrq_u32): Likewise.
3999 (vorrq_u64): Likewise.
4000 (veor_s8): Likewise.
4001 (veor_s16): Likewise.
4002 (veor_s32): Likewise.
4003 (veor_u8): Likewise.
4004 (veor_u16): Likewise.
4005 (veor_u32): Likewise.
4006 (veor_s64): Likewise.
4007 (veor_u64): Likewise.
4008 (veorq_s8): Likewise.
4009 (veorq_s16): Likewise.
4010 (veorq_s32): Likewise.
4011 (veorq_s64): Likewise.
4012 (veorq_u8): Likewise.
4013 (veorq_u16): Likewise.
4014 (veorq_u32): Likewise.
4015 (veorq_u64): Likewise.
4016 (vbic_s8): Likewise.
4017 (vbic_s16): Likewise.
4018 (vbic_s32): Likewise.
4019 (vbic_u8): Likewise.
4020 (vbic_u16): Likewise.
4021 (vbic_u32): Likewise.
4022 (vbic_s64): Likewise.
4023 (vbic_u64): Likewise.
4024 (vbicq_s8): Likewise.
4025 (vbicq_s16): Likewise.
4026 (vbicq_s32): Likewise.
4027 (vbicq_s64): Likewise.
4028 (vbicq_u8): Likewise.
4029 (vbicq_u16): Likewise.
4030 (vbicq_u32): Likewise.
4031 (vbicq_u64): Likewise.
4032 (vorn_s8): Likewise.
4033 (vorn_s16): Likewise.
4034 (vorn_s32): Likewise.
4035 (vorn_u8): Likewise.
4036 (vorn_u16): Likewise.
4037 (vorn_u32): Likewise.
4038 (vorn_s64): Likewise.
4039 (vorn_u64): Likewise.
4040 (vornq_s8): Likewise.
4041 (vornq_s16): Likewise.
4042 (vornq_s32): Likewise.
4043 (vornq_s64): Likewise.
4044 (vornq_u8): Likewise.
4045 (vornq_u16): Likewise.
4046 (vornq_u32): Likewise.
4047 (vornq_u64): Likewise.
4049 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4051 * wide-int.cc (UTItype): Define.
4052 (UDWtype): Define for appropriate W_TYPE_SIZE.
4054 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
4056 PR tree-optimization/59100
4057 * tree-ssa-phiopt.c: Include tree-inline.h.
4058 (neutral_element_p, absorbing_element_p): New functions.
4059 (value_replacement): Handle conditional binary operations with a
4060 neutral or absorbing element.
4062 2014-05-08 Richard Biener <rguenther@suse.de>
4064 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
4065 folding the expression.
4066 (valueize_expr): Remove.
4067 (visit_reference_op_load): Do not valueize the result of
4069 (simplify_binary_expression): Likewise.
4070 (simplify_unary_expression): Likewise.
4072 2014-05-08 Richard Biener <rguenther@suse.de>
4074 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
4075 looking at TYPE_ARG_TYPES.
4077 2014-05-08 Richard Biener <rguenther@suse.de>
4079 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
4080 pointer propagation special-case.
4082 2014-05-08 Bin Cheng <bin.cheng@arm.com>
4084 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
4085 core part of address expressions.
4087 2014-05-08 Alan Modra <amodra@gmail.com>
4090 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
4091 loads and stores when -mno-strict-align at any alignment.
4092 (expand_block_clear): Similarly. Also correct calculation of
4095 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
4098 * tree-complex.c (expand_complex_move): Keep line info when expanding
4100 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
4101 of complex expression. Use new argument to display correct location
4102 for values coming from phi statement.
4103 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
4104 (warn_uninitialized_phi): Pass location of phi argument to
4106 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
4107 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
4109 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
4111 * config/rs6000/predicates.md (indexed_address_mem): New.
4112 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
4113 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
4114 fpstore_ux, fpstore_u.
4115 (sign_extend, indexed, update): New.
4116 (cell_micro): Adjust.
4117 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
4118 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
4119 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
4120 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
4121 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
4122 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
4123 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
4124 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
4125 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
4126 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
4127 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
4128 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
4129 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
4130 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
4131 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
4133 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
4134 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
4135 *vsx_extract_<mode>_store): Adjust.
4136 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
4137 is_cracked_insn, insn_must_be_first_in_group,
4138 insn_must_be_last_in_group): Adjust.
4140 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
4142 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
4143 ppc440-fpstore): Adjust.
4144 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
4145 ppc476-fpstore): Adjust.
4146 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
4147 ppc601-fpstore): Adjust.
4148 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
4150 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
4152 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
4153 ppc7450-fpstore): Adjust.
4154 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
4155 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
4156 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
4158 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
4159 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
4160 cell-fpstore, cell-fpstore-update): Adjust.
4161 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
4162 ppce300c3_store, ppce300c3_fpstore): Adjust.
4163 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
4164 e500mc_fpstore): Adjust.
4165 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
4166 e500mc64_store, e500mc64_fpstore): Adjust.
4167 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
4168 e5500_fpstore): Adjust.
4169 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
4170 e6500_fpstore): Adjust.
4171 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
4173 * config/rs6000/power4.md (power4-load, power4-load-ext,
4174 power4-load-ext-update, power4-load-ext-update-indexed,
4175 power4-load-update-indexed, power4-load-update, power4-fpload,
4176 power4-fpload-update, power4-store, power4-store-update,
4177 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
4179 * config/rs6000/power5.md (power5-load, power5-load-ext,
4180 power5-load-ext-update, power5-load-ext-update-indexed,
4181 power5-load-update-indexed, power5-load-update, power5-fpload,
4182 power5-fpload-update, power5-store, power5-store-update,
4183 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
4185 * config/rs6000/power6.md (power6-load, power6-load-ext,
4186 power6-load-update, power6-load-update-indexed,
4187 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
4188 power6-fpload-update, power6-store, power6-store-update,
4189 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
4191 * config/rs6000/power7.md (power7-load, power7-load-ext,
4192 power7-load-update, power7-load-update-indexed,
4193 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
4194 power7-fpload-update, power7-store, power7-store-update,
4195 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
4197 * config/rs6000/power8.md (power8-load, power8-load-update,
4198 power8-load-ext, power8-load-ext-update, power8-fpload,
4199 power8-fpload-update, power8-store, power8-store-update-indexed,
4200 power8-fpstore, power8-fpstore-update): Adjust.
4201 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
4203 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
4204 titan_lsu_store, titan_lsu_fpstore): Adjust.
4205 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
4207 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
4210 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
4211 unrolled byte insns. Emit address increments after move insns.
4213 2014-05-07 David Malcolm <dmalcolm@redhat.com>
4215 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
4216 const_gimple, rather than a gimple.
4217 (gimple_call_builtin_p): Likewise, for the three variants.
4219 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
4220 (gimple_call_builtin_p): Likewise, for the three variants.
4222 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4224 PR tree-optimization/61095
4225 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
4227 2014-05-07 Richard Biener <rguenther@suse.de>
4229 PR tree-optimization/61034
4230 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
4231 (maybe_skip_until): Use translate to take into account
4232 lattices when trying to do disambiguations.
4233 (get_continuation_for_phi_1): Likewise.
4234 (get_continuation_for_phi): Adjust for added translate arguments.
4235 (walk_non_aliased_vuses): Likewise.
4236 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
4237 (walk_non_aliased_vuses): Likewise.
4238 (call_may_clobber_ref_p_1): Declare.
4239 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
4240 calls. Stop early if we are only supposed to disambiguate.
4241 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
4243 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
4245 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
4246 Emit an error when the function has arguments.
4248 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
4250 * cfgloop.h (unswitch_loops): Remove.
4251 * doc/passes.texi: Remove references to loop-unswitch.c
4252 * timevar.def (TV_LOOP_UNSWITCH): Remove.
4254 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
4256 * tree-vect-data-refs.c (vect_grouped_load_supported): New
4257 check for loads group of length 3.
4258 (vect_permute_load_chain): New permutations for loads group of
4260 * tree-vect-stmts.c (vect_model_load_cost): Change cost
4261 of vec_perm_shuffle for the new permutations.
4263 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
4265 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
4266 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
4267 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
4268 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
4269 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
4270 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
4271 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
4272 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
4274 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
4276 * loop-unswitch.c: Delete.
4278 2014-05-07 Richard Biener <rguenther@suse.de>
4280 * config.gcc: Always set need_64bit_hwint to yes.
4282 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
4284 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
4285 of using optimize_size.
4287 2014-05-06 Mike Stump <mikestump@comcast.net>
4289 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
4291 2014-05-06 Joseph Myers <joseph@codesourcery.com>
4293 * config/i386/sse.md (*mov<mode>_internal)
4294 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
4295 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
4296 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
4297 (*<code><mode>3, *andnot<mode>3<mask_name>)
4298 (<mask_codefor><code><mode>3<mask_name>): Only consider
4299 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
4301 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
4304 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4306 * lra-constraints.c (valid_address_p): Move earlier in file.
4307 Add a constraint argument to the address_info version.
4308 (satisfies_memory_constraint_p): New function.
4309 (satisfies_address_constraint_p): Likewise.
4310 (process_alt_operands, curr_insn_transform): Use them.
4311 (process_address): Pass the constraint to valid_address_p when
4312 checking address operands.
4314 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
4316 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
4317 to their respective blocks. Fix inadvertent use of "node".
4319 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
4321 * emit-rtl.c (init_derived_machine_modes): New functionm, split
4323 (init_emit_once): ...here.
4324 * rtl.h (init_derived_machine_modes): Declare.
4325 * toplev.c (do_compile): Call it even if no_backend.
4327 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
4328 Mike Stump <mikestump@comcast.net>
4329 Richard Sandiford <rdsandiford@googlemail.com>
4330 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4332 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
4333 (rtx_equal_for_memref_p): Update comment.
4334 (adjust_offset_for_component_ref): Use wide-int interfaces.
4335 * builtins.c (get_object_alignment_2): Likewise.
4336 (c_readstr): Likewise.
4337 (target_char_cast): Add comment.
4338 (determine_block_size): Use wide-int interfaces.
4339 (expand_builtin_signbit): Likewise.
4340 (fold_builtin_int_roundingfn): Likewise.
4341 (fold_builtin_bitop): Likewise.
4342 (fold_builtin_bswap): Likewise.
4343 (fold_builtin_logarithm): Use signop.
4344 (fold_builtin_pow): Likewise.
4345 (fold_builtin_memory_op): Use wide-int interfaces.
4346 (fold_builtin_object_size): Likewise.
4347 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
4348 nb_iterations_estimate.
4349 (record_niter_bound): Use wide-int interfaces.
4350 (get_estimated_loop_iterations_int): Likewise.
4351 (get_estimated_loop_iterations): Likewise.
4352 (get_max_loop_iterations): Likewise.
4353 * cfgloop.h: Include wide-int.h.
4354 (struct nb_iter_bound): Change bound to widest_int.
4355 (struct loop): Change nb_iterations_upper_bound and
4356 nb_iterations_estimate to widest_int.
4357 (record_niter_bound): Switch to use widest_int.
4358 (get_estimated_loop_iterations): Likewise.
4359 (get_max_loop_iterations): Likewise.
4360 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
4361 update for wide-int.
4362 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
4363 * combine.c (try_combine): Likewise.
4364 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
4365 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
4367 (aarch64_float_const_representable_p): Likewise.
4368 * config/arc/arc.c: Include wide-int.h.
4369 (arc_can_use_doloop_p): Use wide-int interfaces.
4370 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
4371 (vfp3_const_double_index): Likewise.
4372 * config/avr/avr.c (avr_out_round): Likewise.
4373 (avr_fold_builtin): Likewise.
4374 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
4375 (bfin_can_use_doloop_p): Likewise.
4376 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
4377 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
4378 * config/i386/i386.c: Include wide-int.h.
4379 (ix86_data_alignment): Use wide-int interfaces.
4380 (ix86_local_alignment): Likewise.
4381 (ix86_emit_swsqrtsf): Update real_from_integer.
4382 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
4383 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
4384 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
4385 (zero_constant): Likewise.
4386 (input_operand): Likewise.
4387 (splat_input_operand): Likewise.
4388 (non_logical_cint_operand): Change const_double to const_wide_int.
4389 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
4390 (easy_altivec_constant): Remove comment.
4391 (paired_expand_vector_init): Use CONSTANT_P.
4392 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
4393 (rs6000_emit_move): Update checks.
4394 (rs6000_aggregate_candidate): Use wide-int interfaces.
4395 (rs6000_expand_ternop_builtin): Likewise.
4396 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
4397 (rs6000_assemble_integer): Likewise.
4398 (rs6000_hash_constant): Likewise.
4399 (output_toc): Likewise.
4400 (rs6000_rtx_costs): Likewise.
4401 (rs6000_emit_swrsqrt); Update call to real_from_integer.
4402 * config/rs6000/rs6000-c.c: Include wide-int.h.
4403 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
4404 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
4405 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
4406 Handle CONST_WIDE_INT.
4407 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
4408 Use tree_fits_uhwi_p.
4409 * config/sparc/sparc.c: Include wide-int.h.
4410 (sparc_fold_builtin): Use wide-int interfaces.
4411 * config/vax/vax.c: Include wide-int.h.
4412 (vax_float_literal): Use real_from_integer.
4413 * coretypes.h (struct hwivec_def): New.
4415 (const_hwivec): New.
4416 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
4417 (equiv_constant): Handle CONST_WIDE_INT.
4418 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
4419 (cselib_hash_rtx): Handle CONST_WIDE_INT.
4420 * dbxout.c (stabstr_U): Use wide-int interfaces.
4421 (dbxout_type): Update to use cst_fits_shwi_p.
4422 * defaults.h (LOG2_BITS_PER_UNIT): Define.
4423 (TARGET_SUPPORTS_WIDE_INT): Add default.
4424 * dfp.c: Include wide-int.h.
4425 (decimal_real_to_integer2): Use wide-int interfaces and rename to
4426 decimal_real_to_integer.
4427 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
4428 decimal_real_to_integer.
4429 * doc/generic.texi (Constant expressions): Update for wide_int.
4430 * doc/rtl.texi (const_double): Likewise.
4431 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
4432 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
4433 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
4434 (REAL_VALUE_FROM_INT): Remove.
4435 (TARGET_SUPPORTS_WIDE_INT): New.
4436 * doc/tm.texi: Regenerate.
4437 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
4438 * double-int.h: Include wide-int.h.
4439 (struct wi::int_traits): New.
4440 * dwarf2out.c (get_full_len): New.
4441 (dw_val_equal_p): Add case dw_val_class_wide_int.
4442 (size_of_loc_descr): Likewise.
4443 (output_loc_operands): Likewise.
4444 (insert_double): Remove.
4445 (insert_wide_int): New.
4447 (print_die): Add case dw_val_class_wide_int.
4448 (attr_checksum): Likewise.
4449 (attr_checksum_ordered): Likewise.
4450 (same_dw_val_p): Likewise.
4451 (size_of_die): Likewise.
4452 (value_format): Likewise.
4453 (output_die): Likewise.
4454 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
4456 (clz_loc_descriptor): Use wide-int interfaces.
4457 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
4458 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
4459 (round_up_to_align): Use wide-int interfaces.
4460 (field_byte_offset): Likewise.
4461 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
4462 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
4463 CONST_DOUBLE handling. Use wide-int interfaces.
4464 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
4465 (gen_enumeration_type_die): Use add_AT_wide.
4466 (hash_loc_operands): Add case dw_val_class_wide_int.
4467 (compare_loc_operands): Likewise.
4468 * dwarf2out.h: Include wide-int.h.
4469 (wide_int_ptr): New.
4470 (enum dw_val_class): Add dw_val_class_wide_int.
4471 (struct dw_val_struct): Add val_wide.
4472 * emit-rtl.c (const_wide_int_htab): New.
4473 (const_wide_int_htab_hash): New.
4474 (const_wide_int_htab_eq): New.
4475 (lookup_const_wide_int): New.
4476 (const_double_htab_hash): Use wide-int interfaces.
4477 (const_double_htab_eq): Likewise.
4478 (rtx_to_double_int): Conditionally compile for wide-int.
4479 (immed_double_int_const): Rename to immed_wide_int_const and
4480 update for wide-int.
4481 (immed_double_const): Conditionally compile for wide-int.
4482 (init_emit_once): Use wide-int interfaces.
4483 * explow.c (plus_constant): Likewise.
4484 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
4485 (lshift_value): Use wide-int interfaces.
4486 (expand_mult): Likewise.
4487 (choose_multiplier): Likewise.
4488 (expand_smod_pow2): Likewise.
4489 (make_tree): Likewise.
4490 * expr.c (convert_modes): Consolidate handling of constants.
4491 Use wide-int interfaces.
4492 (emit_group_load_1): Add note.
4493 (store_expr): Update comment.
4494 (get_inner_reference): Use wide-int interfaces.
4495 (expand_constructor): Update comment.
4496 (expand_expr_real_2): Use wide-int interfaces.
4497 (expand_expr_real_1): Likewise.
4498 (reduce_to_bit_field_precision): Likewise.
4499 (const_vector_from_tree): Likewise.
4500 * final.c: Include wide-int-print.h.
4501 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
4502 * fixed-value.c: Include wide-int.h.
4503 (fixed_from_string): Use wide-int interfaces.
4504 (fixed_to_decimal): Likewise.
4505 (fixed_convert_from_real): Likewise.
4506 (real_convert_from_fixed): Likewise.
4507 * fold-const.h (mem_ref_offset): Return an offset_int.
4508 (div_if_zero_remainder): Remove code parameter.
4509 * fold-const.c (div_if_zero_remainder): Remove code parameter.
4510 Use wide-int interfaces.
4511 (may_negate_without_overflow_p): Use wide-int interfaces.
4512 (negate_expr_p): Likewise.
4513 (fold_negate_expr): Likewise.
4514 (int_const_binop_1): Likewise.
4515 (const_binop): Likewise.
4516 (fold_convert_const_int_from_int): Likewise.
4517 (fold_convert_const_int_from_real): Likewise.
4518 (fold_convert_const_int_from_fixed): Likewise.
4519 (fold_convert_const_fixed_from_int): Likewise.
4520 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
4521 (sign_bit_p): Use wide-int interfaces.
4522 (make_range_step): Likewise.
4523 (build_range_check): Likewise. Pass an integer of the correct type
4524 instead of using integer_one_node.
4525 (range_predecessor): Pass an integer of the correct type instead
4526 of using integer_one_node.
4527 (range_successor): Likewise.
4528 (merge_ranges): Likewise.
4529 (unextend): Use wide-int interfaces.
4530 (extract_muldiv_1): Likewise.
4531 (fold_div_compare): Likewise.
4532 (fold_single_bit_test): Likewise.
4533 (fold_sign_changed_comparison): Likewise.
4534 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
4535 (fold_plusminus_mult_expr): Use wide-int interfaces.
4536 (native_encode_int): Likewise.
4537 (native_interpret_int): Likewise.
4538 (fold_unary_loc): Likewise.
4539 (pointer_may_wrap_p): Likewise.
4540 (size_low_cst): Likewise.
4541 (mask_with_tz): Likewise.
4542 (fold_binary_loc): Likewise.
4543 (fold_ternary_loc): Likewise.
4544 (multiple_of_p): Likewise.
4545 (tree_call_nonnegative_warnv_p): Update calls to
4546 tree_int_cst_min_precision and real_from_integer.
4547 (fold_negate_const): Use wide-int interfaces.
4548 (fold_abs_const): Likewise.
4549 (fold_relational_const): Use tree_int_cst_lt.
4550 (round_up_loc): Use wide-int interfaces.
4551 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
4552 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
4553 * gengtype.c: Remove include of double-int.h.
4554 (do_typedef): Use wide-int interfaces.
4555 (open_base_files): Add wide-int.h.
4556 (main): Add offset_int and widest_int typedefs.
4557 * gengtype-lex.l: Handle "^".
4558 (CXX_KEYWORD): Add "static".
4559 * gengtype-parse.c (require3): New.
4560 (require_template_declaration): Handle constant template arguments
4561 and nested templates.
4562 * gengtype-state.c: Don't include "double-int.h".
4563 * genpreds.c (write_one_predicate_function): Update comment.
4564 (write_tm_constrs_h): Add check for hval and lval use in
4566 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
4567 (add_to_sequence): Likewise.
4568 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
4569 and const_double_operand.
4570 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
4572 * gimple-fold.c (get_base_constructor): Likewise.
4573 (fold_array_ctor_reference): Likewise.
4574 (fold_nonarray_ctor_reference): Likewise.
4575 (fold_const_aggregate_ref_1): Likewise.
4576 (gimple_val_nonnegative_real_p): Likewise.
4577 (gimple_fold_indirect_ref): Likewise.
4578 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
4579 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
4580 (struct slsr_cand_d): Change index to be widest_int.
4581 (struct incr_info_d): Change incr to be widest_int.
4582 (alloc_cand_and_find_basis): Use wide-int interfaces.
4583 (slsr_process_phi): Likewise.
4584 (backtrace_base_for_ref): Likewise. Return a widest_int.
4585 (restructure_reference): Take a widest_int instead of a double_int.
4586 (slsr_process_ref): Use wide-int interfaces.
4587 (create_mul_ssa_cand): Likewise.
4588 (create_mul_imm_cand): Likewise.
4589 (create_add_ssa_cand): Likewise.
4590 (create_add_imm_cand): Take a widest_int instead of a double_int.
4591 (slsr_process_add): Use wide-int interfaces.
4592 (slsr_process_cast): Likewise.
4593 (slsr_process_copy): Likewise.
4594 (dump_candidate): Likewise.
4595 (dump_incr_vec): Likewise.
4596 (replace_ref): Likewise.
4597 (cand_increment): Likewise. Return a widest_int.
4598 (cand_abs_increment): Likewise.
4599 (replace_mult_candidate): Take a widest_int instead of a double_int.
4600 (replace_unconditional_candidate): Use wide-int interfaces.
4601 (incr_vec_index): Take a widest_int instead of a double_int.
4602 (create_add_on_incoming_edge): Likewise.
4603 (create_phi_basis): Use wide-int interfaces.
4604 (replace_conditional_candidate): Likewise.
4605 (record_increment): Take a widest_int instead of a double_int.
4606 (record_phi_increments): Use wide-int interfaces.
4607 (phi_incr_cost): Take a widest_int instead of a double_int.
4608 (lowest_cost_path): Likewise.
4609 (total_savings): Likewise.
4610 (analyze_increments): Use wide-int interfaces.
4611 (ncd_with_phi): Take a widest_int instead of a double_int.
4612 (ncd_of_cand_and_phis): Likewise.
4613 (nearest_common_dominator_for_cands): Likewise.
4614 (insert_initializers): Use wide-int interfaces.
4615 (all_phi_incrs_profitable): Likewise.
4616 (replace_one_candidate): Likewise.
4617 (replace_profitable_candidates): Likewise.
4618 * godump.c: Include wide-int-print.h.
4619 (go_output_typedef): Use wide-int interfaces.
4620 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
4621 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
4622 (build_loop_iteration_domains): Likewise.
4623 * hooks.h: Include wide-int.h rather than double-int.h.
4624 (hook_bool_dint_dint_uint_bool_true): Delete.
4625 (hook_bool_wint_wint_uint_bool_true): Declare.
4626 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
4627 (hook_bool_wint_wint_uint_bool_true): New.
4628 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
4630 (ubsan_expand_si_overflow_mul_check): Likewise.
4631 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
4632 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
4633 (get_ancestor_addr_info): Likewise.
4634 (ipa_modify_call_arguments): Likewise.
4635 * loop-doloop.c (doloop_modify): Likewise.
4636 (doloop_optimize): Likewise.
4637 * loop-iv.c (iv_number_of_iterations): Likewise.
4638 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
4639 (unroll_loop_constant_iterations): Likewise.
4640 (decide_unroll_runtime_iterations): Likewise.
4641 (unroll_loop_runtime_iterations): Likewise.
4642 (decide_peel_simple): Likewise.
4643 (decide_unroll_stupid): Likewise.
4644 * lto-streamer-in.c (streamer_read_wi): Add.
4645 (input_cfg): Use wide-int interfaces.
4646 (lto_input_tree_1): Likewise.
4647 * lto-streamer-out.c (streamer_write_wi): Add.
4648 (hash_tree): Use wide-int interfaces.
4649 (output_cfg): Likewise.
4650 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
4651 (GTFILES): Add wide-int.h and signop.h.
4652 (TAGS): Look for .cc files too.
4653 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
4654 * optabs.c (expand_subword_shift): Likewise.
4655 (expand_doubleword_shift): Likewise.
4656 (expand_absneg_bit): Likewise.
4657 (expand_copysign_absneg): Likewise.
4658 (expand_copysign_bit): Likewise.
4659 * postreload.c (reload_cse_simplify_set): Likewise.
4660 * predict.c (predict_iv_comparison): Likewise.
4661 * pretty-print.h: Include wide-int-print.h.
4663 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
4664 * print-tree.c: Include wide-int-print.h.
4665 (print_node_brief): Use wide-int interfaces.
4666 (print_node): Likewise.
4667 * read-rtl.c (validate_const_wide_int): New.
4668 (read_rtx_code): Add CONST_WIDE_INT case.
4669 * real.c: Include wide-int.h.
4670 (real_to_integer2): Delete.
4671 (real_to_integer): New function, returning a wide_int.
4672 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
4673 (ten_to_ptwo): Update call to real_from_integer.
4674 (real_digit): Likewise.
4675 * real.h: Include signop.h, wide-int.h and insn-modes.h.
4676 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
4677 (REAL_VALUE_TO_INT): Delete.
4678 (real_to_integer): Declare a wide-int form.
4679 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
4680 * recog.c (const_int_operand): Improve comment.
4681 (const_scalar_int_operand): New.
4682 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
4683 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
4684 (split_double): Likewise.
4685 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
4686 (rtx_size): Likewise.
4687 (rtx_alloc_stat_v): New.
4688 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
4689 (cwi_output_hex): New.
4690 (iterative_hash_rtx): Handle CONST_WIDE_INT.
4691 (cwi_check_failed_bounds): New.
4692 * rtl.def (CONST_WIDE_INT): New.
4693 * rtl.h: Include <utility> and wide-int.h.
4694 (struct hwivec_def): New.
4695 (CWI_GET_NUM_ELEM): New.
4696 (CWI_PUT_NUM_ELEM): New.
4697 (struct rtx_def): Add num_elem and hwiv.
4698 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
4699 (CASE_CONST_UNIQUE): Likewise.
4700 (CASE_CONST_ANY): Likewise.
4701 (CONST_SCALAR_INT_P): Likewise.
4702 (CONST_WIDE_INT_P): New.
4704 (HWIVEC_CHECK): New.
4705 (cwi_check_failed_bounds): New.
4707 (HWIVEC_CHECK): New.
4708 (CONST_WIDE_INT_VEC) New.
4709 (CONST_WIDE_INT_NUNITS) New.
4710 (CONST_WIDE_INT_ELT) New.
4711 (rtx_mode_t): New type.
4712 (wi::int_traits <rtx_mode_t>): New.
4714 (wi::min_value): New.
4715 (wi::max_value): New.
4717 (const_wide_int_alloc): New.
4718 (immed_wide_int_const): New.
4719 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
4720 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
4721 * signop.h: New file.
4722 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
4723 (simplify_const_unary_operation): Use wide-int interfaces.
4724 (simplify_binary_operation_1): Likewise.
4725 (simplify_const_binary_operation): Likewise.
4726 (simplify_const_relational_operation): Likewise.
4727 (simplify_immed_subreg): Likewise.
4728 * stmt.c (expand_case): Likewise.
4729 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
4730 signop rather than a bool.
4731 * stor-layout.c (layout_type): Use wide-int interfaces.
4732 (initialize_sizetypes): Update calls to
4733 set_min_and_max_values_for_integral_type.
4734 (set_min_and_max_values_for_integral_type): Take a signop rather
4735 than a bool. Use wide-int interfaces.
4736 (fixup_signed_type): Update accordingly. Remove
4737 HOST_BITS_PER_DOUBLE_INT limit.
4738 (fixup_unsigned_type): Likewise.
4739 * system.h (STATIC_CONSTANT_P): New.
4740 (STATIC_ASSERT): New.
4741 * target.def (can_use_doloop_p): Take widest_ints rather than
4743 * target.h: Include wide-int.h rather than double-int.h.
4744 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
4746 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
4747 rather than INT_CST_LT_UNSIGNED.
4748 (can_use_doloop_if_innermost): Take widest_ints rather than
4750 * tree-affine.c: Include wide-int-print.h.
4751 (double_int_ext_for_comb): Delete.
4752 (wide_int_ext_for_comb): New.
4753 (aff_combination_zero): Use wide-int interfaces.
4754 (aff_combination_const): Take a widest_int instead of a double_int.
4755 (aff_combination_elt): Use wide-int interfaces.
4756 (aff_combination_scale): Take a widest_int instead of a double_int.
4757 (aff_combination_add_elt): Likewise.
4758 (aff_combination_add_cst): Likewise.
4759 (aff_combination_add): Use wide-int interfaces.
4760 (aff_combination_convert): Likewise.
4761 (tree_to_aff_combination): Likewise.
4762 (add_elt_to_tree): Take a widest_int instead of a double_int.
4763 (aff_combination_to_tree): Use wide-int interfaces.
4764 (aff_combination_remove_elt): Likewise.
4765 (aff_combination_add_product): Take a widest_int instead of
4767 (aff_combination_mult): Use wide-int interfaces.
4768 (aff_combination_expand): Likewise.
4769 (double_int_constant_multiple_p): Delete.
4770 (wide_int_constant_multiple_p): New.
4771 (aff_combination_constant_multiple_p): Take a widest_int pointer
4772 instead of a double_int pointer.
4773 (print_aff): Use wide-int interfaces.
4774 (get_inner_reference_aff): Take a widest_int pointer
4775 instead of a double_int pointer.
4776 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
4777 * tree-affine.h: Include wide-int.h.
4778 (struct aff_comb_elt): Change type of coef to widest_int.
4779 (struct affine_tree_combination): Change type of offset to widest_int.
4780 (double_int_ext_for_comb): Delete.
4781 (wide_int_ext_for_comb): New.
4782 (aff_combination_const): Use widest_int instead of double_int.
4783 (aff_combination_scale): Likewise.
4784 (aff_combination_add_elt): Likewise.
4785 (aff_combination_constant_multiple_p): Likewise.
4786 (get_inner_reference_aff): Likewise.
4787 (aff_comb_cannot_overlap_p): Likewise.
4788 (aff_combination_zero_p): Use wide-int interfaces.
4789 * tree.c: Include tree.h.
4790 (init_ttree): Use make_int_cst.
4791 (tree_code_size): Removed code for INTEGER_CST case.
4792 (tree_size): Add INTEGER_CST case.
4793 (make_node_stat): Update comment.
4794 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
4795 (build_int_cst_type): Use wide-int interfaces.
4796 (double_int_to_tree): Likewise.
4797 (double_int_fits_to_tree_p): Delete.
4798 (force_fit_type_double): Delete.
4799 (force_fit_type): New.
4800 (int_cst_hash_hash): Use wide-int interfaces.
4801 (int_cst_hash_eq): Likewise.
4802 (build_int_cst_wide): Delete.
4803 (wide_int_to_tree): New.
4804 (cache_integer_cst): Use wide-int interfaces.
4805 (build_low_bits_mask): Likewise.
4806 (cst_and_fits_in_hwi): Likewise.
4807 (real_value_from_int_cst): Likewise.
4808 (make_int_cst_stat): New.
4809 (integer_zerop): Use wide_int interfaces.
4810 (integer_onep): Likewise.
4811 (integer_all_onesp): Likewise.
4812 (integer_pow2p): Likewise.
4813 (integer_nonzerop): Likewise.
4814 (tree_log2): Likewise.
4815 (tree_floor_log2): Likewise.
4816 (tree_ctz): Likewise.
4817 (int_size_in_bytes): Likewise.
4818 (mem_ref_offset): Return an offset_int rather than a double_int.
4819 (build_type_attribute_qual_variant): Use wide_int interfaces.
4820 (type_hash_eq): Likewise
4821 (tree_int_cst_equal): Likewise.
4822 (tree_int_cst_lt): Delete.
4823 (tree_int_cst_compare): Likewise.
4824 (tree_fits_shwi_p): Use wide_int interfaces.
4825 (tree_fits_uhwi_p): Likewise.
4826 (tree_int_cst_sign_bit): Likewise.
4827 (tree_int_cst_sgn): Likewise.
4828 (tree_int_cst_min_precision): Take a signop rather than a bool.
4829 (simple_cst_equal): Use wide_int interfaces.
4830 (compare_tree_int): Likewise.
4831 (iterative_hash_expr): Likewise.
4832 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
4834 (get_type_static_bounds): Use wide_int interfaces.
4835 (tree_int_cst_elt_check_failed): New.
4836 (build_common_tree_nodes): Reordered to set prec before filling in
4838 (int_cst_value): Check cst_and_fits_in_hwi.
4839 (widest_int_cst_value): Use wide_int interfaces.
4840 (upper_bound_in_type): Likewise.
4841 (lower_bound_in_type): Likewise.
4842 (num_ending_zeros): Likewise.
4843 (drop_tree_overflow): Likewise.
4844 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
4845 (gen_conditions_for_pow_cst_base): Likewise.
4846 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
4847 (group_case_labels_stmt): Use wide-int interfaces.
4848 (verify_gimple_assign_binary): Likewise.
4849 (print_loop): Likewise.
4850 * tree-chrec.c (tree_fold_binomial): Likewise.
4851 * tree-core.h (struct tree_base): Add int_length.
4852 (struct tree_int_cst): Change rep of value.
4853 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
4854 (dr_may_alias_p): Likewise.
4855 (max_stmt_executions_tree): Likewise.
4856 * tree.def (INTEGER_CST): Update comment.
4857 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
4858 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
4859 * tree-dump.c: Include wide-int.h and wide-int-print.h.
4860 (dequeue_and_dump): Use wide-int interfaces.
4861 * tree.h: Include wide-int.h.
4862 (NULL_TREE): Moved to earlier loc in file.
4863 (TREE_INT_CST_ELT_CHECK): New.
4864 (tree_int_cst_elt_check_failed): New.
4866 (TREE_INT_CST): Delete.
4867 (TREE_INT_CST_LOW): Use wide-int interfaces.
4868 (TREE_INT_CST_HIGH): Delete.
4869 (TREE_INT_CST_NUNITS): New.
4870 (TREE_INT_CST_EXT_NUNITS): Likewise.
4871 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
4872 (TREE_INT_CST_ELT): Likewise.
4873 (INT_CST_LT): Delete.
4874 (tree_int_cst_elt_check): New (two forms).
4875 (type_code_size): Update comment.
4876 (make_int_cst_stat, make_int_cst): New.
4877 (tree_to_double_int): Delete.
4878 (double_int_fits_to_tree_p): Delete.
4879 (force_fit_type_double): Delete.
4880 (build_int_cstu): Replace with out-of-line function.
4881 (build_int_cst_wide): Delete.
4882 (tree_int_cst_lt): Define inline.
4883 (tree_int_cst_le): New.
4884 (tree_int_cst_compare): Define inline.
4885 (tree_int_cst_min_precision): Take a signop rather than a bool.
4886 (wi::int_traits <const_tree>): New.
4887 (wi::int_traits <tree>): New.
4888 (wi::extended_tree): New.
4889 (wi::int_traits <wi::extended_tree>): New.
4890 (wi::to_widest): New.
4891 (wi::to_offset): New.
4892 (wi::fits_to_tree_p): New.
4893 (wi::min_value): New.
4894 (wi::max_value): New.
4895 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
4896 (copy_tree_body_r): Likewise.
4897 * tree-object-size.c (compute_object_offset): Likewise.
4898 (addr_object_size): Likewise.
4899 * tree-predcom.c: Include wide-int-print.h.
4900 (struct dref_d): Change type of offset to widest_int.
4901 (dump_dref): Call wide-int printer.
4902 (aff_combination_dr_offset): Use wide-int interfaces.
4903 (determine_offset): Take a widest_int pointer rather than a
4905 (split_data_refs_to_components): Use wide-int interfaces.
4906 (suitable_component_p): Likewise.
4907 (order_drefs): Likewise.
4908 (add_ref_to_chain): Likewise.
4909 (valid_initializer_p): Likewise.
4910 (determine_roots_comp): Likewise.
4911 * tree-pretty-print.c: Include wide-int-print.h.
4912 (dump_generic_node): Use wide-int interfaces.
4913 * tree-sra.c (sra_ipa_modify_expr): Likewise.
4914 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
4915 (move_fixed_address_to_symbol): Likewise.
4916 (move_hint_to_base): Likewise.
4917 (move_pointer_to_base): Likewise.
4918 (move_variant_to_index): Likewise.
4919 (most_expensive_mult_to_index): Likewise.
4920 (addr_to_parts): Likewise.
4921 (copy_ref_info): Likewise.
4922 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
4923 (indirect_refs_may_alias_p): Likewise.
4924 (stmt_kills_ref_p_1): Likewise.
4925 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
4926 * tree-ssa-ccp.c: Update comment at top of file. Include
4928 (struct prop_value_d): Change type of mask to widest_int.
4929 (extend_mask): New function.
4930 (dump_lattice_value): Use wide-int interfaces.
4931 (get_default_value): Likewise.
4932 (set_constant_value): Likewise.
4933 (set_value_varying): Likewise.
4934 (valid_lattice_transition): Likewise.
4935 (set_lattice_value): Likewise.
4936 (value_to_double_int): Delete.
4937 (value_to_wide_int): New.
4938 (get_value_from_alignment): Use wide-int interfaces.
4939 (get_value_for_expr): Likewise.
4940 (do_dbg_cnt): Likewise.
4941 (ccp_finalize): Likewise.
4942 (ccp_lattice_meet): Likewise.
4943 (bit_value_unop_1): Use widest_ints rather than double_ints.
4944 (bit_value_binop_1): Likewise.
4945 (bit_value_unop): Use wide-int interfaces.
4946 (bit_value_binop): Likewise.
4947 (bit_value_assume_aligned): Likewise.
4948 (evaluate_stmt): Likewise.
4949 (ccp_fold_stmt): Likewise.
4950 (visit_cond_stmt): Likewise.
4951 (ccp_visit_stmt): Likewise.
4952 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
4953 (constant_pointer_difference): Likewise.
4954 (associate_pointerplus): Likewise.
4955 (combine_conversions): Likewise.
4956 * tree-ssa-loop.h: Include wide-int.h.
4957 (struct tree_niter_desc): Change type of max to widest_int.
4958 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
4959 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
4960 (remove_redundant_iv_tests): Likewise.
4961 (canonicalize_loop_induction_variables): Likewise.
4962 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
4963 (constant_multiple_of): Take a widest_int pointer instead of
4964 a double_int pointer.
4965 (get_computation_aff): Use wide-int interfaces.
4966 (ptr_difference_cost): Likewise.
4967 (difference_cost): Likewise.
4968 (get_loop_invariant_expr_id): Likewise.
4969 (get_computation_cost_at): Likewise.
4970 (iv_elimination_compare_lt): Likewise.
4971 (may_eliminate_iv): Likewise.
4972 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
4973 instead of double_int.
4974 (max_loop_iterations): Likewise.
4975 (max_stmt_executions): Likewise.
4976 (estimated_stmt_executions): Likewise.
4977 * tree-ssa-loop-niter.c: Include wide-int-print.h.
4978 (split_to_var_and_offset): Use wide-int interfaces.
4979 (determine_value_range): Likewise.
4980 (bound_difference_of_offsetted_base): Likewise.
4981 (bounds_add): Take a widest_int instead of a double_int.
4982 (number_of_iterations_ne_max): Use wide-int interfaces.
4983 (number_of_iterations_ne): Likewise.
4984 (number_of_iterations_lt_to_ne): Likewise.
4985 (assert_loop_rolls_lt): Likewise.
4986 (number_of_iterations_lt): Likewise.
4987 (number_of_iterations_le): Likewise.
4988 (number_of_iterations_cond): Likewise.
4989 (number_of_iterations_exit): Likewise.
4990 (finite_loop_p): Likewise.
4991 (derive_constant_upper_bound_assign): Likewise.
4992 (derive_constant_upper_bound): Return a widest_int.
4993 (derive_constant_upper_bound_ops): Likewise.
4994 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
4995 (record_estimate): Take a widest_int rather than a double_int.
4996 (record_nonwrapping_iv): Use wide-int interfaces.
4997 (double_int_cmp): Delete.
4998 (wide_int_cmp): New.
4999 (bound_index): Take a widest_int rather than a double_int.
5000 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
5001 (maybe_lower_iteration_bound): Likewise.
5002 (estimate_numbers_of_iterations_loop): Likewise.
5003 (estimated_loop_iterations): Take a widest_int pointer than than
5004 a double_int pointer.
5005 (estimated_loop_iterations_int): Use wide-int interfaces.
5006 (max_loop_iterations): Take a widest_int pointer than than
5007 a double_int pointer.
5008 (max_loop_iterations_int): Use wide-int interfaces.
5009 (max_stmt_executions): Take a widest_int pointer than than
5010 a double_int pointer.
5011 (estimated_stmt_executions): Likewise.
5012 (n_of_executions_at_most): Use wide-int interfaces.
5013 (scev_probably_wraps_p): Likewise.
5014 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
5016 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
5018 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
5019 double_ints. Adjust for trailing_wide_ints <3> representation.
5020 (set_nonzero_bits): Likewise.
5021 (get_range_info): Return wide_ints rather than double_ints.
5022 Adjust for trailing_wide_ints <3> representation.
5023 (get_nonzero_bits): Likewise.
5024 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
5026 * tree-ssanames.h (struct range_info_def): Replace min, max and
5027 nonzero_bits with a trailing_wide_ints <3>.
5028 (set_range_info): Use wide_int_refs rather than double_ints.
5029 (set_nonzero_bits): Likewise.
5030 (get_range_info): Return wide_ints rather than double_ints.
5031 (get_nonzero_bits): Likewise.
5032 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
5033 * tree-ssa-pre.c (phi_translate_1): Likewise.
5034 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
5035 (acceptable_pow_call): Likewise.
5036 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
5038 (vn_reference_fold_indirect): Likewise.
5039 (vn_reference_maybe_forwprop_address): Likewise.
5040 (valueize_refs_1): Likewise.
5041 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
5042 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
5043 tree_int_cst_lt and tree_int_cst_le.
5044 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
5046 (streamer_alloc_tree): Likewise.
5047 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
5048 (streamer_write_tree_header): Likewise.
5049 (streamer_write_integer_cst): Likewise.
5050 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
5051 (build_constructors): Likewise.
5052 (array_value_type): Likewise.
5053 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
5054 (vect_check_gather): Likewise.
5055 * tree-vect-generic.c (build_replicated_const): Likewise.
5056 (expand_vector_divmod): Likewise.
5057 * tree-vect-loop.c (vect_transform_loop): Likewise.
5058 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
5059 (vect_do_peeling_for_alignment): Likewise.
5060 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
5061 * tree-vrp.c: Include wide-int.h.
5062 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
5063 (extract_range_from_assert): Use wide-int interfaces.
5064 (vrp_int_const_binop): Likewise.
5065 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
5066 double_int pointers.
5067 (ranges_from_anti_range): Use wide-int interfaces.
5068 (quad_int_cmp): Delete.
5069 (quad_int_pair_sort): Likewise.
5070 (extract_range_from_binary_expr_1): Use wide-int interfaces.
5071 (extract_range_from_unary_expr_1): Likewise.
5072 (adjust_range_with_scev): Likewise.
5073 (masked_increment): Take and return wide_ints rather than double_ints.
5074 (register_edge_assert_for_2): Use wide-int interfaces.
5075 (check_array_ref): Likewise.
5076 (search_for_addr_array): Likewise.
5077 (maybe_set_nonzero_bits): Likewise.
5078 (union_ranges): Pass an integer of the correct type instead of
5079 using integer_one_node.
5080 (intersect_ranges): Likewise.
5081 (simplify_truth_ops_using_ranges): Likewise.
5082 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
5083 (range_fits_type_p): Likewise.
5084 (simplify_cond_using_ranges): Likewise. Take a signop rather than
5086 (simplify_conversion_using_ranges): Use wide-int interfaces.
5087 (simplify_float_conversion_using_ranges): Likewise.
5088 (vrp_finalize): Likewise.
5089 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
5090 (gimple_stringops_transform): Likewise.
5091 * varasm.c (decode_addr_const): Likewise.
5092 (const_hash_1): Likewise.
5093 (const_rtx_hash_1): Likewise
5094 (output_constant): Likewise.
5095 (array_size_for_constructor): Likewise.
5096 (output_constructor_regular_field): Likewise.
5097 (output_constructor_bitfield): Likewise.
5098 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
5099 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
5101 * gencheck.c: Define BITS_PER_UNIT.
5104 * wide-int-print.cc: New.
5105 * wide-int-print.h: New.
5107 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5109 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
5111 2014-05-06 Richard Biener <rguenther@suse.de>
5113 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
5114 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
5115 (TODO_verify_all): Adjust.
5116 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
5117 TODO_verify_stmts and TODO_verify_rtl_sharing.
5118 * bb-reorder.c: Likewise.
5119 * cfgexpand.c: Likewise.
5120 * cprop.c: Likewise.
5122 * function.c: Likewise.
5123 * fwprop.c: Likewise.
5125 * gimple-ssa-isolate-paths.c: Likewise.
5126 * gimple-ssa-strength-reduction.c: Likewise.
5127 * ipa-split.c: Likewise.
5128 * loop-init.c: Likewise.
5129 * loop-unroll.c: Likewise.
5130 * lower-subreg.c: Likewise.
5131 * modulo-sched.c: Likewise.
5132 * postreload-gcse.c: Likewise.
5133 * predict.c: Likewise.
5134 * recog.c: Likewise.
5135 * sched-rgn.c: Likewise.
5136 * store-motion.c: Likewise.
5137 * tracer.c: Likewise.
5138 * trans-mem.c: Likewise.
5139 * tree-call-cdce.c: Likewise.
5140 * tree-cfg.c: Likewise.
5141 * tree-cfgcleanup.c: Likewise.
5142 * tree-complex.c: Likewise.
5143 * tree-eh.c: Likewise.
5144 * tree-emutls.c: Likewise.
5145 * tree-if-conv.c: Likewise.
5146 * tree-into-ssa.c: Likewise.
5147 * tree-loop-distribution.c: Likewise.
5148 * tree-object-size.c: Likewise.
5149 * tree-parloops.c: Likewise.
5150 * tree-pass.h: Likewise.
5151 * tree-sra.c: Likewise.
5152 * tree-ssa-ccp.c: Likewise.
5153 * tree-ssa-copy.c: Likewise.
5154 * tree-ssa-copyrename.c: Likewise.
5155 * tree-ssa-dce.c: Likewise.
5156 * tree-ssa-dom.c: Likewise.
5157 * tree-ssa-dse.c: Likewise.
5158 * tree-ssa-forwprop.c: Likewise.
5159 * tree-ssa-ifcombine.c: Likewise.
5160 * tree-ssa-loop-ch.c: Likewise.
5161 * tree-ssa-loop-ivcanon.c: Likewise.
5162 * tree-ssa-loop.c: Likewise.
5163 * tree-ssa-math-opts.c: Likewise.
5164 * tree-ssa-phiopt.c: Likewise.
5165 * tree-ssa-phiprop.c: Likewise.
5166 * tree-ssa-pre.c: Likewise.
5167 * tree-ssa-reassoc.c: Likewise.
5168 * tree-ssa-sink.c: Likewise.
5169 * tree-ssa-strlen.c: Likewise.
5170 * tree-ssa-tail-merge.c: Likewise.
5171 * tree-ssa-uncprop.c: Likewise.
5172 * tree-switch-conversion.c: Likewise.
5173 * tree-tailcall.c: Likewise.
5174 * tree-vect-generic.c: Likewise.
5175 * tree-vectorizer.c: Likewise.
5176 * tree-vrp.c: Likewise.
5178 * var-tracking.c: Likewise.
5179 * bt-load.c: Likewise.
5180 * cfgcleanup.c: Likewise.
5181 * combine-stack-adj.c: Likewise.
5182 * combine.c: Likewise.
5183 * compare-elim.c: Likewise.
5184 * config/epiphany/resolve-sw-modes.c: Likewise.
5185 * config/i386/i386.c: Likewise.
5186 * config/mips/mips.c: Likewise.
5187 * config/s390/s390.c: Likewise.
5188 * config/sh/sh_treg_combine.cc: Likewise.
5189 * config/sparc/sparc.c: Likewise.
5192 * final.c: Likewise.
5193 * ifcvt.c: Likewise.
5194 * mode-switching.c: Likewise.
5195 * passes.c: Likewise.
5196 * postreload.c: Likewise.
5198 * reg-stack.c: Likewise.
5199 * regcprop.c: Likewise.
5200 * regrename.c: Likewise.
5203 2014-05-06 Richard Biener <rguenther@suse.de>
5206 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
5207 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
5209 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
5212 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
5214 2014-05-05 Radovan Obradovic <robradovic@mips.com>
5215 Tom de Vries <tom@codesourcery.com>
5217 * target.def (call_fusage_contains_non_callee_clobbers): New
5219 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
5221 (@node Miscellaneous Register Hooks): New node.
5222 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
5223 * doc/tm.texi: Regenerate.
5225 2014-05-05 Marek Polacek <polacek@redhat.com>
5228 * opts.c (common_handle_option): Call error_at instead of warning_at.
5230 2014-05-05 Richard Biener <rguenther@suse.de>
5232 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
5233 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
5234 under the TODO_verify_il umbrella.
5236 2014-05-05 Richard Biener <rguenther@suse.de>
5238 * passes.c (execute_function_todo): Move TODO_verify_flow under
5239 the TODO_verify_ul umbrella.
5241 2014-05-05 Richard Biener <rguenther@suse.de>
5244 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
5245 X & CST away from a CST that is the mask of a mode.
5247 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5249 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
5250 int argument to enum machine_mode.
5251 (picochip_class_max_nregs): Ditto.
5252 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
5253 (picochip_class_max_nregs): Ditto.
5255 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5257 * target.def: Add new target hook.
5258 * doc/tm.texi: Regenerate.
5259 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
5260 * targhooks.c (default_keep_leaf_when_profiled): New function.
5262 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
5263 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
5265 2014-05-05 Bin Cheng <bin.cheng@arm.com>
5267 PR tree-optimization/60363
5268 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
5269 (copy_phi_args): New parameters. Call get_value_locus_in_path.
5270 (update_destination_phis): New parameter.
5271 (create_edge_and_update_destination_phis): Ditto.
5272 (ssa_fix_duplicate_block_edges): Pass new arguments.
5273 (thread_single_edge): Ditto.
5275 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
5277 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
5278 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
5279 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
5280 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
5281 Use RS6000_BTM_HARD_FLOAT.
5282 (BU_MISC_2): Likewise.
5283 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
5284 RS6000_BTM_HARD_FLOAT.
5285 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
5287 (rs6000_invalid_builtin): Add hard floating builtin support.
5288 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
5289 hard float builtins.
5290 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
5292 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
5294 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
5295 Add missing function* argument.
5297 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
5299 * lra-constraints.c (valid_address_p): Move earlier in file.
5300 Add a constraint argument to the address_info version.
5301 (satisfies_memory_constraint_p): New function.
5302 (satisfies_address_constraint_p): Likewise.
5303 (process_alt_operands, curr_insn_transform): Use them.
5304 (process_address): Pass the constraint to valid_address_p when
5305 checking address operands.
5307 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
5309 * config/mips/mips.c (mips_isa_rev): New variable.
5310 (mips_set_architecture): Set it.
5311 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
5313 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
5314 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
5315 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
5316 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
5317 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
5318 conditions in terms of mips_isa_rev.
5319 (mips_isa_rev): Declare.
5321 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
5323 * config/sh/sh-mem.cc: Use tabs instead of spaces.
5324 (prob_unlikely, prob_likely): Make variables const.
5326 2014-05-03 Denis Chertykov <chertykov@gmail.com>
5328 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
5330 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
5332 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
5334 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
5336 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
5337 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
5338 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
5340 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
5341 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
5343 Replace usage of ROUND_REG with sh_round_reg.
5344 Use CEIL instead of ROUND_ADVANCE.
5346 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
5349 * config/sh/sh.c: Include stdlib headers before everything else.
5351 2014-05-02 Jakub Jelinek <jakub@redhat.com>
5353 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
5354 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
5355 (gimplify_adjust_omp_clauses): Simd region is never
5356 directly nested in combined parallel. Instead, for linear
5357 with copyin/copyout, if in combined for simd loop, make decl
5358 firstprivate/lastprivate on OMP_FOR.
5359 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
5360 expand_omp_for_static_chunk): When setting endvar, also set
5361 fd->loop.v to the same value.
5363 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5365 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
5367 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
5369 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
5372 2014-05-02 Marek Polacek <polacek@redhat.com>
5374 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
5376 2014-05-02 Kito Cheng <kito@0xlab.org>
5378 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
5379 to a C expression marco.
5380 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
5381 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
5382 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
5383 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
5384 HONOR_REG_ALLOC_ORDER.
5385 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
5387 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5389 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
5391 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5393 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
5395 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
5397 * tree-if-conv.c (is_cond_scalar_reduction): New function.
5398 (convert_scalar_cond_reduction): Likewise.
5399 (predicate_scalar_phi): Add recognition and transformation
5400 of simple conditioanl reduction to be vectorizable.
5402 2014-05-01 Marek Polacek <polacek@redhat.com>
5405 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
5407 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
5409 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
5410 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
5411 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
5412 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
5413 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
5414 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
5415 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
5416 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
5418 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
5420 * config/arc/arc.opt (mlra): Move comment above option name
5421 to avoid mis-parsing as language options.
5423 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5425 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
5426 * config/sol2.h: ... here.
5427 * config/sol2-10.h: Remove.
5429 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
5430 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
5431 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
5432 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
5433 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
5434 * config/sol2.h: ... here.
5435 (SECTION_NAME_FORMAT): Don't redefine.
5436 (STARTFILE_ARCH32_SPEC): Rename to ...
5437 (STARTFILE_ARCH_SPEC): ... this.
5438 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
5439 * config/sparc/sol2.h: ... here.
5440 (SECTION_NAME_FORMAT): Don't undef.
5441 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
5442 (SUBTARGET_EXTRA_SPECS): Remove.
5443 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
5445 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
5446 (MD_STARTFILE_PREFIX): Remove.
5447 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
5448 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
5449 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
5450 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
5451 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
5452 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
5453 * config/i386/sol2.h: ... here.
5454 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
5455 * config/i386/sol2-bi.h: Remove.
5456 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
5457 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
5459 * config/i386/t-sol2-64: Rename to ...
5460 * config/i386/t-sol2: ... this.
5461 * config/sparc/t-sol2-64: Rename to ...
5462 * config/sparc/t-sol2: ... this.
5464 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
5465 sol2_tm_file_head, sol2_tm_file_tail.
5466 Include ${cpu_type}/sol2.h before sol2.h.
5468 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
5469 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
5470 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
5471 Reflect i386/t-sol2-64 renaming.
5472 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
5473 Reflect sparc/t-sol2-64 renaming.
5475 2014-04-30 Richard Biener <rguenther@suse.de>
5477 * passes.c (execute_function_todo): Move TODO_verify_stmts
5478 and TODO_verify_ssa under the TODO_verify_il umbrella.
5479 * tree-ssa.h (verify_ssa): Adjust prototype.
5480 * tree-ssa.c (verify_ssa): Add parameter to tell whether
5481 we should verify SSA operands.
5482 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
5483 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
5484 whether we should verify whether not throwing stmts have EH info.
5485 * graphite-scop-detection.c (create_sese_edges): Adjust.
5486 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
5487 * tree-eh.c (lower_try_finally_switch): Do not add the
5488 default case label twice.
5490 2014-04-30 Marek Polacek <polacek@redhat.com>
5492 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
5493 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
5494 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
5495 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
5497 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
5499 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
5500 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
5501 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
5502 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
5503 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
5504 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
5505 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
5506 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
5508 2014-04-29 David Malcolm <dmalcolm@redhat.com>
5510 * tree-cfg.c (dump_function_to_file): Dump the return type of
5511 functions, in a line to itself before the function body, mimicking
5512 the layout of a C function.
5514 2014-04-29 Jakub Jelinek <jakub@redhat.com>
5516 PR tree-optimization/60971
5517 * tree-tailcall.c (process_assignment): Reject conversions which
5520 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
5522 * calls.c (initialize_argument_information): Always treat
5523 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
5524 (expand_call): Likewise.
5525 (emit_library_call_calue_1): Likewise.
5526 * expr.c (PUSH_ARGS_REVERSED): Do not define.
5527 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
5530 2014-04-29 Nick Clifton <nickc@redhat.com>
5532 * config/msp430/msp430.md (umulsidi): Fix typo.
5533 (mulhisi3): Enable even inside interrupt handlers.
5534 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
5535 bigger return address pushed in large mode.
5537 2014-04-29 Nick Clifton <nickc@redhat.com>
5539 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
5540 (arc_init_reg_tables): Use a machine_mode enum to iterate over
5542 * config/m32r/m32r.c (init_reg_tables): Likewise.
5543 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
5544 enum to hold the modes.
5546 2014-04-29 Richard Biener <rguenther@suse.de>
5548 * dominance.c (free_dominance_info): Add overload with
5550 (dom_info_state): Likewise.
5551 (dom_info_available_p): Likewise.
5552 * basic-block.h (free_dominance_info, dom_info_state,
5553 dom_info_available_p): Declare overloads.
5554 * passes.c (execute_function_todo): Verify that verifiers
5555 don't change dominator info state. Drop dominator info
5556 for IPA pass invocations.
5557 * cgraph.c (release_function_body): Restore asserts that
5558 dominator information is released.
5560 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
5562 * doc/invoke.texi: Fix typo.
5563 * tree-vrp.c: Fix typos.
5564 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
5566 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5568 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
5570 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
5572 * config/aarch64/aarch64-builtins.c
5573 (aarch64_types_storestruct_lane_qualifiers): New.
5574 (TYPES_STORESTRUCT_LANE): Likewise.
5575 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
5576 (st3_lane): Likewise.
5577 (st4_lane): Likewise.
5578 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
5579 (vec_store_lanesci_lane<mode>): Likewise.
5580 (vec_store_lanesxi_lane<mode>): Likewise.
5581 (aarch64_st2_lane<VQ:mode>): Likewise.
5582 (aarch64_st3_lane<VQ:mode>): Likewise.
5583 (aarch64_st4_lane<VQ:mode>): Likewise.
5584 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
5585 * config/aarch64/arm_neon.h
5586 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
5587 use new macro arguments.
5588 (__ST3_LANE_FUNC): Likewise.
5589 (__ST4_LANE_FUNC): Likewise.
5590 * config/aarch64/iterators.md (V_TWO_ELEM): New.
5591 (V_THREE_ELEM): Likewise.
5592 (V_FOUR_ELEM): Likewise.
5594 2014-04-28 David Malcolm <dmalcolm@redhat.com>
5596 * doc/gimple.texi: Replace the description of the now-defunct
5597 union gimple_statement_d with a diagram showing the
5598 gimple_statement_base class hierarchy and its relationships to
5599 the GSS_ and GIMPLE_ enums.
5601 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
5603 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
5604 * config/aarch64/aarch64.c
5605 (aarch64_cannot_change_mode_class): Weaken conditions.
5606 (aarch64_modes_tieable_p): New.
5607 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
5609 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
5611 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5612 (loadsync_<mode>): Change mode.
5613 (load_quadpti, store_quadpti): New.
5614 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5615 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5617 2014-04-28 Martin Jambor <mjambor@suse.cz>
5619 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
5620 same alias type as the original statement.
5621 (subreplacement_assignment_data): New type.
5622 (handle_unscalarized_data_in_subtree): New type of parameter,
5623 generate new memory accesses with same alias type as the original
5625 (load_assign_lhs_subreplacements): Likewise.
5626 (sra_modify_constructor_assign): Generate new memory accesses with
5627 same alias type as the original statement.
5629 2014-04-28 Richard Biener <rguenther@suse.de>
5631 * tree-pass.h (TODO_verify_il): Define.
5632 (TODO_verify_all): Complete properly.
5633 * passes.c (execute_function_todo): Move existing loop-closed
5634 SSA verification under TODO_verify_il.
5635 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
5636 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
5637 Fix tree sharing issue.
5639 2014-04-28 Richard Biener <rguenther@suse.de>
5642 * builtins.def (DEF_C11_BUILTIN): Add.
5643 (BUILT_IN_ALIGNED_ALLOC): Likewise.
5644 * coretypes.h (enum function_class): Add function_c11_misc.
5645 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
5646 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
5647 (call_may_clobber_ref_p_1): Likewise.
5648 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
5649 (mark_all_reaching_defs_necessary_1): Likewise.
5650 (propagate_necessity): Likewise.
5651 (eliminate_unnecessary_stmts): Likewise.
5652 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
5654 2014-04-28 Richard Biener <rguenther@suse.de>
5656 * tree-vrp.c (vrp_var_may_overflow): Remove.
5657 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
5658 with overflow immediately bump to one before that value and
5659 let iteration figure out overflow status.
5661 2014-04-28 Richard Biener <rguenther@suse.de>
5663 * configure.ac: Do valgrind header checks unconditionally.
5664 Add --enable-valgrind-annotations.
5665 * system.h: Guard valgrind header inclusion with
5666 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
5667 * alloc-pool.c (pool_alloc, pool_free): Use
5668 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
5669 to guard possibly dead code.
5670 * config.in: Regenerated.
5671 * configure: Likewise.
5673 2014-04-28 Jeff Law <law@redhat.com>
5675 PR tree-optimization/60902
5676 * tree-ssa-threadedge.c
5677 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
5678 over real defs when invalidating outputs from statements that do not
5679 produce useful outputs for threading.
5681 2014-04-28 Richard Biener <rguenther@suse.de>
5683 PR tree-optimization/60979
5684 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
5685 SCOPs that end in a block with a successor with abnormal
5688 2014-04-28 Richard Biener <rguenther@suse.de>
5690 * tree-pass.h (execute_pass_list): Adjust prototype.
5691 * passes.c (pass_manager::execute_early_local_passes): Adjust.
5692 (do_per_function): Change callback signature, push all actual
5693 work to the callbals.
5694 (do_per_function_toporder): Likewise.
5695 (execute_function_dump): Adjust.
5696 (execute_function_todo): Likewise.
5697 (clear_last_verified): Likewise.
5698 (verify_curr_properties): Likewise.
5699 (update_properties_after_pass): Likewise.
5700 (execute_pass_list_1): Split out from ...
5701 (execute_pass_list): ... here. Adjust.
5702 (execute_ipa_pass_list): Likewise.
5703 * cgraphunit.c (cgraph_add_new_function): Adjust.
5704 (analyze_function): Likewise.
5705 (expand_function): Likewise.
5706 * cgraph.c (release_function_body): Free dominance info
5707 here instead of asserting it was magically freed elsewhere.
5709 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
5711 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
5712 * configure: Regenerate.
5713 * config/sparc/sparc.opt (muser-mode): New option.
5714 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
5716 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
5717 * doc/invoke.texi (SPARC options): Document -muser-mode.
5719 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
5721 * cselib.c (find_slot_memmode): Delete.
5722 (cselib_hasher): Change compare_type to a struct.
5723 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
5725 (preserve_constants_and_equivs): Adjust for new compare_type.
5726 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
5727 (wrap_constant): Delete.
5728 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
5730 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
5732 * doc/install.texi (Building with profile feedback): Remove
5735 2014-04-26 Tom de Vries <tom@codesourcery.com>
5737 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
5740 2014-04-25 Cary Coutant <ccoutant@google.com>
5743 * dwarf2out.c (should_move_die_to_comdat): A type definition
5744 can contain a subprogram definition, but don't move it to a
5746 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
5747 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
5749 (clone_tree_hash): Rename to...
5750 (clone_tree_partial): ...this; change callers. Copy
5751 DW_TAG_subprogram DIEs as declarations.
5752 (copy_decls_walk): Don't copy children of a declaration into a
5755 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
5758 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
5761 2014-04-25 Jiong Wang <jiong.wang@arm.com>
5763 * config/arm/predicates.md (call_insn_operand): Add long_call check.
5764 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
5766 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
5769 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5771 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
5773 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5775 PR tree-optimization/60930
5776 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
5777 creating a multiply candidate by folding two constant
5778 multiplicands when the result overflows.
5780 2014-04-25 Jakub Jelinek <jakub@redhat.com>
5782 PR tree-optimization/60960
5783 * tree-vect-generic.c (expand_vector_operation): Only call
5784 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
5786 2014-04-25 Tom de Vries <tom@codesourcery.com>
5788 * expr.c (clobber_reg_mode): New function.
5789 * expr.h (clobber_reg): New function.
5791 2014-04-25 Tom de Vries <tom@codesourcery.com>
5793 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
5796 2014-04-25 Radovan Obradovic <robradovic@mips.com>
5797 Tom de Vries <tom@codesourcery.com>
5799 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
5801 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
5802 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
5803 new argument to find_all_hard_reg_sets call.
5805 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5807 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
5808 Use HOST_WIDE_INT_C for mask literal.
5809 (aarch_rev16_shleft_mask_imm_p): Likewise.
5811 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
5814 * config/sparc/sparc.md (ashlsi3_extend): Delete.
5816 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
5818 PR preprocessor/56540
5819 * config/i386/i386-c.c (ix86_target_macros): Define
5820 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
5822 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5824 * configure.ac (tga_func): Remove.
5825 (LIB_TLS_SPEC): Remove.
5826 * configure: Regenerate.
5827 * config.in: Regenerate.
5828 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
5830 2014-04-25 Richard Biener <rguenther@suse.de>
5833 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
5834 call stmt use/clobber sets during stmt walk instead of
5835 walking the possibly incomplete set of caller edges.
5837 2014-04-25 Richard Biener <rguenther@suse.de>
5840 * passes.c (apply_ipa_transforms): Inline into only caller ...
5841 (execute_one_pass): ... here. Properly bring in function
5842 bodies for nodes we want to apply IPA transforms to.
5844 2014-04-24 Cong Hou <congh@google.com>
5846 PR tree-optimization/60896
5847 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
5848 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
5849 (vect_mark_pattern_stmts): Set the def type of all statements in
5850 PATTERN_DEF_SEQ as vect_internal_def.
5852 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
5854 * doc/extend.texi (PowerPC Built-in Functions): Document new
5855 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
5856 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
5858 * config/rs6000/predicates.md (const_0_to_3_operand): New
5859 predicate to match 0..3 integer constants.
5861 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
5862 to support adding miscellaneous builtin functions.
5863 (BU_DFP_MISC_2): Likewise.
5864 (BU_P7_MISC_1): Likewise.
5865 (BU_P7_MISC_2): Likewise.
5866 (BU_P8V_MISC_3): Likewise.
5867 (BU_MISC_1): Likewise.
5868 (BU_MISC_2): Likewise.
5869 (DIVWE): Add extended divide builtin functions.
5872 (DIVWEUO): Likewise.
5876 (DIVDEUO): Likewise.
5877 (DXEX): Add decimal floating-point builtin functions.
5880 (DDEDPDQ): Likewise.
5882 (DENBCDQ): Likewise.
5889 (CDTBCD): Add new BCD builtin functions.
5893 (BCDADD_LT): Likewise.
5894 (BCDADD_EQ): Likewise.
5895 (BCDADD_GT): Likewise.
5896 (BCDADD_OV): Likewise.
5898 (BCDSUB_LT): Likewise.
5899 (BCDSUB_EQ): Likewise.
5900 (BCDSUB_GT): Likewise.
5901 (BCDSUB_OV): Likewise.
5902 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
5903 (UNPACK_TD): Likewise.
5904 (PACK_TF): Likewise.
5905 (UNPACK_TF): Likewise.
5906 (UNPACK_TF_0): Likewise.
5907 (UNPACK_TF_1): Likewise.
5908 (PACK_V1TI): Likewise.
5909 (UNPACK_V1TI): Likewise.
5911 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
5912 support for decimal floating point builtin functions.
5913 (rs6000_expand_ternop_builtin): Add checks for the new builtin
5914 functions that take constant arguments.
5915 (rs6000_invalid_builtin): Add decimal floating point builtin support.
5916 (rs6000_init_builtins): Setup long double, _Decimal64, and
5917 _Decimal128 types for new builtin functions.
5918 (builtin_function_type): Set the unsigned flags appropriately for
5919 the new builtin functions.
5920 (rs6000_opt_masks): Add support for decimal floating point builtin
5923 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
5924 floating point builtin functions.
5925 (RS6000_BTM_COMMON): Likewise.
5926 (RS6000_BTI_long_double): Likewise.
5927 (RS6000_BTI_dfloat64): Likewise.
5928 (RS6000_BTI_dfloat128): Likewise.
5929 (long_double_type_internal_node): Likewise.
5930 (dfloat64_type_internal_node): Likewise.
5931 (dfloat128_type_internal_node): Likewise.
5933 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
5934 2.07 bcd arithmetic instructions.
5935 (UNSPEC_BCDSUB): Likewise.
5936 (UNSPEC_BCD_OVERFLOW): Likewise.
5937 (UNSPEC_BCD_ADD_SUB): Likewise.
5938 (bcd_add_sub): Likewise.
5939 (BCD_TEST): Likewise.
5940 (bcd<bcd_add_sub>): Likewise.
5941 (bcd<bcd_add_sub>_test): Likewise.
5942 (bcd<bcd_add_sub>_test2): Likewise.
5943 (bcd<bcd_add_sub>_<code>): Likewise.
5944 (peephole2 for combined bcd ops): Likewise.
5946 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
5947 decimal floating point builtin functions.
5948 (UNSPEC_DENBCD): Likewise.
5949 (UNSPEC_DXEX): Likewise.
5950 (UNSPEC_DIEX): Likewise.
5951 (UNSPEC_DSCLI): Likewise.
5952 (UNSPEC_DSCRI): Likewise.
5953 (D64_D128): Likewise.
5954 (dfp_suffix): Likewise.
5955 (dfp_ddedpd_<mode>): Likewise.
5956 (dfp_denbcd_<mode>): Likewise.
5957 (dfp_dxex_<mode>): Likewise.
5958 (dfp_diex_<mode>): Likewise.
5959 (dfp_dscli_<mode>): Likewise.
5960 (dfp_dscri_<mode>): Likewise.
5962 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
5964 (UNSPEC_CDTBCD): Likewise.
5965 (UNSPEC_CBCDTD): Likewise.
5966 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
5967 (UNSPEC_DIVEO): Likewise.
5968 (UNSPEC_DIVEU): Likewise.
5969 (UNSPEC_DIVEUO): Likewise.
5970 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
5971 pack/unpack 128-bit types.
5972 (UNSPEC_PACK_128BIT): Likewise.
5973 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
5974 (udiv<mode>3): Use idiv_ldiv mode attribute.
5975 (div<mode>3): Likewise.
5976 (addg6s): Add new BCD builtin functions.
5979 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
5980 (div_extend): Likewise.
5981 (div<div_extend>_<mode>"): Likewise.
5982 (FP128_64): Add support for new builtin functions to pack/unpack
5984 (unpack<mode>): Likewise.
5985 (unpacktf_0): Likewise.
5986 (unpacktf_1): Likewise.
5987 (unpack<mode>_dm): Likewise.
5988 (unpack<mode>_nodm): Likewise.
5989 (pack<mode>): Likewise.
5990 (unpackv1ti): Likewise.
5991 (packv1ti): Likewise.
5993 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
5995 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
5998 2014-04-24 Jakub Jelinek <jakub@redhat.com>
6000 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
6001 * gimplify.c (omp_is_private): Change last argument's type to int.
6002 Only diagnose lastprivate if the simd argument is 1, only diagnose
6003 linear if the simd argument is 2.
6004 (gimplify_omp_for): Adjust omp_is_private callers. When adding
6005 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
6006 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
6007 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
6008 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
6009 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
6010 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
6011 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
6012 * tree-nested.c (convert_nonlocal_omp_clauses,
6013 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
6015 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
6018 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
6021 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
6023 * flag-types.h (enum ivar_visibility): Add.
6025 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
6027 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
6028 function * argument.
6030 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
6032 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
6034 2014-04-24 Radovan Obradovic <robradovic@mips.com>
6035 Tom de Vries <tom@codesourcery.com>
6037 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
6038 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
6040 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
6041 * emit-rtl.c (try_split): Same.
6043 2014-04-24 Radovan Obradovic <robradovic@mips.com>
6044 Tom de Vries <tom@codesourcery.com>
6046 * common.opt (fuse-caller-save): New option.
6048 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
6050 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
6051 elements for big-endian.
6053 2014-04-24 Richard Biener <rguenther@suse.de>
6055 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
6056 during TER and instead use the sepops interface for expanding
6057 non-GIMPLE_SINGLE_RHS.
6059 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6061 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
6062 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
6064 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6066 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
6067 assembler 64-bit option.
6068 * configure: Regenerate.
6070 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6072 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
6073 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
6074 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
6075 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
6076 (TARGET_CRYPTO): Take TARGET_SIMD into account.
6078 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6080 * config/aarch64/aarch64-builtins.c
6081 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
6082 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
6083 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
6084 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
6086 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
6087 (Vrevsuff): New mode attribute.
6089 2014-04-24 Terry Guo <terry.guo@arm.com>
6091 * config/arm/arm.h (machine_function): Define variable
6092 after_arm_reorg here.
6093 * config/arm/arm.c (after_arm_reorg): Remove the definition.
6094 (arm_split_constant): Update the way to access variable
6097 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
6099 2014-04-23 Tom de Vries <tom@codesourcery.com>
6101 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
6103 2014-04-23 David Malcolm <dmalcolm@redhat.com>
6105 * is-a.h: Update comments to reflect the following changes to the
6106 "pointerness" of the API, making the template parameter match the
6107 return type, allowing use of is-a.h with typedefs of pointers.
6108 (is_a_helper::cast): Return a T rather then a pointer to a T, so
6109 that the return type matches the parameter to the is_a_helper.
6111 (dyn_cast): Likewise.
6113 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
6114 pointer from the is-a.h API.
6116 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
6117 (is_a_helper <cgraph_node *>::test): ...this, matching change to
6119 (is_a_helper <varpool_node>::test): Likewise, convert to...
6120 (is_a_helper <varpool_node *>::test): ...this.
6122 (varpool_first_variable): Update for removal of implicit pointer
6123 from the is-a.h API.
6124 (varpool_next_variable): Likewise.
6125 (varpool_first_static_initializer): Likewise.
6126 (varpool_next_static_initializer): Likewise.
6127 (varpool_first_defined_variable): Likewise.
6128 (varpool_next_defined_variable): Likewise.
6129 (cgraph_first_defined_function): Likewise.
6130 (cgraph_next_defined_function): Likewise.
6131 (cgraph_first_function): Likewise.
6132 (cgraph_next_function): Likewise.
6133 (cgraph_first_function_with_gimple_body): Likewise.
6134 (cgraph_next_function_with_gimple_body): Likewise.
6135 (cgraph_alias_target): Likewise.
6136 (varpool_alias_target): Likewise.
6137 (cgraph_function_or_thunk_node): Likewise.
6138 (varpool_variable_node): Likewise.
6139 (symtab_real_symbol_p): Likewise.
6140 * cgraphunit.c (referred_to_p): Likewise.
6141 (analyze_functions): Likewise.
6142 (handle_alias_pairs): Likewise.
6143 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
6144 * gimple-ssa.h (gimple_vuse_op): Likewise.
6145 (gimple_vdef_op): Likewise.
6146 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
6147 * gimple.c (gimple_build_asm_1): Likewise.
6148 (gimple_build_try): Likewise.
6149 (gimple_build_resx): Likewise.
6150 (gimple_build_eh_dispatch): Likewise.
6151 (gimple_build_omp_for): Likewise.
6152 (gimple_omp_for_set_clauses): Likewise.
6154 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
6155 (is_a_helper <gimple_statement_asm *>::test): ...this.
6156 (is_a_helper <gimple_statement_bind>::test): Convert to...
6157 (is_a_helper <gimple_statement_bind *>::test): ...this.
6158 (is_a_helper <gimple_statement_call>::test): Convert to...
6159 (is_a_helper <gimple_statement_call *>::test): ...this.
6160 (is_a_helper <gimple_statement_catch>::test): Convert to...
6161 (is_a_helper <gimple_statement_catch *>::test): ...this.
6162 (is_a_helper <gimple_statement_resx>::test): Convert to...
6163 (is_a_helper <gimple_statement_resx *>::test): ...this.
6164 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
6165 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
6166 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
6167 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
6168 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
6169 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
6170 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
6171 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
6172 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
6173 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
6174 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
6175 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
6176 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
6177 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
6178 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
6179 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
6180 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
6181 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
6182 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
6183 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
6184 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
6185 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
6186 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
6187 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
6188 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
6189 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
6190 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
6191 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
6192 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
6193 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
6194 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
6195 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
6196 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
6197 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
6198 (is_a_helper <gimple_statement_phi>::test): Convert to...
6199 (is_a_helper <gimple_statement_phi *>::test): ...this.
6200 (is_a_helper <gimple_statement_transaction>::test): Convert to...
6201 (is_a_helper <gimple_statement_transaction *>::test): ...this.
6202 (is_a_helper <gimple_statement_try>::test): Convert to...
6203 (is_a_helper <gimple_statement_try *>::test): ...this.
6204 (is_a_helper <gimple_statement_wce>::test): Convert to...
6205 (is_a_helper <gimple_statement_wce *>::test): ...this.
6206 (is_a_helper <const gimple_statement_asm>::test): Convert to...
6207 (is_a_helper <const gimple_statement_asm *>::test): ...this.
6208 (is_a_helper <const gimple_statement_bind>::test): Convert to...
6209 (is_a_helper <const gimple_statement_bind *>::test): ...this.
6210 (is_a_helper <const gimple_statement_call>::test): Convert to...
6211 (is_a_helper <const gimple_statement_call *>::test): ...this.
6212 (is_a_helper <const gimple_statement_catch>::test): Convert to...
6213 (is_a_helper <const gimple_statement_catch *>::test): ...this.
6214 (is_a_helper <const gimple_statement_resx>::test): Convert to...
6215 (is_a_helper <const gimple_statement_resx *>::test): ...this.
6216 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
6217 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
6218 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
6219 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
6220 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
6222 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
6224 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
6226 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
6228 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
6229 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
6230 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
6232 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
6233 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
6235 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
6236 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
6237 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
6238 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
6239 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
6240 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
6242 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
6243 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
6244 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
6245 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
6247 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
6248 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
6249 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
6250 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
6251 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
6252 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
6253 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
6254 (is_a_helper <const gimple_statement_phi>::test): Convert to...
6255 (is_a_helper <const gimple_statement_phi *>::test): ...this.
6256 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
6257 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
6258 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
6259 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
6260 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
6261 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
6262 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
6264 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
6266 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
6267 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
6269 (gimple_use_ops): Update for removal of implicit pointer from the
6271 (gimple_set_use_ops): Likewise.
6272 (gimple_vuse): Likewise.
6273 (gimple_vdef): Likewise.
6274 (gimple_vuse_ptr): Likewise.
6275 (gimple_vdef_ptr): Likewise.
6276 (gimple_set_vuse): Likewise.
6277 (gimple_set_vdef): Likewise.
6278 (gimple_omp_return_set_lhs): Likewise.
6279 (gimple_omp_return_lhs): Likewise.
6280 (gimple_omp_return_lhs_ptr): Likewise.
6281 (gimple_call_fntype): Likewise.
6282 (gimple_call_set_fntype): Likewise.
6283 (gimple_call_set_internal_fn): Likewise.
6284 (gimple_call_use_set): Likewise.
6285 (gimple_call_clobber_set): Likewise.
6286 (gimple_bind_vars): Likewise.
6287 (gimple_bind_set_vars): Likewise.
6288 (gimple_bind_body_ptr): Likewise.
6289 (gimple_bind_set_body): Likewise.
6290 (gimple_bind_add_stmt): Likewise.
6291 (gimple_bind_block): Likewise.
6292 (gimple_bind_set_block): Likewise.
6293 (gimple_asm_ninputs): Likewise.
6294 (gimple_asm_noutputs): Likewise.
6295 (gimple_asm_nclobbers): Likewise.
6296 (gimple_asm_nlabels): Likewise.
6297 (gimple_asm_input_op): Likewise.
6298 (gimple_asm_input_op_ptr): Likewise.
6299 (gimple_asm_output_op): Likewise.
6300 (gimple_asm_output_op_ptr): Likewise.
6301 (gimple_asm_set_output_op): Likewise.
6302 (gimple_asm_clobber_op): Likewise.
6303 (gimple_asm_set_clobber_op): Likewise.
6304 (gimple_asm_label_op): Likewise.
6305 (gimple_asm_set_label_op): Likewise.
6306 (gimple_asm_string): Likewise.
6307 (gimple_catch_types): Likewise.
6308 (gimple_catch_types_ptr): Likewise.
6309 (gimple_catch_handler_ptr): Likewise.
6310 (gimple_catch_set_types): Likewise.
6311 (gimple_catch_set_handler): Likewise.
6312 (gimple_eh_filter_types): Likewise.
6313 (gimple_eh_filter_types_ptr): Likewise.
6314 (gimple_eh_filter_failure_ptr): Likewise.
6315 (gimple_eh_filter_set_types): Likewise.
6316 (gimple_eh_filter_set_failure): Likewise.
6317 (gimple_eh_must_not_throw_fndecl): Likewise.
6318 (gimple_eh_must_not_throw_set_fndecl): Likewise.
6319 (gimple_eh_else_n_body_ptr): Likewise.
6320 (gimple_eh_else_e_body_ptr): Likewise.
6321 (gimple_eh_else_set_n_body): Likewise.
6322 (gimple_eh_else_set_e_body): Likewise.
6323 (gimple_try_eval_ptr): Likewise.
6324 (gimple_try_cleanup_ptr): Likewise.
6325 (gimple_try_set_eval): Likewise.
6326 (gimple_try_set_cleanup): Likewise.
6327 (gimple_wce_cleanup_ptr): Likewise.
6328 (gimple_wce_set_cleanup): Likewise.
6329 (gimple_phi_capacity): Likewise.
6330 (gimple_phi_num_args): Likewise.
6331 (gimple_phi_result): Likewise.
6332 (gimple_phi_result_ptr): Likewise.
6333 (gimple_phi_set_result): Likewise.
6334 (gimple_phi_arg): Likewise.
6335 (gimple_phi_set_arg): Likewise.
6336 (gimple_resx_region): Likewise.
6337 (gimple_resx_set_region): Likewise.
6338 (gimple_eh_dispatch_region): Likewise.
6339 (gimple_eh_dispatch_set_region): Likewise.
6340 (gimple_omp_critical_name): Likewise.
6341 (gimple_omp_critical_name_ptr): Likewise.
6342 (gimple_omp_critical_set_name): Likewise.
6343 (gimple_omp_for_clauses): Likewise.
6344 (gimple_omp_for_clauses_ptr): Likewise.
6345 (gimple_omp_for_set_clauses): Likewise.
6346 (gimple_omp_for_collapse): Likewise.
6347 (gimple_omp_for_index): Likewise.
6348 (gimple_omp_for_index_ptr): Likewise.
6349 (gimple_omp_for_set_index): Likewise.
6350 (gimple_omp_for_initial): Likewise.
6351 (gimple_omp_for_initial_ptr): Likewise.
6352 (gimple_omp_for_set_initial): Likewise.
6353 (gimple_omp_for_final): Likewise.
6354 (gimple_omp_for_final_ptr): Likewise.
6355 (gimple_omp_for_set_final): Likewise.
6356 (gimple_omp_for_incr): Likewise.
6357 (gimple_omp_for_incr_ptr): Likewise.
6358 (gimple_omp_for_set_incr): Likewise.
6359 (gimple_omp_for_pre_body_ptr): Likewise.
6360 (gimple_omp_for_set_pre_body): Likewise.
6361 (gimple_omp_parallel_clauses): Likewise.
6362 (gimple_omp_parallel_clauses_ptr): Likewise.
6363 (gimple_omp_parallel_set_clauses): Likewise.
6364 (gimple_omp_parallel_child_fn): Likewise.
6365 (gimple_omp_parallel_child_fn_ptr): Likewise.
6366 (gimple_omp_parallel_set_child_fn): Likewise.
6367 (gimple_omp_parallel_data_arg): Likewise.
6368 (gimple_omp_parallel_data_arg_ptr): Likewise.
6369 (gimple_omp_parallel_set_data_arg): Likewise.
6370 (gimple_omp_task_clauses): Likewise.
6371 (gimple_omp_task_clauses_ptr): Likewise.
6372 (gimple_omp_task_set_clauses): Likewise.
6373 (gimple_omp_task_child_fn): Likewise.
6374 (gimple_omp_task_child_fn_ptr): Likewise.
6375 (gimple_omp_task_set_child_fn): Likewise.
6376 (gimple_omp_task_data_arg): Likewise.
6377 (gimple_omp_task_data_arg_ptr): Likewise.
6378 (gimple_omp_task_set_data_arg): Likewise.
6379 (gimple_omp_taskreg_clauses): Likewise.
6380 (gimple_omp_taskreg_clauses_ptr): Likewise.
6381 (gimple_omp_taskreg_set_clauses): Likewise.
6382 (gimple_omp_taskreg_child_fn): Likewise.
6383 (gimple_omp_taskreg_child_fn_ptr): Likewise.
6384 (gimple_omp_taskreg_set_child_fn): Likewise.
6385 (gimple_omp_taskreg_data_arg): Likewise.
6386 (gimple_omp_taskreg_data_arg_ptr): Likewise.
6387 (gimple_omp_taskreg_set_data_arg): Likewise.
6388 (gimple_omp_task_copy_fn): Likewise.
6389 (gimple_omp_task_copy_fn_ptr): Likewise.
6390 (gimple_omp_task_set_copy_fn): Likewise.
6391 (gimple_omp_task_arg_size): Likewise.
6392 (gimple_omp_task_arg_size_ptr): Likewise.
6393 (gimple_omp_task_set_arg_size): Likewise.
6394 (gimple_omp_task_arg_align): Likewise.
6395 (gimple_omp_task_arg_align_ptr): Likewise.
6396 (gimple_omp_task_set_arg_align): Likewise.
6397 (gimple_omp_single_clauses): Likewise.
6398 (gimple_omp_single_clauses_ptr): Likewise.
6399 (gimple_omp_single_set_clauses): Likewise.
6400 (gimple_omp_target_clauses): Likewise.
6401 (gimple_omp_target_clauses_ptr): Likewise.
6402 (gimple_omp_target_set_clauses): Likewise.
6403 (gimple_omp_target_child_fn): Likewise.
6404 (gimple_omp_target_child_fn_ptr): Likewise.
6405 (gimple_omp_target_set_child_fn): Likewise.
6406 (gimple_omp_target_data_arg): Likewise.
6407 (gimple_omp_target_data_arg_ptr): Likewise.
6408 (gimple_omp_target_set_data_arg): Likewise.
6409 (gimple_omp_teams_clauses): Likewise.
6410 (gimple_omp_teams_clauses_ptr): Likewise.
6411 (gimple_omp_teams_set_clauses): Likewise.
6412 (gimple_omp_sections_clauses): Likewise.
6413 (gimple_omp_sections_clauses_ptr): Likewise.
6414 (gimple_omp_sections_set_clauses): Likewise.
6415 (gimple_omp_sections_control): Likewise.
6416 (gimple_omp_sections_control_ptr): Likewise.
6417 (gimple_omp_sections_set_control): Likewise.
6418 (gimple_omp_for_set_cond): Likewise.
6419 (gimple_omp_for_cond): Likewise.
6420 (gimple_omp_atomic_store_set_val): Likewise.
6421 (gimple_omp_atomic_store_val): Likewise.
6422 (gimple_omp_atomic_store_val_ptr): Likewise.
6423 (gimple_omp_atomic_load_set_lhs): Likewise.
6424 (gimple_omp_atomic_load_lhs): Likewise.
6425 (gimple_omp_atomic_load_lhs_ptr): Likewise.
6426 (gimple_omp_atomic_load_set_rhs): Likewise.
6427 (gimple_omp_atomic_load_rhs): Likewise.
6428 (gimple_omp_atomic_load_rhs_ptr): Likewise.
6429 (gimple_omp_continue_control_def): Likewise.
6430 (gimple_omp_continue_control_def_ptr): Likewise.
6431 (gimple_omp_continue_set_control_def): Likewise.
6432 (gimple_omp_continue_control_use): Likewise.
6433 (gimple_omp_continue_control_use_ptr): Likewise.
6434 (gimple_omp_continue_set_control_use): Likewise.
6435 (gimple_transaction_body_ptr): Likewise.
6436 (gimple_transaction_label): Likewise.
6437 (gimple_transaction_label_ptr): Likewise.
6438 (gimple_transaction_set_body): Likewise.
6439 (gimple_transaction_set_label): Likewise.
6441 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
6442 * ipa-inline-analysis.c (inline_write_summary): Likewise.
6443 * ipa-ref.c (ipa_record_reference): Likewise.
6444 * ipa-reference.c (analyze_function): Likewise.
6445 (ipa_reference_write_optimization_summary): Likewise.
6446 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
6447 (address_taken_from_non_vtable_p): Likewise.
6448 (comdat_can_be_unshared_p_1): Likewise.
6449 * lto-cgraph.c (lto_output_ref): Likewise.
6450 (add_references): Likewise.
6451 (compute_ltrans_boundary): Likewise.
6452 (output_symtab): Likewise.
6453 (input_ref): Likewise.
6454 (input_cgraph_1): Likewise.
6455 (output_cgraph_opt_summary): Likewise.
6456 * lto-streamer-out.c (lto_output): Likewise.
6457 (output_symbol_p): Likewise.
6458 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
6459 (lsei_start_function_in_partition): Likewise.
6460 (lsei_next_variable_in_partition): Likewise.
6461 (lsei_start_variable_in_partition): Likewise.
6462 * symtab.c (insert_to_assembler_name_hash): Likewise.
6463 (unlink_from_assembler_name_hash): Likewise.
6464 (symtab_unregister_node): Likewise.
6465 (symtab_remove_node): Likewise.
6466 (dump_symtab_node): Likewise.
6467 (verify_symtab_base): Likewise.
6468 (verify_symtab_node): Likewise.
6469 (symtab_make_decl_local): Likewise.
6470 (symtab_alias_ultimate_target): Likewise.
6471 (symtab_resolve_alias): Likewise.
6472 (symtab_get_symbol_partitioning_class): Likewise.
6473 * tree-phinodes.c (allocate_phi_node): Likewise.
6474 (reserve_phi_args_for_new_edge): Likewise.
6475 (remove_phi_args): Likewise.
6476 * varpool.c (varpool_node_for_asm): Likewise.
6477 (varpool_remove_unreferenced_decls): Likewise.
6479 2014-04-23 Jeff Law <law@redhat.com>
6481 PR tree-optimization/60902
6482 * tree-ssa-threadedge.c
6483 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
6484 invalidate outputs from statements that do not produce useful
6485 outputs for threading.
6487 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6489 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
6490 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
6491 machine descriptions for Stack Smashing Protector.
6493 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
6495 * aarch64.md (<optab>_rol<mode>3): New pattern.
6496 (<optab>_rolsi3_uxtw): Likewise.
6497 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
6499 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
6501 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
6502 (arm_cortex_a12_tune): Likewise.
6504 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6506 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
6508 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6510 * config/arm/arm.md (arm_rev16si2): New pattern.
6511 (arm_rev16si2_alt): Likewise.
6512 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
6514 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6516 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
6517 (rev16<mode>2_alt): Likewise.
6518 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
6519 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
6520 (aarch_rev16_shleft_mask_imm_p): Likewise.
6521 (aarch_rev16_p_1): Likewise.
6522 (aarch_rev16_p): Likewise.
6523 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
6524 (aarch_rev16_shright_mask_imm_p): Likewise.
6525 (aarch_rev16_shleft_mask_imm_p): Likewise.
6527 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6529 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
6530 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
6532 (cortex_a53_extra_costs): Likewise.
6533 (cortex_a57_extra_costs): Likewise.
6534 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
6535 (cortexa7_extra_costs): Likewise.
6536 (cortexa8_extra_costs): Likewise.
6537 (cortexa12_extra_costs): Likewise.
6538 (cortexa15_extra_costs): Likewise.
6539 (v7m_extra_costs): Likewise.
6540 (arm_new_rtx_costs): Handle BSWAP.
6542 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6544 * config/arm/arm.c (cortexa8_extra_costs): New table.
6545 (arm_cortex_a8_tune): New tuning struct.
6546 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
6548 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6550 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
6552 2014-04-23 Richard Biener <rguenther@suse.de>
6554 * Makefile.in (OBJS): Remove loop-unswitch.o.
6555 * tree-pass.h (make_pass_rtl_unswitch): Remove.
6556 * passes.def (pass_rtl_unswitch): Likewise.
6557 * loop-init.c (gate_rtl_unswitch): Likewise.
6558 (rtl_unswitch): Likewise.
6559 (pass_data_rtl_unswitch): Likewise.
6560 (pass_rtl_unswitch): Likewise.
6561 (make_pass_rtl_unswitch): Likewise.
6562 * rtl.h (reversed_condition): Likewise.
6563 (compare_and_jump_seq): Likewise.
6564 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
6566 * loop-unroll.c (compare_and_jump_seq): Likewise.
6568 2014-04-23 Richard Biener <rguenther@suse.de>
6570 PR tree-optimization/60903
6571 * tree-ssa-loop-im.c (analyze_memory_references): Remove
6572 commented code block.
6573 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
6574 loop flags to newly created BBs and edges.
6576 2014-04-23 Nick Clifton <nickc@redhat.com>
6578 * config/msp430/msp430.c (msp430_handle_option): Move function
6580 (msp430_option_override): Simplify mcu and mcpu option handling.
6581 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
6582 support for -mhwmult command line option.
6583 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
6584 -mhwmult command line option.
6585 (msp430_hwmult_enabled): Delete.
6586 (msp43o_output_labelref): Add support for -mhwmult command line option.
6587 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
6588 (umulsidi3): Likewise.
6589 * config/msp430/msp430.opt (mmcu): Add Report attribute.
6590 (mcpu, mlarge, msmall): Likewise.
6591 (mhwmult): New option.
6592 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
6594 (msp430_is_f5_mcu): Remove prototype.
6595 (msp430_use_f5_series_hwmult): Add prototype.
6596 * config/msp430/msp430-opts.h: New file.
6597 * common/config/msp430: New directory.
6598 * common/config/msp430/msp430-common.c: New file.
6599 * config.gcc (msp430): Remove target_has_targetm_common.
6600 * doc/invoke.texi: Document -mhwmult command line option.
6602 2014-04-23 Nick Clifton <nickc@redhat.com>
6604 * config/i386/cygwin.h (ENDFILE_SPEC): Include
6605 default-manifest.o if it can be found in the search path.
6606 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
6608 2014-04-23 Terry Guo <terry.guo@arm.com>
6610 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
6612 2014-04-23 Richard Biener <rguenther@suse.de>
6615 * tree-inline.c (declare_return_variable): Use mark_addressable.
6617 2014-04-23 Richard Biener <rguenther@suse.de>
6620 * loop-init.c (loop_optimizer_init): Make sure to apply
6621 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
6623 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6626 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
6628 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
6629 if flag_sanitize_undefined_trap_on_error.
6630 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
6631 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
6632 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
6633 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
6634 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
6635 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
6636 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
6637 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
6638 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
6639 * ubsan.c (ubsan_instrument_unreachable): Return
6640 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
6641 (ubsan_expand_null_ifn): Emit __builtin_trap ()
6642 if flag_sanitize_undefined_trap_on_error and
6643 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
6644 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
6645 instrument_bool_enum_load): Emit __builtin_trap () if
6646 flag_sanitize_undefined_trap_on_error and
6647 __builtin_handle_*_abort () if !flag_sanitize_recover.
6648 * doc/invoke.texi (-fsanitize-recover,
6649 -fsanitize-undefined-trap-on-error): Document.
6651 2014-04-22 Christian Bruel <christian.bruel@st.com>
6653 * config/sh/sh.md (mov<mode>): Replace movQIHI.
6654 Force immediates to SImode.
6656 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
6658 * config/nios2/nios2.md (UNSPEC_ROUND): New.
6660 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
6661 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
6662 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
6663 (nios2_fpu_insn): Add entry for round.
6664 (N2FPU_NO_ERRNO_P): Define.
6665 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
6667 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
6669 2014-04-22 Richard Henderson <rth@redhat.com>
6671 * config/aarch64/aarch64 (addti3, subti3): New expanders.
6672 (add<GPI>3_compare0): Remove leading * from name.
6673 (add<GPI>3_carryin): Likewise.
6674 (sub<GPI>3_compare0): Likewise.
6675 (sub<GPI>3_carryin): Likewise.
6676 (<su_optab>mulditi3): New expander.
6677 (multi3): New expander.
6678 (madd<GPI>): Remove leading * from name.
6680 2014-04-22 Martin Jambor <mjambor@suse.cz>
6682 * cgraphclones.c (cgraph_function_versioning): Copy
6683 ipa_transforms_to_apply instead of asserting it is empty.
6685 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
6688 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
6689 on count_exp to get mode.
6691 2014-04-22 Andrew Pinski <apinski@cavium.com>
6693 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6694 Handle TLS for ILP32.
6695 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
6696 (tlsie_small_<mode>): this and handle PTR.
6697 (tlsie_small_sidi): New pattern.
6698 (tlsle_small): Change to an expand to handle ILP32.
6699 (tlsle_small_<mode>): New pattern.
6700 (tlsdesc_small): Rename to ...
6701 (tlsdesc_small_<mode>): this and handle PTR.
6703 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6705 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
6707 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6709 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6710 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
6711 (aarch64_types_signed_poly_qualifiers): Likewise.
6712 (aarch64_types_unsigned_signed_qualifiers): Likewise.
6713 (aarch64_types_poly_signed_qualifiers): Likewise.
6714 (TYPES_REINTERP_SS): Type macro added.
6715 (TYPES_REINTERP_SU): Likewise.
6716 (TYPES_REINTERP_SP): Likewise.
6717 (TYPES_REINTERP_US): Likewise.
6718 (TYPES_REINTERP_PS): Likewise.
6719 (aarch64_fold_builtin): New expression folding added.
6720 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
6721 Declarations removed.
6722 (REINTERP_SS): Declarations added.
6723 (REINTERP_US): Likewise.
6724 (REINTERP_PS): Likewise.
6725 (REINTERP_SU): Likewise.
6726 (REINTERP_SP): Likewise.
6727 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
6728 (vreinterpretq_p8_f64): Likewise.
6729 (vreinterpret_p16_f64): Likewise.
6730 (vreinterpretq_p16_f64): Likewise.
6731 (vreinterpret_f32_f64): Likewise.
6732 (vreinterpretq_f32_f64): Likewise.
6733 (vreinterpret_f64_f32): Likewise.
6734 (vreinterpret_f64_p8): Likewise.
6735 (vreinterpret_f64_p16): Likewise.
6736 (vreinterpret_f64_s8): Likewise.
6737 (vreinterpret_f64_s16): Likewise.
6738 (vreinterpret_f64_s32): Likewise.
6739 (vreinterpret_f64_s64): Likewise.
6740 (vreinterpret_f64_u8): Likewise.
6741 (vreinterpret_f64_u16): Likewise.
6742 (vreinterpret_f64_u32): Likewise.
6743 (vreinterpret_f64_u64): Likewise.
6744 (vreinterpretq_f64_f32): Likewise.
6745 (vreinterpretq_f64_p8): Likewise.
6746 (vreinterpretq_f64_p16): Likewise.
6747 (vreinterpretq_f64_s8): Likewise.
6748 (vreinterpretq_f64_s16): Likewise.
6749 (vreinterpretq_f64_s32): Likewise.
6750 (vreinterpretq_f64_s64): Likewise.
6751 (vreinterpretq_f64_u8): Likewise.
6752 (vreinterpretq_f64_u16): Likewise.
6753 (vreinterpretq_f64_u32): Likewise.
6754 (vreinterpretq_f64_u64): Likewise.
6755 (vreinterpret_s64_f64): Likewise.
6756 (vreinterpretq_s64_f64): Likewise.
6757 (vreinterpret_u64_f64): Likewise.
6758 (vreinterpretq_u64_f64): Likewise.
6759 (vreinterpret_s8_f64): Likewise.
6760 (vreinterpretq_s8_f64): Likewise.
6761 (vreinterpret_s16_f64): Likewise.
6762 (vreinterpretq_s16_f64): Likewise.
6763 (vreinterpret_s32_f64): Likewise.
6764 (vreinterpretq_s32_f64): Likewise.
6765 (vreinterpret_u8_f64): Likewise.
6766 (vreinterpretq_u8_f64): Likewise.
6767 (vreinterpret_u16_f64): Likewise.
6768 (vreinterpretq_u16_f64): Likewise.
6769 (vreinterpret_u32_f64): Likewise.
6770 (vreinterpretq_u32_f64): Likewise.
6772 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6774 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6775 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
6776 (vreinterpret_p8_s8): Likewise.
6777 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
6778 (vreinterpret_p8_s16): Likewise.
6779 (vreinterpret_p8_s32): Likewise.
6780 (vreinterpret_p8_s64): Likewise.
6781 (vreinterpret_p8_f32): Likewise.
6782 (vreinterpret_p8_u8): Likewise.
6783 (vreinterpret_p8_u16): Likewise.
6784 (vreinterpret_p8_u32): Likewise.
6785 (vreinterpret_p8_u64): Likewise.
6786 (vreinterpret_p8_p16): Likewise.
6787 (vreinterpretq_p8_s8): Likewise.
6788 (vreinterpretq_p8_s16): Likewise.
6789 (vreinterpretq_p8_s32): Likewise.
6790 (vreinterpretq_p8_s64): Likewise.
6791 (vreinterpretq_p8_f32): Likewise.
6792 (vreinterpretq_p8_u8): Likewise.
6793 (vreinterpretq_p8_u16): Likewise.
6794 (vreinterpretq_p8_u32): Likewise.
6795 (vreinterpretq_p8_u64): Likewise.
6796 (vreinterpretq_p8_p16): Likewise.
6797 (vreinterpret_p16_s8): Likewise.
6798 (vreinterpret_p16_s16): Likewise.
6799 (vreinterpret_p16_s32): Likewise.
6800 (vreinterpret_p16_s64): Likewise.
6801 (vreinterpret_p16_f32): Likewise.
6802 (vreinterpret_p16_u8): Likewise.
6803 (vreinterpret_p16_u16): Likewise.
6804 (vreinterpret_p16_u32): Likewise.
6805 (vreinterpret_p16_u64): Likewise.
6806 (vreinterpret_p16_p8): Likewise.
6807 (vreinterpretq_p16_s8): Likewise.
6808 (vreinterpretq_p16_s16): Likewise.
6809 (vreinterpretq_p16_s32): Likewise.
6810 (vreinterpretq_p16_s64): Likewise.
6811 (vreinterpretq_p16_f32): Likewise.
6812 (vreinterpretq_p16_u8): Likewise.
6813 (vreinterpretq_p16_u16): Likewise.
6814 (vreinterpretq_p16_u32): Likewise.
6815 (vreinterpretq_p16_u64): Likewise.
6816 (vreinterpretq_p16_p8): Likewise.
6817 (vreinterpret_f32_s8): Likewise.
6818 (vreinterpret_f32_s16): Likewise.
6819 (vreinterpret_f32_s32): Likewise.
6820 (vreinterpret_f32_s64): Likewise.
6821 (vreinterpret_f32_u8): Likewise.
6822 (vreinterpret_f32_u16): Likewise.
6823 (vreinterpret_f32_u32): Likewise.
6824 (vreinterpret_f32_u64): Likewise.
6825 (vreinterpret_f32_p8): Likewise.
6826 (vreinterpret_f32_p16): Likewise.
6827 (vreinterpretq_f32_s8): Likewise.
6828 (vreinterpretq_f32_s16): Likewise.
6829 (vreinterpretq_f32_s32): Likewise.
6830 (vreinterpretq_f32_s64): Likewise.
6831 (vreinterpretq_f32_u8): Likewise.
6832 (vreinterpretq_f32_u16): Likewise.
6833 (vreinterpretq_f32_u32): Likewise.
6834 (vreinterpretq_f32_u64): Likewise.
6835 (vreinterpretq_f32_p8): Likewise.
6836 (vreinterpretq_f32_p16): Likewise.
6837 (vreinterpret_s64_s8): Likewise.
6838 (vreinterpret_s64_s16): Likewise.
6839 (vreinterpret_s64_s32): Likewise.
6840 (vreinterpret_s64_f32): Likewise.
6841 (vreinterpret_s64_u8): Likewise.
6842 (vreinterpret_s64_u16): Likewise.
6843 (vreinterpret_s64_u32): Likewise.
6844 (vreinterpret_s64_u64): Likewise.
6845 (vreinterpret_s64_p8): Likewise.
6846 (vreinterpret_s64_p16): Likewise.
6847 (vreinterpretq_s64_s8): Likewise.
6848 (vreinterpretq_s64_s16): Likewise.
6849 (vreinterpretq_s64_s32): Likewise.
6850 (vreinterpretq_s64_f32): Likewise.
6851 (vreinterpretq_s64_u8): Likewise.
6852 (vreinterpretq_s64_u16): Likewise.
6853 (vreinterpretq_s64_u32): Likewise.
6854 (vreinterpretq_s64_u64): Likewise.
6855 (vreinterpretq_s64_p8): Likewise.
6856 (vreinterpretq_s64_p16): Likewise.
6857 (vreinterpret_u64_s8): Likewise.
6858 (vreinterpret_u64_s16): Likewise.
6859 (vreinterpret_u64_s32): Likewise.
6860 (vreinterpret_u64_s64): Likewise.
6861 (vreinterpret_u64_f32): Likewise.
6862 (vreinterpret_u64_u8): Likewise.
6863 (vreinterpret_u64_u16): Likewise.
6864 (vreinterpret_u64_u32): Likewise.
6865 (vreinterpret_u64_p8): Likewise.
6866 (vreinterpret_u64_p16): Likewise.
6867 (vreinterpretq_u64_s8): Likewise.
6868 (vreinterpretq_u64_s16): Likewise.
6869 (vreinterpretq_u64_s32): Likewise.
6870 (vreinterpretq_u64_s64): Likewise.
6871 (vreinterpretq_u64_f32): Likewise.
6872 (vreinterpretq_u64_u8): Likewise.
6873 (vreinterpretq_u64_u16): Likewise.
6874 (vreinterpretq_u64_u32): Likewise.
6875 (vreinterpretq_u64_p8): Likewise.
6876 (vreinterpretq_u64_p16): Likewise.
6877 (vreinterpret_s8_s16): Likewise.
6878 (vreinterpret_s8_s32): Likewise.
6879 (vreinterpret_s8_s64): Likewise.
6880 (vreinterpret_s8_f32): Likewise.
6881 (vreinterpret_s8_u8): Likewise.
6882 (vreinterpret_s8_u16): Likewise.
6883 (vreinterpret_s8_u32): Likewise.
6884 (vreinterpret_s8_u64): Likewise.
6885 (vreinterpret_s8_p8): Likewise.
6886 (vreinterpret_s8_p16): Likewise.
6887 (vreinterpretq_s8_s16): Likewise.
6888 (vreinterpretq_s8_s32): Likewise.
6889 (vreinterpretq_s8_s64): Likewise.
6890 (vreinterpretq_s8_f32): Likewise.
6891 (vreinterpretq_s8_u8): Likewise.
6892 (vreinterpretq_s8_u16): Likewise.
6893 (vreinterpretq_s8_u32): Likewise.
6894 (vreinterpretq_s8_u64): Likewise.
6895 (vreinterpretq_s8_p8): Likewise.
6896 (vreinterpretq_s8_p16): Likewise.
6897 (vreinterpret_s16_s8): Likewise.
6898 (vreinterpret_s16_s32): Likewise.
6899 (vreinterpret_s16_s64): Likewise.
6900 (vreinterpret_s16_f32): Likewise.
6901 (vreinterpret_s16_u8): Likewise.
6902 (vreinterpret_s16_u16): Likewise.
6903 (vreinterpret_s16_u32): Likewise.
6904 (vreinterpret_s16_u64): Likewise.
6905 (vreinterpret_s16_p8): Likewise.
6906 (vreinterpret_s16_p16): Likewise.
6907 (vreinterpretq_s16_s8): Likewise.
6908 (vreinterpretq_s16_s32): Likewise.
6909 (vreinterpretq_s16_s64): Likewise.
6910 (vreinterpretq_s16_f32): Likewise.
6911 (vreinterpretq_s16_u8): Likewise.
6912 (vreinterpretq_s16_u16): Likewise.
6913 (vreinterpretq_s16_u32): Likewise.
6914 (vreinterpretq_s16_u64): Likewise.
6915 (vreinterpretq_s16_p8): Likewise.
6916 (vreinterpretq_s16_p16): Likewise.
6917 (vreinterpret_s32_s8): Likewise.
6918 (vreinterpret_s32_s16): Likewise.
6919 (vreinterpret_s32_s64): Likewise.
6920 (vreinterpret_s32_f32): Likewise.
6921 (vreinterpret_s32_u8): Likewise.
6922 (vreinterpret_s32_u16): Likewise.
6923 (vreinterpret_s32_u32): Likewise.
6924 (vreinterpret_s32_u64): Likewise.
6925 (vreinterpret_s32_p8): Likewise.
6926 (vreinterpret_s32_p16): Likewise.
6927 (vreinterpretq_s32_s8): Likewise.
6928 (vreinterpretq_s32_s16): Likewise.
6929 (vreinterpretq_s32_s64): Likewise.
6930 (vreinterpretq_s32_f32): Likewise.
6931 (vreinterpretq_s32_u8): Likewise.
6932 (vreinterpretq_s32_u16): Likewise.
6933 (vreinterpretq_s32_u32): Likewise.
6934 (vreinterpretq_s32_u64): Likewise.
6935 (vreinterpretq_s32_p8): Likewise.
6936 (vreinterpretq_s32_p16): Likewise.
6937 (vreinterpret_u8_s8): Likewise.
6938 (vreinterpret_u8_s16): Likewise.
6939 (vreinterpret_u8_s32): Likewise.
6940 (vreinterpret_u8_s64): Likewise.
6941 (vreinterpret_u8_f32): Likewise.
6942 (vreinterpret_u8_u16): Likewise.
6943 (vreinterpret_u8_u32): Likewise.
6944 (vreinterpret_u8_u64): Likewise.
6945 (vreinterpret_u8_p8): Likewise.
6946 (vreinterpret_u8_p16): Likewise.
6947 (vreinterpretq_u8_s8): Likewise.
6948 (vreinterpretq_u8_s16): Likewise.
6949 (vreinterpretq_u8_s32): Likewise.
6950 (vreinterpretq_u8_s64): Likewise.
6951 (vreinterpretq_u8_f32): Likewise.
6952 (vreinterpretq_u8_u16): Likewise.
6953 (vreinterpretq_u8_u32): Likewise.
6954 (vreinterpretq_u8_u64): Likewise.
6955 (vreinterpretq_u8_p8): Likewise.
6956 (vreinterpretq_u8_p16): Likewise.
6957 (vreinterpret_u16_s8): Likewise.
6958 (vreinterpret_u16_s16): Likewise.
6959 (vreinterpret_u16_s32): Likewise.
6960 (vreinterpret_u16_s64): Likewise.
6961 (vreinterpret_u16_f32): Likewise.
6962 (vreinterpret_u16_u8): Likewise.
6963 (vreinterpret_u16_u32): Likewise.
6964 (vreinterpret_u16_u64): Likewise.
6965 (vreinterpret_u16_p8): Likewise.
6966 (vreinterpret_u16_p16): Likewise.
6967 (vreinterpretq_u16_s8): Likewise.
6968 (vreinterpretq_u16_s16): Likewise.
6969 (vreinterpretq_u16_s32): Likewise.
6970 (vreinterpretq_u16_s64): Likewise.
6971 (vreinterpretq_u16_f32): Likewise.
6972 (vreinterpretq_u16_u8): Likewise.
6973 (vreinterpretq_u16_u32): Likewise.
6974 (vreinterpretq_u16_u64): Likewise.
6975 (vreinterpretq_u16_p8): Likewise.
6976 (vreinterpretq_u16_p16): Likewise.
6977 (vreinterpret_u32_s8): Likewise.
6978 (vreinterpret_u32_s16): Likewise.
6979 (vreinterpret_u32_s32): Likewise.
6980 (vreinterpret_u32_s64): Likewise.
6981 (vreinterpret_u32_f32): Likewise.
6982 (vreinterpret_u32_u8): Likewise.
6983 (vreinterpret_u32_u16): Likewise.
6984 (vreinterpret_u32_u64): Likewise.
6985 (vreinterpret_u32_p8): Likewise.
6986 (vreinterpret_u32_p16): Likewise.
6987 (vreinterpretq_u32_s8): Likewise.
6988 (vreinterpretq_u32_s16): Likewise.
6989 (vreinterpretq_u32_s32): Likewise.
6990 (vreinterpretq_u32_s64): Likewise.
6991 (vreinterpretq_u32_f32): Likewise.
6992 (vreinterpretq_u32_u8): Likewise.
6993 (vreinterpretq_u32_u16): Likewise.
6994 (vreinterpretq_u32_u64): Likewise.
6995 (vreinterpretq_u32_p8): Likewise.
6996 (vreinterpretq_u32_p16): Likewise.
6998 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
7000 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
7002 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
7004 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
7005 (vqnegd_s64): Likewise.
7006 (vqabs_s64): Likewise.
7007 (vqabsd_s64): Likewise.
7009 2014-04-22 Richard Henderson <rth@redhat.com>
7011 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
7012 computation to the top of the loop.
7014 2014-04-22 Renlin <renlin.li@arm.com>
7015 Jiong Wang <jiong.wang@arm.com>
7017 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
7018 * config/aarch64/aarch64.c (aarch64_layout_frame)
7019 (aarch64_initial_elimination_offset): Likewise.
7021 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
7023 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
7026 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
7028 * machmode.h (bitwise_mode_for_mode): Declare.
7029 * stor-layout.h (bitwise_type_for_mode): Likewise.
7030 * stor-layout.c (bitwise_mode_for_mode): New function.
7031 (bitwise_type_for_mode): Likewise.
7032 * builtins.c (fold_builtin_memory_op): Use it instead of
7033 int_mode_for_mode and build_nonstandard_integer_type.
7035 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7037 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
7038 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
7039 (*-*-solaris2*): Simplify.
7040 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
7041 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
7042 *-*-solaris2.9* handling.
7044 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
7046 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
7047 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
7049 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
7050 * configure: Regenerate.
7052 * config/i386/sol2-9.h: Remove.
7054 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
7055 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
7056 Remove Solaris 9 references.
7058 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
7060 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
7061 (floatuns<GPI:mode><GPF:mode>2): Remove.
7062 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
7063 and floatuns conversions.
7064 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
7065 and floatuns conversions.
7066 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
7067 (w1,w2): New mode attributes for inequal width conversions.
7069 2014-04-22 Renlin Li <Renlin.Li@arm.com>
7071 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
7072 the output asm format.
7074 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
7076 * config/aarch64/aarch64-simd.md
7077 (aarch64_cm<optab>di): Always split.
7078 (*aarch64_cm<optab>di): New.
7079 (aarch64_cmtstdi): Always split.
7080 (*aarch64_cmtstdi): New.
7082 2014-04-22 Jakub Jelinek <jakub@redhat.com>
7084 PR tree-optimization/60823
7085 * omp-low.c (ipa_simd_modify_function_body): Go through
7086 all SSA_NAMEs and for those refering to vector arguments
7087 which are going to be replaced adjust SSA_NAME_VAR and,
7088 if it is a default definition, change it into a non-default
7089 definition assigned at the beginning of function from new_decl.
7090 (ipa_simd_modify_stmt_ops): Rewritten.
7091 * tree-dfa.c (set_ssa_default_def): When removing default def,
7092 check for NULL loc instead of NULL *loc.
7094 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7096 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
7097 restrictions on core registers for DImode values in Thumb2.
7099 2014-04-22 Ian Bolton <ian.bolton@arm.com>
7101 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
7102 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
7104 2014-04-22 Ian Bolton <ian.bolton@arm.com>
7106 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
7107 (*iordi_notzesidi_di): Likewise.
7108 (*iordi_notsesidi_di): Likewise.
7110 2014-04-22 Ian Bolton <ian.bolton@arm.com>
7112 * config/arm/arm-protos.h (tune_params): New struct members.
7113 * config/arm/arm.c: Initialise tune_params per processor.
7114 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
7115 for speed, based on new tune_params.
7117 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
7119 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
7120 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
7121 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
7122 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
7123 * config/aarch64/arm_neon.h (vrnd_f64): Added.
7124 (vrnda_f64): Likewise.
7125 (vrndi_f64): Likewise.
7126 (vrndm_f64): Likewise.
7127 (vrndn_f64): Likewise.
7128 (vrndp_f64): Likewise.
7129 (vrndx_f64): Likewise.
7131 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7133 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
7134 GET_MODE_SIZE argument is enum machine_mode.
7136 2014-04-22 Jakub Jelinek <jakub@redhat.com>
7139 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
7140 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
7142 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
7145 * asan.c (asan_emit_stack_protection): Force the base to align to
7146 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
7147 appropriate bits if STRICT_ALIGNMENT.
7148 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
7150 (expand_used_vars): Leave a space in the stack frame for alignment
7151 if STRICT_ALIGNMENT.
7153 2014-04-21 David Malcolm <dmalcolm@redhat.com>
7155 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
7157 (gimple_store_p): Likewise.
7158 (gimple_assign_load_p): Likewise.
7159 (gimple_assign_cast_p): Likewise.
7160 (gimple_clobber_p): Likewise.
7162 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
7163 rather than a gimple.
7164 (gimple_assign_cast_p): Likewise.
7166 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
7169 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
7170 If mode is DDmode and TARGET_E500_DOUBLE allow move.
7172 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
7173 more debug information for E500 if -mdebug=reg.
7175 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
7178 * config/i386/i386.c (ix86_expand_builtin)
7179 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
7180 register for target RTX.
7181 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
7183 2014-04-18 Cong Hou <congh@google.com>
7185 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
7186 the widen-mult pattern by handling two operands with different sizes,
7187 and operands whose size is smaller than half of the result type.
7189 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
7191 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
7192 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
7193 (do_estimate_edge_time): Compute it.
7194 * ipa-inline.c (want_inline_small_function_p): Bypass
7195 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
7197 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
7199 * ipa-inline.c (spec_rem): New static variable.
7200 (dump_overall_stats): New function.
7201 (dump_inline_stats): New function.
7203 2014-04-18 Richard Henderson <rth@redhat.com>
7205 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
7206 to GET_MODE_SIZE, not a reg_class_t.
7208 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7210 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
7211 (vsx_xxmrglw_<mode>): Likewise.
7213 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
7216 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
7217 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
7218 (rs6000_init_hard_regno_mode_ok): Likewise.
7220 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
7222 * ipa-inline.c (inline_small_functions): Account only non-cold
7224 * doc/invoke.texi (inline-unit-growth): Update documentation.
7226 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
7228 * config/rs6000/rs6000.md (addti3, subti3): New.
7230 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
7233 * config/i386/i386.c (ix86_expand_clear): Remove outdated
7234 comment. Check optimize_insn_for_size_p instead of
7235 optimize_insn_for_speed_p.
7237 2014-04-17 Martin Jambor <mjambor@suse.cz>
7239 * gimple-iterator.c (gsi_start_edge): New function.
7240 * gimple-iterator.h (gsi_start_edge): Declare.
7241 * tree-sra.c (single_non_eh_succ): New function.
7242 (disqualify_ops_if_throwing_stmt): Renamed to
7243 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
7244 having one non-EH successor BB.
7245 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
7246 generate loads into replacements.
7247 (sra_modify_assign): Likewise and and also use the simple path for
7249 (sra_modify_function_body): Commit statements on edges.
7251 2014-04-17 Richard Biener <rguenther@suse.de>
7254 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
7255 comparison results and add clarifying comment.
7257 2014-04-17 Jakub Jelinek <jakub@redhat.com>
7259 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
7260 (blank_mode): Initialize it.
7261 (emit_mode_size_inline, emit_mode_nunits_inline,
7262 emit_mode_inner_inline): New functions.
7263 (emit_insn_modes_h): Call them and surround their output with
7264 #if GCC_VERSION >= 4001 ... #endif.
7265 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
7266 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
7267 mode_* arrays if the argument is __builtin_constant_p.
7268 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
7269 is enum machine_mode.
7271 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7273 * passes.c (opt_pass::execute): Adjust.
7274 (pass_manager::execute_pass_mode_switching): Likewise.
7275 (early_local_passes::execute): Likewise.
7276 (execute_one_pass): Pass cfun to the pass's execute method.
7277 * tree-pass.h (opt_pass::execute): Add function * argument.
7278 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
7279 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
7280 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
7281 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
7282 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
7283 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
7284 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
7285 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
7286 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
7287 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
7288 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
7289 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
7290 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
7291 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
7292 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
7293 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
7294 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
7295 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
7296 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
7297 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7298 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
7299 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
7300 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
7301 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7302 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7303 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
7304 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
7305 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
7306 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
7309 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7311 * passes.c (opt_pass::gate): Take function * argument.
7312 (gate_all_early_local_passes): Merge into
7313 (early_local_passes::gate): this.
7314 (gate_all_early_optimizations): Merge into
7315 (all_early_optimizations::gate): this.
7316 (gate_all_optimizations): Mege into
7317 (all_optimizations::gate): this.
7318 (gate_all_optimizations_g): Merge into
7319 (all_optimizations_g::gate): this.
7320 (gate_rest_of_compilation): Mege into
7321 (rest_of_compilation::gate): this.
7322 (gate_postreload): Merge into
7323 (postreload::gate): this.
7324 (dump_one_pass): Pass cfun to the pass's gate method.
7325 (execute_ipa_summary_passes): Likewise.
7326 (execute_one_pass): Likewise.
7327 (ipa_write_summaries_2): Likewise.
7328 (ipa_write_optimization_summaries_1): Likewise.
7329 (ipa_read_summaries_1): Likewise.
7330 (ipa_read_optimization_summaries_1): Likewise.
7331 (execute_ipa_stmt_fixups): Likewise.
7332 * tree-pass.h (opt_pass::gate): Add function * argument.
7333 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
7334 combine-stack-adj.c, combine.c, compare-elim.c,
7335 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
7336 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
7337 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
7338 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
7339 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
7340 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
7341 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
7342 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
7343 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
7344 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
7345 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
7346 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
7347 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
7348 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
7349 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
7350 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7351 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
7352 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
7353 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
7354 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7355 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7356 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
7357 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
7358 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
7359 var-tracking.c, vtable-verify.c, web.c: Adjust.
7361 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7363 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
7364 * configure: Regenerate.
7366 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7368 * passes.c (dump_one_pass): don't check pass->has_gate.
7369 (execute_ipa_summary_passes): Likewise.
7370 (execute_one_pass): Likewise.
7371 (ipa_write_summaries_2): Likewise.
7372 (ipa_write_optimization_summaries_1): Likewise.
7373 (ipa_read_optimization_summaries_1): Likewise.
7374 (execute_ipa_stmt_fixups): Likewise.
7375 * tree-pass.h (pass_data::has_gate): Remove.
7376 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
7377 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
7378 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
7379 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
7380 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
7381 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
7382 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
7383 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
7384 gimple-low.c, gimple-ssa-isolate-paths.c,
7385 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
7386 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
7387 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
7388 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
7389 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
7390 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
7391 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
7392 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
7393 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
7394 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
7395 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
7396 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
7397 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7398 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
7399 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
7400 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
7401 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7402 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7403 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
7404 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
7405 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
7406 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
7409 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7411 * pass_manager.h (pass_manager::register_dump_files_1): Remove
7413 * passes.c (pass_manager::register_dump_files_1): Merge into
7414 (pass_manager::register_dump_files): this, and remove its handling of
7415 properties since the pass always has the properties anyway.
7416 (pass_manager::pass_manager): Adjust.
7418 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7420 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
7421 * passes.c (pass_manager::register_dump_files_1): Remove dead code
7422 dealing with properties.
7423 (pass_manager::register_dump_files): Adjust.
7425 2014-03-20 Mark Wielaard <mjw@redhat.com>
7427 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
7428 then represent the bound as normal constant value.
7430 2014-04-17 Jakub Jelinek <jakub@redhat.com>
7433 Forward port from 4.8 branch
7434 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
7436 * config/i386/bmiintrin.h (_blsi_u32): New.
7440 (_blsmsk_u32): Ditto.
7441 (_blsmsk_u64): Ditto.
7442 (_tzcnt_u32): Ditto.
7443 (_tzcnt_u64): Ditto.
7445 2014-04-17 Kito Cheng <kito@0xlab.org>
7447 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
7449 2014-04-17 Richard Biener <rguenther@suse.de>
7452 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
7453 boolean results for comparisons.
7455 2014-04-17 Richard Biener <rguenther@suse.de>
7457 PR tree-optimization/60836
7458 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
7459 initial PHI args to be gimple values.
7461 2014-04-17 Richard Biener <rguenther@suse.de>
7463 PR tree-optimization/60841
7464 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
7465 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
7466 of stmts to SLP build.
7467 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
7468 (vect_analyze_slp): Likewise.
7469 (vect_analyze_slp_instance): Likewise.
7470 (vect_build_slp_tree): Limit overall SLP tree growth.
7471 * tree-vectorizer.h (vect_analyze_data_refs,
7472 vect_analyze_slp): Adjust prototypes.
7474 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
7476 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
7479 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
7481 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
7482 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
7483 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
7484 for TARGET_SLOW_PSHUFB
7486 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
7488 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
7489 * config/i386/i386.c (intel_cost): Ditto.
7491 2014-04-17 Joey Ye <joey.ye@arm.com>
7493 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
7495 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
7497 * opts.c (common_handle_option): Disable -fipa-reference coorectly
7500 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
7502 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
7503 (type_all_derivations_known_p): New predicate.
7504 (type_all_ctors_visible_p): New predicate.
7505 (type_possibly_instantiated_p): New predicate.
7506 (get_odr_type): Compute all_derivations_known.
7507 (dump_odr_type): Dump the flag.
7508 (maybe_record_type): Cleanup.
7509 (record_target_from_binfo): Add bases_to_consider array;
7510 record bases for types w/o instances and skip CXX destructor.
7511 (possible_polymorphic_call_targets_1): Add bases_to_consider
7512 and consider_construction parameters; check if type may have instance.
7513 (get_polymorphic_call_info): Set maybe_in_construction to true
7514 when we know nothing.
7515 (record_targets_from_bases): Skip CXX destructors; they are
7516 never called for types in construction.
7517 (possible_polymorphic_call_targets): Do not record target when
7518 type may not have instance.
7520 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
7523 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
7524 external aliases alive, too.
7526 2014-04-16 Andrew Pinski <apinski@cavium.com>
7528 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
7531 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
7533 * final.c (compute_alignments): Do not apply loop alignment to a block
7534 falling through to the exit.
7536 2014-04-16 Catherine Moore <clm@codesourcery.com>
7538 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
7539 Adjust constraints for microMIPS store patterns.
7541 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
7543 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
7545 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
7547 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
7548 (append_use): Run at -O0.
7549 (append_vdef): Likewise.
7550 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
7551 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
7553 2014-04-16 Jakub Jelinek <jakub@redhat.com>
7555 PR tree-optimization/60844
7556 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
7557 (propagate_op_to_single_use, remove_visited_stmt_chain,
7558 linearize_expr, repropagate_negates, reassociate_bb): Use it
7559 instead of gsi_remove.
7561 2014-04-16 Martin Jambor <mjambor@suse.cz>
7563 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
7564 ipa_transforms_to_apply.
7565 (cgraph_function_versioning): Assert that old_node has empty
7566 ipa_transforms_to_apply.
7567 * trans-mem.c (ipa_tm_create_version): Likewise.
7568 * tree-inline.c (tree_function_versioning): Do not duplicate
7569 ipa_transforms_to_apply.
7571 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7574 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
7576 Pass necessary as flags on 64-bit Solaris/x86.
7577 Use lowercase relocs for x86_64-*-*.
7578 * configure: Regenerate.
7580 2014-04-15 Jan Hubicka <jh@suse.cz>
7582 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
7583 (maybe_record_node, likely_target_p): Use it.
7585 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7588 Revert following patch
7590 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7593 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7594 software floating point or no floating point registers, do not
7595 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7596 in GPRs that occurs after we tested for GPRs that would never be
7599 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7600 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7601 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7602 specifically allow DDmode, since that does not use the SPE SIMD
7605 2014-03-21 Mark Wielaard <mjw@redhat.com>
7607 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
7608 as unsigned or int depending on type and value used.
7610 2014-04-15 Richard Biener <rguenther@suse.de>
7612 PR rtl-optimization/56965
7613 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
7614 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
7616 * alias.c (true_dependence_1): Do not call
7617 nonoverlapping_component_refs_p.
7618 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
7619 nonoverlapping_component_refs_p.
7620 (indirect_refs_may_alias_p): Likewise.
7622 2014-04-15 Teresa Johnson <tejohnson@google.com>
7624 * cfg.c (dump_bb_info): Fix flags check.
7625 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
7627 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7629 PR rtl-optimization/60663
7630 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
7633 2014-04-15 Richard Biener <rguenther@suse.de>
7635 * lto-streamer.h (LTO_major_version): Bump to 4.
7637 2014-04-15 Richard Biener <rguenther@suse.de>
7639 * common.opt (lto_partition_model): New enum.
7640 (flto-partition=): Merge separate options with a single with argument,
7641 add -flto-partition=one support.
7642 * flag-types.h (enum lto_partition_model): Declare.
7643 * opts.c (finish_options): Remove duplicate -flto-partition=
7645 * lto-wrapper.c (run_gcc): Adjust.
7647 2014-04-15 Richard Biener <rguenther@suse.de>
7649 * alias.c (ncr_compar): New function.
7650 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
7652 2014-04-15 Richard Biener <rguenther@suse.de>
7654 * alias.c (record_component_aliases): Do not walk BINFOs.
7656 2014-04-15 Richard Biener <rguenther@suse.de>
7658 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7659 Add struct function argument and adjust.
7660 (find_func_aliases_for_call): Likewise.
7661 (find_func_aliases): Likewise.
7662 (find_func_clobbers): Likewise.
7663 (intra_create_variable_infos): Likewise.
7664 (compute_points_to_sets): Likewise.
7665 (ipa_pta_execute): Adjust. Do not push/pop cfun.
7667 2014-04-15 Richard Biener <rguenther@suse.de>
7669 * tree.c (iterative_hash_expr): Use enum tree_code_class
7670 to store TREE_CODE_CLASS.
7671 (tree_block): Likewise.
7672 (tree_set_block): Likewise.
7673 * tree.h (fold_build_pointer_plus_loc): Use
7674 convert_to_ptrofftype_loc.
7676 2014-04-15 Jakub Jelinek <jakub@redhat.com>
7679 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
7682 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
7684 * cfgloop.h (struct loop): Move force_vectorize down.
7685 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
7686 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
7687 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
7688 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
7689 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
7690 * tree-core.h (enum annot_expr_kind): Add new kind values.
7691 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
7692 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
7694 * tree.def (ANNOTATE_EXPR): Tweak comment.
7696 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7698 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
7701 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
7703 * tree.h (TYPE_IDENTIFIER): Declare.
7704 * tree.c (subrange_type_for_debug_p): Use it.
7705 * godump.c (go_format_type): Likewise.
7706 * dwarf2out.c (is_cxx_auto, modified_type_die,
7707 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
7708 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
7710 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7713 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
7715 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
7717 * config/i386/i386.c (examine_argument): Return bool. Return true if
7718 parameter should be passed in memory.
7719 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
7720 (construct_container): Update calls to examine_argument.
7721 (function_arg_advance_64): Ditto.
7722 (return_in_memory_32): Merge with ix86_return_in_memory.
7723 (return_in_memory_64): Ditto.
7724 (return_in_memory_ms_64): Ditto.
7726 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7728 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
7729 * coverage.c (coverage_compute_profile_id): Handle externally visible
7732 2014-04-14 Martin Jambor <mjambor@suse.cz>
7734 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
7735 DECL_DISREGARD_INLINE_LIMITS functions.
7737 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7740 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
7742 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7745 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
7746 optimize_insn_for_speed_p instead of
7747 optimize_function_for_speed_p.
7749 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
7751 * doc/invoke.texi (free): Document AArch64.
7753 2014-04-14 Richard Biener <rguenther@suse.de>
7755 PR tree-optimization/60042
7756 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
7757 (insert_into_preds_of_block): Do not prevent PHI insertion
7758 for REFERENCE exprs here ...
7759 (eliminate_dom_walker::before_dom_children): ... but prevent
7760 their use here under similar conditions when applied to the
7761 IL after PRE optimizations.
7763 2014-04-14 Richard Biener <rguenther@suse.de>
7765 * passes.def: Move early points-to after early SRA.
7767 2014-04-14 Richard Biener <rguenther@suse.de>
7769 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
7770 check for which sign-changes we allow when forwarding
7771 a converted value into a switch.
7773 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7775 * stor-layout.c (place_field): Finalize non-constant offset for the
7778 2014-04-14 Richard Biener <rguenther@suse.de>
7780 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
7782 (expand_switch_using_bit_tests_p): Likewise.
7783 (process_switch): Compute and pass on speed_p based on the
7785 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
7786 optimize_bb_for_speed_p.
7788 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7790 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
7791 * function.h (struct function): Rename has_force_vect_loops into
7792 has_force_vectorize_loops.
7793 * lto-streamer-in.c (input_cfg): Adjust for renaming.
7794 (input_struct_function_base): Likewise.
7795 * lto-streamer-out.c (output_cfg): Likewise.
7796 (output_struct_function_base): Likewise.
7797 * omp-low.c (expand_omp_simd): Likewise.
7798 * tree-cfg.c (move_sese_region_to_fn): Likewise.
7799 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
7800 (version_loop_for_if_conversion): Likewise.
7801 (tree_if_conversion): Likewise.
7802 (main_tree_if_conversion): Likewise.
7803 (gate_tree_if_conversion): Likewise.
7804 * tree-inline.c (copy_loops): Likewise.
7805 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
7806 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
7807 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
7808 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
7809 * tree-vectorizer.c (vectorize_loops): Likewise.
7810 * tree-vectorizer.h (unlimited_cost_model): Likewise.
7812 2014-04-14 Richard Biener <rguenther@suse.de>
7815 * lto-streamer-out.c (wrap_refs): New function.
7816 (lto_output): Wrap symbol references in global initializes in
7817 type-preserving MEM_REFs.
7819 2014-04-14 Christian Bruel <christian.bruel@st.com>
7821 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
7823 2014-04-14 Christian Bruel <christian.bruel@st.com>
7825 * config/sh/sh.md (setmemqi): New expand pattern.
7826 * config/sh/sh.h (CLEAR_RATIO): Define.
7827 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
7828 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
7830 2014-04-14 Richard Biener <rguenther@suse.de>
7833 * fold-const.c (negate_expr_p): Don't negate directional rounding
7835 (fold_negate_expr): Likewise.
7837 2014-04-14 Richard Biener <rguenther@suse.de>
7839 PR tree-optimization/59817
7840 PR tree-optimization/60453
7841 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
7842 recursion to catch all CHRECs in the scalar evolution and restrict
7843 the predicate for the remains appropriately.
7845 2014-04-12 Catherine Moore <clm@codesourcery.com>
7847 * config/mips/constraints.md: Add new register constraint "kb".
7848 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
7849 (*movhi_internal): Likewise.
7850 (*movqi_internal): Likewise.
7851 * config/mips/mips.h (M16_STORE_REGS): New register class.
7852 (REG_CLASS_NAMES): Add M16_STORE_REGS.
7853 (REG_CLASS_CONTENTS): Likewise.
7854 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
7856 2014-04-11 Tobias Burnus <burnus@net-b.de>
7859 * doc/invoke.texi (-Wformat-signedness): Document it.
7860 (Wformat=2): Mention that this enables -Wformat-signedness.
7862 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7864 * common/config/epiphany/epiphany-common.c
7865 (epiphany_option_optimization_table): Enable section anchors by
7866 default at -O1 or higher.
7867 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
7868 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
7869 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
7870 carries no extra cost.
7871 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
7872 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
7873 * config/epiphany/predicates.md (memclob_operand): New predicate.
7874 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
7875 Use memclob_operand predicate and X constraint for operand 3.
7877 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7879 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
7880 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
7883 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7885 PR rtl-optimization/60651
7886 * mode-switching.c (optimize_mode_switching): Make sure to emit
7887 sets of a lower numbered entity before sets of a higher numbered
7888 entity to a mode of the same or lower priority.
7889 When creating a seginfo for a basic block that starts with a code
7890 label, move the insertion point past the code label.
7891 (new_seginfo): Document and enforce requirement that
7892 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
7893 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
7894 * doc/tm.texi: Regenerate.
7896 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
7899 * config/arc/arc.c (arc_save_restore): Fix assert typo.
7901 2013-04-11 Jakub Jelinek <jakub@redhat.com>
7903 * BASE-VER: Set to 4.10.0.
7905 2014-04-11 Tobias Burnus <burnus@net-b.de>
7908 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
7909 * doc/gcc.texi (Service): Update description in the @menu
7910 * doc/invoke.texi (Option Summary): Remove misplaced and
7913 2014-04-11 Steve Ellcey <sellcey@mips.com>
7914 Jakub Jelinek <jakub@redhat.com>
7917 * expr.c (convert_move): Use emit_store_flag_force instead of
7918 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
7921 2014-04-11 Richard Biener <rguenther@suse.de>
7924 * varasm.c (assemble_alias): Avoid endless error reporting
7925 recursion by setting TREE_ASM_WRITTEN.
7927 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7929 * config/s390/s390.md: Add a splitter for NOT rtx.
7931 2014-04-11 Jakub Jelinek <jakub@redhat.com>
7933 PR rtl-optimization/60663
7934 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
7936 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
7937 Jakub Jelinek <jakub@redhat.com>
7940 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
7941 flag from decl_node to node.
7943 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7946 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
7947 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
7948 ameliorating the cases where it can be.
7950 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
7953 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7955 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7956 (loadsync_<mode>): Change mode.
7957 (load_quadpti, store_quadpti): New.
7958 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7959 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7960 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
7962 2014-04-09 Cong Hou <congh@google.com>
7965 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
7968 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7970 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
7971 instead of vnor to exploit possible fusion opportunity in the
7973 (altivec_expand_vec_perm_const_le): Likewise.
7975 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7977 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7978 (loadsync_<mode>): Change mode.
7979 (load_quadpti, store_quadpti): New.
7980 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7981 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7983 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
7986 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
7987 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
7988 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
7990 2014-04-08 Richard Biener <rguenther@suse.de>
7993 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
7994 a 64bit widest int print double-int similar to on HWI64 hosts.
7996 2014-04-08 Richard Biener <rguenther@suse.de>
7998 PR tree-optimization/60785
7999 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
8000 default defs properly.
8002 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
8004 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
8005 (Weffc++): Likewise.
8007 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
8009 * ipa-devirt.c (maybe_record_node): When node is not recorded,
8010 set completep to false rather than true.
8012 2014-04-07 Douglas B Rupp <rupp@adacore.com>
8015 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
8016 ARM_TARGET2_DWARF_FORMAT.
8018 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
8021 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
8022 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
8025 2014-04-07 Richard Biener <rguenther@suse.de>
8027 PR tree-optimization/60766
8028 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
8029 (may_eliminate_iv): Convert cand_value_at result to desired type.
8031 2014-04-07 Jason Merrill <jason@redhat.com>
8034 * common.opt (-fno-gnu-unique): Add.
8035 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
8037 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8039 * haifa-sched.c: Fix outdated function reference and minor
8040 grammar errors in introductory comment.
8042 2014-04-07 Richard Biener <rguenther@suse.de>
8045 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
8047 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
8049 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
8052 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
8053 size accounting for thunks.
8054 (pa_asm_output_mi_thunk): Use final_start_function() and
8055 final_end_function() to output function start and end directives.
8057 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
8059 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
8060 device specific ISA/ feature information. Remove short_sp and
8061 errata_skip ds. Add avr_device_specific_features enum to have device
8063 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
8064 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
8065 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
8066 updated device specific info.
8067 * config/avr/avr-mcus.def: Merge device specific details to
8068 dev_attribute field.
8069 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
8071 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
8072 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
8073 assembler if RMW isa supported by current device.
8074 * config/avr/genmultilib.awk: Update as device info structure changed.
8075 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
8077 2014-04-04 Cong Hou <congh@google.com>
8079 PR tree-optimization/60656
8080 * tree-vect-stmts.c (supportable_widening_operation):
8081 Fix a bug that elements in a vector with vect_used_by_reduction
8082 property are incorrectly reordered when the operation on it is not
8083 consistant with the one in reduction operation.
8085 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
8087 PR rtl-optimization/60155
8088 * gcse.c (record_set_data): New function.
8089 (single_set_gcse): New function.
8090 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
8091 (hoist_code): Likewise.
8092 (get_pressure_class_and_nregs): Likewise.
8094 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
8096 * explow.c (probe_stack_range): Emit a final optimization blockage.
8098 2014-04-04 Anthony Green <green@moxielogic.com>
8100 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
8103 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
8106 * lto-cgraph.c (input_overwrite_node): Check that partitioning
8107 flags are set only during streaming.
8108 * ipa.c (process_references, walk_polymorphic_call_targets,
8109 symtab_remove_unreachable_nodes): Drop bodies of always inline
8110 after early inlining.
8111 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
8113 2014-04-04 Jakub Jelinek <jakub@redhat.com>
8114 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8117 * dwarf2out.c (const_ok_for_output_1): Reject expressions
8120 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8123 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
8125 (cortex_a53_fdivd): Likewise.
8127 2014-04-04 Martin Jambor <mjambor@suse.cz>
8130 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
8132 * cgraph.c (clone_of_p): Also return true if thunks match.
8133 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
8134 cgraph_function_or_thunk_node and an obsolete comment.
8135 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
8137 (build_function_decl_skip_args): Likewise.
8138 (set_new_clone_decl_and_node_flags): New function.
8139 (duplicate_thunk_for_node): Likewise.
8140 (redirect_edge_duplicating_thunks): Likewise.
8141 (cgraph_clone_node): New parameter args_to_skip, pass it to
8142 redirect_edge_duplicating_thunks which is called instead of
8143 cgraph_redirect_edge_callee.
8144 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
8145 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
8147 2014-04-04 Jeff Law <law@redhat.com>
8150 * config/arm/predicates.md (const_int_I_operand): New predicate.
8151 (const_int_M_operand): Similarly.
8152 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
8154 (insv_t2, extv_reg, extzv_t2): Likewise.
8155 (load_multiple_with_writeback): Similarly for const_int_I_operand.
8156 (pop_multiple_with_writeback_and_return): Likewise.
8157 (vfp_pop_multiple_with_writeback): Likewise
8159 2014-04-04 Richard Biener <rguenther@suse.de>
8162 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
8163 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
8165 * gimplify.h (gimple_add_tmp_var_fn): Declare.
8166 * gimplify.c (gimple_add_tmp_var_fn): New function.
8167 * gimple-expr.h (create_tmp_reg_fn): Declare.
8168 * gimple-expr.c (create_tmp_reg_fn): New function.
8169 * gimple-low.c (record_vars_into): Don't change cfun.
8170 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
8171 code generation without cfun.
8173 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
8176 * Makefile.in (install-driver): Fix shell scripting.
8178 2014-04-03 Cong Hou <congh@google.com>
8180 PR tree-optimization/60505
8181 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
8182 threshold of number of iterations below which no vectorization
8184 * tree-vect-loop.c (new_loop_vec_info):
8185 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
8186 * tree-vect-loop.c (vect_analyze_loop_operations):
8187 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
8188 * tree-vect-loop.c (vect_transform_loop):
8189 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
8190 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
8191 of iterations of the loop and see if we should build the epilogue.
8193 2014-04-03 Richard Biener <rguenther@suse.de>
8195 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
8196 (streamer_tree_cache_create): Adjust.
8197 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
8198 to allow optional nodes array.
8199 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
8200 (streamer_tree_cache_append): Likewise.
8201 (streamer_tree_cache_create): Create nodes array optionally
8202 as specified by parameter.
8203 * lto-streamer-out.c (create_output_block): Avoid maintaining
8204 the node array in the writer cache.
8205 (DFS_write_tree): Remove assertion.
8206 (produce_asm_for_decls): Free the out decl state hash table early.
8207 * lto-streamer-in.c (lto_data_in_create): Adjust for
8208 streamer_tree_cache_create prototype change.
8210 2014-04-03 Richard Biener <rguenther@suse.de>
8212 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
8213 set TREE_CHAIN to NULL_TREE.
8215 2014-04-03 Richard Biener <rguenther@suse.de>
8217 PR tree-optimization/60740
8218 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
8219 over all GIMPLE_COND operands.
8221 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
8223 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
8224 (Weffc++): Remove Scott's numbering, merge lists and reference
8227 2014-04-03 Nick Clifton <nickc@redhat.com>
8229 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
8232 2014-04-03 Martin Jambor <mjambor@suse.cz>
8234 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
8235 mention gcc_unreachable before failing.
8236 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
8239 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
8242 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
8243 inconsistent code and instead mark the context inconsistent.
8244 (possible_polymorphic_call_targets): For inconsistent contexts
8245 return empty complete list.
8247 2014-04-02 Anthony Green <green@moxielogic.com>
8249 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
8250 (extendqisi2, extendhisi2): Define.
8251 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
8252 (WCHAR_TYPE): Change to unsigned int.
8254 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8256 PR tree-optimization/60733
8257 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
8258 insertion point for PHI candidates to be the end of the feeding
8259 block for the PHI argument.
8261 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
8263 PR rtl-optimization/60650
8264 * lra-constraints.c (process_alt_operands): Decrease reject for
8265 earlyclobber matching.
8267 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8269 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
8271 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8273 * config/spu/spu.c (pad_bb): Do not crash when the last
8274 insn is CODE_FOR_blockage.
8276 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8278 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
8279 lies outside the target mode.
8281 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
8284 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
8285 software floating point or no floating point registers, do not
8286 allow any type in the FPRs. Eliminate a test for SPE SIMD types
8287 in GPRs that occurs after we tested for GPRs that would never be
8290 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
8291 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
8292 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
8293 specifically allow DDmode, since that does not use the SPE SIMD
8296 2014-04-02 Richard Biener <rguenther@suse.de>
8299 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
8300 MODE_INTs. Properly use negv_optab.
8301 (expand_abs): Likewise.
8303 2014-04-02 Richard Biener <rguenther@suse.de>
8306 * Makefile.in (install-driver): Guard extra installs with special
8309 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
8311 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8314 2014-04-01 Richard Henderson <rth@redhat.com>
8317 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
8318 alternative enabled before register allocation.
8320 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
8322 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
8323 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
8325 (nios2_large_got_address): Remove unneeded 'sym' parameter.
8326 (nios2_got_address): Update nios2_large_got_address call site.
8327 (nios2_delegitimize_address): New function.
8328 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
8329 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
8330 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
8332 2014-04-01 Martin Husemann <martin@duskware.de>
8334 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
8337 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
8339 PR rtl-optimization/60604
8340 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
8341 check from register_operand.
8342 (register_operand): Redefine in terms of general_operand.
8343 (nonmemory_operand): Use register_operand for the non-constant cases.
8345 2014-04-01 Richard Biener <rguenther@suse.de>
8347 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
8349 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
8351 * doc/invoke.texi (mapp-regs): Clarify.
8353 2014-03-31 Ulrich Drepper <drepper@gmail.com>
8355 * config/i386/avx512fintrin.h (__v32hi): Define type.
8356 (__v64qi): Likewise.
8357 (_mm512_set1_epi8): Define.
8358 (_mm512_set1_epi16): Define.
8359 (_mm512_set4_epi32): Define.
8360 (_mm512_set4_epi64): Define.
8361 (_mm512_set4_pd): Define.
8362 (_mm512_set4_ps): Define.
8363 (_mm512_setr4_epi64): Define.
8364 (_mm512_setr4_epi32): Define.
8365 (_mm512_setr4_pd): Define.
8366 (_mm512_setr4_ps): Define.
8367 (_mm512_setzero_epi32): Define.
8369 2014-03-31 Martin Jambor <mjambor@suse.cz>
8372 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
8373 callsite_arguments_match_p. Updated all callers. Also check types of
8374 corresponding formal parameters and actual arguments.
8375 (not_all_callers_have_enough_arguments_p) Renamed to
8376 some_callers_have_mismatched_arguments_p.
8378 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
8380 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
8382 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
8385 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
8388 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
8390 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
8391 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
8393 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
8394 Use FMAMODE_NOVF512 mode iterator.
8395 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
8396 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
8397 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
8398 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
8400 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
8401 Use VF_128_256 mode iterator.
8402 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
8405 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
8407 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
8408 static chain if needed.
8410 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
8413 * lra-constraints.c (index_part_to_reg): New.
8414 (process_address): Use it.
8416 2014-03-27 Jeff Law <law@redhat.com>
8417 Jakub Jelinek <jakub@redhat.com>
8420 * expr.c (do_tablejump): Use simplify_gen_binary rather than
8421 gen_rtx_{PLUS,MULT} to build up the address expression.
8423 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
8424 creating non-canonical RTL.
8426 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
8429 * ipa-inline.c (want_inline_small_function_p): Short circuit large
8430 functions; reorganize to make cheap checks first.
8431 (inline_small_functions): Do not estimate growth when dumping;
8433 * ipa-inline.h (inline_summary): Add min_size.
8434 (growth_likely_positive): New function.
8435 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
8436 (set_cond_stmt_execution_predicate): Cleanup.
8437 (estimate_edge_size_and_time): Compute min_size.
8438 (estimate_calls_size_and_time): Likewise.
8439 (estimate_node_size_and_time): Likewise.
8440 (inline_update_overall_summary): Update min_size.
8441 (do_estimate_edge_time): Likewise.
8442 (do_estimate_edge_size): Update.
8443 (do_estimate_edge_hints): Update.
8444 (growth_likely_positive): New function.
8446 2014-03-28 Jakub Jelinek <jakub@redhat.com>
8449 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
8450 also if addr has VOIDmode.
8452 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8454 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
8455 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
8457 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
8458 instructions as well as AdvancedSIMD loads.
8460 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8462 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
8463 Use crypto_aese type.
8464 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
8465 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
8466 crypto_aese, crypto_aesmc. Move to types.md.
8467 * config/arm/types.md (crypto_aes): Split into crypto_aese,
8469 * config/arm/iterators.md (crypto_type): Likewise.
8471 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
8473 * cgraph.c: Include expr.h and tree-dfa.h.
8474 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
8477 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
8480 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
8481 regs from checking multi-reg pseudos.
8483 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8485 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
8487 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8489 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
8490 if it would clobber the stack pointer, even temporarily.
8492 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
8494 * mode-switching.c: Make small adjustments to the top comment.
8496 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
8498 * config/rs6000/constraints.md (wD constraint): New constraint to
8499 match the constant integer to get the top DImode/DFmode out of a
8500 vector in a VSX register.
8502 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
8503 match the constant integer to get the top DImode/DFmode out of a
8504 vector in a VSX register.
8506 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
8509 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8512 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
8513 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
8515 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
8516 Optimize vec_extract of 64-bit values, where the value being
8517 extracted is in the top word, where we can use scalar
8518 instructions. Add direct move and store support. Combine the big
8519 endian/little endian vector select load support into a single insn.
8520 (vsx_extract_<mode>_internal1): Likewise.
8521 (vsx_extract_<mode>_internal2): Likewise.
8522 (vsx_extract_<mode>_load): Likewise.
8523 (vsx_extract_<mode>_store): Likewise.
8524 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
8525 combined into vsx_extract_<mode>_load.
8526 (vsx_extract_<mode>_one_le): Likewise.
8528 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
8529 define the top 64-bit vector element.
8531 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
8534 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8535 Document vec_vbpermq builtin.
8538 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
8539 enable use of xxsldwi and xxpermdi builtin functions.
8540 (vec_xxpermdi): Likewise.
8542 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8543 Document use of vec_xxsldwi and vec_xxpermdi builtins.
8545 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
8547 PR rtl-optimization/60650
8548 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
8550 (find_spills_for): New.
8551 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
8552 Spill all pseudos on the second iteration.
8554 2014-03-27 Marek Polacek <polacek@redhat.com>
8557 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
8560 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8562 * config/s390/s390.c (s390_can_use_return_insn): Check for
8563 call-saved FPRs on 31 bit.
8565 2014-03-27 Jakub Jelinek <jakub@redhat.com>
8568 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
8569 if they need regimplification, just drop them instead of
8570 calling gimple_regimplify_operands on them.
8572 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
8575 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
8576 (aarch64_frame_pointer_required): Adjust logic.
8577 (aarch64_can_eliminate): Adjust logic.
8578 (aarch64_override_options_after_change): Adjust logic.
8580 2014-03-27 Dehao Chen <dehao@google.com>
8582 * ipa-inline.c (early_inliner): Update node's inline info.
8584 2014-03-26 Dehao Chen <dehao@google.com>
8586 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
8587 compiler inserted conditional jumps for NAN float check.
8589 2014-03-26 Jakub Jelinek <jakub@redhat.com>
8591 * ubsan.h (ubsan_create_data): Change second argument's type
8592 to const location_t *.
8593 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
8595 (ubsan_create_data): Change second argument to const location_t *PLOC.
8596 Create Loc field whenever PLOC is non-NULL.
8597 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
8598 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
8602 * real.c (real_to_integer2): Change type of low to UHWI.
8604 2014-03-26 Tobias Burnus <burnus@net-b.de>
8606 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
8607 (CILK_SELF_SPECS): New define.
8608 (driver_self_specs): Use it.
8610 2014-03-26 Richard Biener <rguenther@suse.de>
8612 * tree-pretty-print.c (percent_K_format): Implement special
8613 case for LTO and its stripped down BLOCK tree.
8615 2014-03-26 Jakub Jelinek <jakub@redhat.com>
8618 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
8620 * tree-vrp.c (simplify_internal_call_using_ranges): If only
8621 one range is range_int_cst_p, but not both, at least optimize
8622 addition/subtraction of 0 and multiplication by 0 or 1.
8623 * gimple-fold.c (gimple_fold_call): Fold
8624 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
8625 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
8626 INTEGER_CSTs, try to fold at least x * 0 and y - y.
8628 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
8630 PR rtl-optimization/60452
8631 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
8632 <case REG>: Return 1 for invalid offsets from the frame pointer.
8634 2014-03-26 Marek Polacek <polacek@redhat.com>
8637 * doc/extend.texi (C Extensions): Mention variable-length arrays in
8640 2014-03-26 Marek Polacek <polacek@redhat.com>
8643 * doc/extend.texi (Designated Inits): Describe what happens to omitted
8646 2014-03-26 Marek Polacek <polacek@redhat.com>
8649 * ira-color.c (update_conflict_hard_regno_costs): Perform the
8650 multiplication in unsigned type.
8652 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
8654 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
8656 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
8658 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
8660 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
8663 * cif-code.def (UNREACHABLE) New code.
8664 * ipa-inline.c (inline_small_functions): Skip edges to
8665 __builtlin_unreachable.
8666 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
8667 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
8668 predicate to __bulitin_unreachable.
8669 (set_cond_stmt_execution_predicate): Fix issue when
8670 invert_tree_comparison returns ERROR_MARK.
8671 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
8672 propagate to inline clones.
8673 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
8675 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
8676 * cgraphclones.c (cgraph_clone_node): If call destination is already
8677 ureachable, do not redirect it back.
8678 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
8681 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
8683 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
8684 Do not modify inline clones.
8686 2014-03-25 Jakub Jelinek <jakub@redhat.com>
8688 * config/i386/i386.md (general_sext_operand): New mode attr.
8689 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
8690 don't generate (sign_extend (const_int)).
8691 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
8692 operands[2]. Use We constraint instead of <i> and
8693 <general_sext_operand> predicate instead of <general_operand>.
8694 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
8695 * config/i386/constraints.md (We): New constraint.
8696 * config/i386/predicates.md (x86_64_sext_operand,
8697 sext_operand): New predicates.
8699 2014-03-25 Martin Jambor <mjambor@suse.cz>
8702 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
8703 inconsistent devirtualizations to __builtin_unreachable.
8705 2014-03-25 Marek Polacek <polacek@redhat.com>
8708 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
8710 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
8712 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
8713 order of elements for big-endian.
8715 2014-03-25 Richard Biener <rguenther@suse.de>
8718 * gimplify-me.c (gimple_regimplify_operands): Update the
8721 2014-03-25 Martin Jambor <mjambor@suse.cz>
8724 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
8725 (lto_output_varpool_node): Likewise.
8726 (input_overwrite_node): Likewise.
8727 (input_varpool_node): Likewise.
8729 2014-03-25 Richard Biener <rguenther@suse.de>
8731 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
8732 (run_gcc): Likewise.
8734 2014-03-25 Jakub Jelinek <jakub@redhat.com>
8736 * combine.c (simplify_compare_const): Add MODE argument.
8737 Handle mode_width 0 as very large mode_width.
8738 (try_combine, simplify_comparison): Adjust callers.
8740 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
8741 type to avoid signed integer overflow.
8742 * explow.c (plus_constant): Likewise.
8744 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
8746 * doc/generic.texi: Correct typos.
8748 2014-03-24 Tobias Burnus <burnus@net-b.de>
8750 * doc/invoke.texi (-flto): Expand section about
8751 using static libraries with LTO.
8753 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8755 PR rtl-optimization/60501
8756 * optabs.def (addptr3_optab): New optab.
8757 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
8758 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
8759 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
8761 * lra.c (emit_add3_insn): Use the addptr pattern if available.
8763 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
8765 2014-03-24 Ulrich Drepper <drepper@gmail.com>
8767 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
8770 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
8771 (_mm256_undefined_ps): Define.
8772 (_mm256_undefined_pd): Define.
8773 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
8774 (_mm_undefined_pd): Define.
8775 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
8776 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
8777 (_mm512_undefined_ps): Define.
8778 (_mm512_undefined_pd): Define.
8779 Use _mm*_undefined_*.
8780 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
8782 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
8784 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
8785 (lshr_simd): DI mode added.
8786 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
8787 (aarch64_ushr_simddi): Likewise.
8788 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
8789 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
8790 (vshrd_n_u64): Likewise.
8792 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8794 * Makefile.in (s-macro_list): Depend on cc1.
8796 2014-03-23 Teresa Johnson <tejohnson@google.com>
8798 * ipa-utils.c (ipa_print_order): Use specified dump file.
8800 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
8802 PR rtl-optimization/60601
8803 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
8805 * gcc.c (eval_spec_function): Initialize save_growing_value.
8807 2014-03-22 Jakub Jelinek <jakub@redhat.com>
8810 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
8811 code == MINUS_EXPR, never swap op0 with op1.
8813 * toplev.c (init_local_tick): Avoid signed integer multiplication
8815 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
8816 shift by first operand's bitsize.
8818 2014-03-21 Jakub Jelinek <jakub@redhat.com>
8821 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
8823 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
8824 TARGET_ISA_64BIT_P(x).
8826 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8828 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
8829 pattern for vector nor instead of subtract from splat(-1).
8830 (altivec_expand_vec_perm_const_le): Likewise.
8832 2014-03-21 Richard Henderson <rth@twiddle.net>
8835 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
8836 related insns after epilogue_completed.
8838 2014-03-21 Martin Jambor <mjambor@suse.cz>
8841 * cgraph.h (symtab_node): New flag body_removed.
8842 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
8843 when removing bodies.
8844 * symtab.c (dump_symtab_base): Dump body_removed flag.
8845 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
8846 had their bodies removed.
8848 2014-03-21 Martin Jambor <mjambor@suse.cz>
8851 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
8854 2014-03-21 Richard Biener <rguenther@suse.de>
8856 PR tree-optimization/60577
8857 * tree-core.h (struct tree_base): Document nothrow_flag use
8858 in VAR_DECL_NONALIASED.
8859 * tree.h (VAR_DECL_NONALIASED): New.
8860 (may_be_aliased): Adjust.
8861 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
8863 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8865 * expr.c (expand_expr_real_1): Remove outdated comment.
8867 2014-03-20 Jakub Jelinek <jakub@redhat.com>
8870 * ira.c (adjust_cleared_regs): Call copy_rtx on
8871 *reg_equiv[REGNO (loc)].src_p before passing it to
8872 simplify_replace_fn_rtx.
8875 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
8876 into CONST, put pic register as first operand of PLUS. Use
8877 gen_const_mem for both 32-bit and 64-bit PIC got loads.
8879 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8881 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
8883 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8885 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
8886 around for store forwarding issue in the FPU on the UT699.
8887 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
8888 loads and operations if -mfix-ut699 is specified.
8889 (divtf3_hq): Tweak attribute.
8890 (sqrttf2_hq): Likewise.
8892 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8894 * calls.c (store_one_arg): Remove incorrect const qualification on the
8895 type of the temporary.
8896 * cfgexpand.c (expand_return): Likewise.
8897 * expr.c (expand_constructor): Likewise.
8898 (expand_expr_real_1): Likewise.
8900 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8902 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
8905 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
8908 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
8910 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
8912 * config/arm/aarch-common-protos.h
8913 (alu_cost_table): Fix spelling of "extend".
8914 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
8916 2014-03-19 Richard Biener <rguenther@suse.de>
8919 * tree-core.h (tree_type_common): Re-order pointer members
8920 to reduce recursion depth during GC walks.
8922 2014-03-19 Marek Polacek <polacek@redhat.com>
8925 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
8926 before accessing it.
8928 2014-03-19 Richard Biener <rguenther@suse.de>
8931 * lto-streamer-in.c (input_function): In WPA stage do not drop
8934 2014-03-19 Jakub Jelinek <jakub@redhat.com>
8936 PR tree-optimization/60559
8937 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
8938 with build_zero_cst assignment.
8940 2014-03-18 Kai Tietz <ktietz@redhat.com>
8942 PR rtl-optimization/56356
8943 * sdbout.c (sdbout_parms): Verify that parms'
8944 incoming argument is valid.
8945 (sdbout_reg_parms): Likewise.
8947 2014-03-18 Richard Henderson <rth@redhat.com>
8950 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
8951 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
8952 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
8954 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
8956 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
8957 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
8958 Italicize plugin event names in description. Explain that
8959 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
8960 Remind that no GCC functions should be called after PLUGIN_FINISH.
8961 Explain what pragmas with expansion are.
8963 2014-03-18 Martin Liska <mliska@suse.cz>
8965 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
8966 gimple call statement is update.
8967 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
8968 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
8970 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8973 * ubsan.c (ubsan_instrument_unreachable): Call
8974 initialize_sanitizer_builtins.
8975 (ubsan_pass): Likewise.
8978 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
8979 varpool_finalize_decl instead of rest_of_decl_compilation.
8981 2014-03-18 Richard Biener <rguenther@suse.de>
8983 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
8984 by using bitmap_and_compl instead of bitmap_and_compl_into.
8985 (df_rd_transfer_function): Likewise.
8987 2014-03-18 Richard Biener <rguenther@suse.de>
8989 * doc/lto.texi (fresolution): Fix typo.
8991 2014-03-18 Richard Biener <rguenther@suse.de>
8993 * doc/invoke.texi (flto): Update for changes in 4.9.
8995 2014-03-18 Richard Biener <rguenther@suse.de>
8997 * doc/loop.texi: Remove section on the removed lambda framework.
8998 Update loop docs with recent changes in preserving loop structure.
9000 2014-03-18 Richard Biener <rguenther@suse.de>
9002 * doc/lto.texi (-fresolution): Document.
9004 2014-03-18 Richard Biener <rguenther@suse.de>
9006 * doc/contrib.texi: Adjust my name.
9008 2014-03-18 Jakub Jelinek <jakub@redhat.com>
9011 * internal-fn.c: Include diagnostic-core.h.
9012 (expand_BUILTIN_EXPECT): New function.
9013 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
9014 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
9015 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
9016 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
9018 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
9019 Revert 3 argument __builtin_expect code.
9020 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
9021 * gimple-fold.c (gimple_fold_call): Likewise.
9022 * tree.h (fold_builtin_expect): New prototype.
9023 * builtins.c (build_builtin_expect_predicate): Add predictor
9024 argument, if non-NULL, create 3 argument __builtin_expect.
9025 (fold_builtin_expect): No longer static. Add ARG2 argument,
9026 pass it through to build_builtin_expect_predicate.
9027 (fold_builtin_2): Adjust caller.
9028 (fold_builtin_3): Handle BUILT_IN_EXPECT.
9029 * internal-fn.def (BUILTIN_EXPECT): New.
9031 2014-03-18 Tobias Burnus <burnus@net-b.de>
9034 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
9035 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
9036 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
9038 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
9041 * predict.c (combine_predictions_for_bb): Fix up formatting.
9042 (expr_expected_value_1, expr_expected_value): Add predictor argument,
9043 fill what it points to if non-NULL.
9044 (tree_predict_by_opcode): Adjust caller, use the predictor.
9045 * predict.def (PRED_COMPARE_AND_SWAP): Add.
9047 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
9049 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
9050 proper constant for the store mode.
9052 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
9054 * symtab.c (change_decl_assembler_name): Fix transparent alias
9057 2014-03-16 Renlin Li <Renlin.Li@arm.com>
9059 * config/aarch64/aarch64.c: Correct the comments about the
9060 aarch64 stack layout.
9062 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
9064 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
9065 check for GF_OMP_FOR_KIND_FOR.
9067 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
9069 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
9070 ymm and zmm register names.
9072 2014-03-17 Jakub Jelinek <jakub@redhat.com>
9075 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
9076 note creation for the 2010-08-31 changes.
9078 2014-03-17 Marek Polacek <polacek@redhat.com>
9081 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
9082 as -fno-tree-loop-vectorize.
9083 (expand_omp_simd): Likewise.
9085 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
9087 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
9088 (eligible_for_call_delay): New prototype.
9089 * config/sparc/sparc.c (tls_call_delay): Rename into...
9090 (eligible_for_call_delay): ...this. Return false if the instruction
9091 cannot be put in the delay slot of a branch.
9092 (eligible_for_restore_insn): Simplify.
9093 (eligible_for_return_delay): Return false if the instruction cannot be
9094 put in the delay slot of a branch and simplify.
9095 (eligible_for_sibcall_delay): Return false if the instruction cannot be
9096 put in the delay slot of a branch.
9097 * config/sparc/sparc.md (fix_ut699): New attribute.
9098 (tls_call_delay): Delete.
9099 (in_call_delay): Reimplement.
9100 (eligible_for_sibcall_delay): Rename into...
9101 (in_sibcall_delay): ...this.
9102 (eligible_for_return_delay): Rename into...
9103 (in_return_delay): ...this.
9104 (in_branch_delay): Reimplement.
9105 (in_uncond_branch_delay): Delete.
9106 (in_annul_branch_delay): Delete.
9108 2014-03-14 Richard Henderson <rth@redhat.com>
9111 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
9112 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
9113 (*floathi<X87MODEF>2_i387_with_temp): Remove.
9114 (floathi splitters): Remove.
9115 (float<SWI48x>xf2): New pattern.
9116 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
9117 code that tried to handle DImode for 32-bit, but which was excluded
9118 by the pattern's condition. Drop allocation of stack temporary.
9119 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
9120 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
9121 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
9122 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
9123 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
9124 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
9125 (*float<SWI48><MODEF>2_sse_interunit): Remove.
9126 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
9127 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
9128 (*float<SWI48x><X87MODEF>2_i387): Remove.
9129 (all float _with_temp splitters): Remove.
9130 (*float<SWI48x><MODEF>2_i387): New pattern.
9131 (*float<SWI48><MODEF>2_sse): New pattern.
9132 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
9133 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
9135 2014-03-14 Jakub Jelinek <jakub@redhat.com>
9136 Marek Polacek <polacek@redhat.com>
9139 * common.opt (dump_base_name_prefixed): New Variable.
9140 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
9141 if x_dump_base_name_prefixed is already set, set it at the end.
9143 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
9145 PR rtl-optimization/60508
9146 * lra-constraints.c (get_reload_reg): Add new parameter
9148 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
9149 Pass the new parameter values.
9151 2014-03-14 Richard Biener <rguenther@suse.de>
9153 * common.opt: Revert unintented changes from r205065.
9156 2014-03-14 Richard Biener <rguenther@suse.de>
9159 * cfghooks.c (split_block): Properly adjust all loops the
9160 block was a latch of.
9162 2014-03-14 Martin Jambor <mjambor@suse.cz>
9165 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
9168 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
9171 * config/avr/avr.c (avr_set_current_function): Pass function name
9172 through default_strip_name_encoding before sanity checking instead
9173 of skipping the first char of the assembler name.
9175 2014-03-13 Richard Henderson <rth@redhat.com>
9178 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
9179 (ix86_force_to_memory, ix86_free_from_memory): Remove.
9180 * config/i386/i386-protos.h: Likewise.
9181 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
9182 in the expander instead of a splitter.
9183 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
9184 any possibility of requiring a memory.
9185 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
9186 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
9187 (fp branch splitters): Update for ix86_split_fp_branch.
9188 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
9189 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
9190 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
9191 (*fop_<MODEF>_2_i387): Remove f/r alternative.
9192 (*fop_<MODEF>_3_i387): Likewise.
9193 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
9194 (splitters for the fop_* register patterns): Remove.
9195 (fscalexf4_i387): Rename from *fscalexf4_i387.
9196 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
9198 2014-03-13 Jakub Jelinek <jakub@redhat.com>
9200 PR tree-optimization/59779
9201 * tree-dfa.c (get_ref_base_and_extent): Use double_int
9202 type for bitsize and maxsize instead of HOST_WIDE_INT.
9204 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
9206 PR rtl-optimization/57320
9207 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
9208 the CFG after thread_prologue_and_epilogue_insns.
9210 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
9212 PR rtl-optimization/57189
9213 * lra-constraints.c (process_alt_operands): Disfavor spilling
9216 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
9218 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
9220 2014-03-13 Jakub Jelinek <jakub@redhat.com>
9222 PR tree-optimization/59025
9224 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
9225 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
9227 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
9230 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
9231 calls of avr_out_plus_1.
9233 2014-03-13 Bin Cheng <bin.cheng@arm.com>
9235 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
9236 BB's single pred and update the father loop's latch info later.
9238 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
9240 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
9244 (VEC_base): Likewise.
9245 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
9246 registers, we need to swap double words in little endian mode.
9248 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
9249 to be a container mode for 128-bit integer operations added in ISA
9250 2.07. Unlike TImode and PTImode, the preferred register set is
9251 the Altivec/VMX registers for the 128-bit operations.
9253 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
9255 (rs6000_split_128bit_ok_p): Likewise.
9257 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
9258 macros for creating ISA 2.07 normal and overloaded builtin
9259 functions with 3 arguments.
9260 (BU_P8V_OVERLOAD_3): Likewise.
9261 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
9262 for use as overloaded functions.
9263 (VPERM_1TI_UNS): Likewise.
9264 (VSEL_1TI): Likewise.
9265 (VSEL_1TI_UNS): Likewise.
9266 (ST_INTERNAL_1ti): Likewise.
9267 (LD_INTERNAL_1ti): Likewise.
9268 (XXSEL_1TI): Likewise.
9269 (XXSEL_1TI_UNS): Likewise.
9270 (VPERM_1TI): Likewise.
9271 (VPERM_1TI_UNS): Likewise.
9272 (XXPERMDI_1TI): Likewise.
9273 (SET_1TI): Likewise.
9274 (LXVD2X_V1TI): Likewise.
9275 (STXVD2X_V1TI): Likewise.
9276 (VEC_INIT_V1TI): Likewise.
9277 (VEC_SET_V1TI): Likewise.
9278 (VEC_EXT_V1TI): Likewise.
9279 (EQV_V1TI): Likewise.
9280 (NAND_V1TI): Likewise.
9281 (ORC_V1TI): Likewise.
9282 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
9283 added in ISA 2.07. Add both normal 'altivec' builtins, and the
9285 (VADDUQM): Likewise.
9286 (VSUBCUQ): Likewise.
9287 (VADDEUQM): Likewise.
9288 (VADDECUQ): Likewise.
9289 (VSUBEUQM): Likewise.
9290 (VSUBECUQ): Likewise.
9292 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
9293 __int128_t and __uint128_t types.
9294 (__uint128_type): Likewise.
9295 (altivec_categorize_keyword): Add support for vector __int128_t,
9296 vector __uint128_t, vector __int128, and vector unsigned __int128
9297 as a container type for TImode operations that need to be done in
9298 VSX/Altivec registers.
9299 (rs6000_macro_to_expand): Likewise.
9300 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
9301 to support 128-bit integer instructions vaddcuq, vadduqm,
9302 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
9303 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
9305 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
9306 for V1TImode, and set up preferences to use VSX/Altivec registers.
9307 Setup VSX reload handlers.
9308 (rs6000_debug_reg_global): Likewise.
9309 (rs6000_init_hard_regno_mode_ok): Likewise.
9310 (rs6000_preferred_simd_mode): Likewise.
9311 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
9312 (easy_altivec_constant): Likewise.
9313 (output_vec_const_move): Likewise.
9314 (rs6000_expand_vector_set): Convert V1TImode set and extract to
9316 (rs6000_expand_vector_extract): Likewise.
9317 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
9319 (rs6000_const_vec): Add support for V1TImode.
9320 (rs6000_emit_le_vsx_load): Swap double words when loading or
9321 storing TImode/V1TImode.
9322 (rs6000_emit_le_vsx_store): Likewise.
9323 (rs6000_emit_le_vsx_move): Likewise.
9324 (rs6000_emit_move): Add support for V1TImode.
9325 (altivec_expand_ld_builtin): Likewise.
9326 (altivec_expand_st_builtin): Likewise.
9327 (altivec_expand_vec_init_builtin): Likewise.
9328 (altivec_expand_builtin): Likewise.
9329 (rs6000_init_builtins): Add support for V1TImode type. Add
9330 support for ISA 2.07 128-bit integer builtins. Define type names
9331 for the VSX/Altivec vector types.
9332 (altivec_init_builtins): Add support for overloaded vector
9333 functions with V1TImode type.
9334 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
9335 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
9337 (rs6000_split_128bit_ok_p): Likewise.
9338 (rs6000_handle_altivec_attribute): Create V1TImode from vector
9339 __int128_t and vector __uint128_t.
9341 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
9342 and mode attributes.
9349 (VS_scalar): Likewise.
9350 (VS_double): Likewise.
9351 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
9353 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
9354 we support the ISA 2.07 128-bit integer arithmetic instructions.
9355 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
9356 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
9357 and TImode types for use with the builtin functions.
9358 (V1TI_type_node): Likewise.
9359 (unsigned_V1TI_type_node): Likewise.
9360 (intTI_type_internal_node): Likewise.
9361 (uintTI_type_internal_node): Likewise.
9363 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
9364 128-bit builtin functions.
9365 (UNSPEC_VADDEUQM): Likewise.
9366 (UNSPEC_VADDECUQ): Likewise.
9367 (UNSPEC_VSUBCUQ): Likewise.
9368 (UNSPEC_VSUBEUQM): Likewise.
9369 (UNSPEC_VSUBECUQ): Likewise.
9370 (VM): Add V1TImode to vector mode iterators.
9372 (VI_unit): Likewise.
9373 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
9374 (altivec_vaddcuq): Likewise.
9375 (altivec_vsubuqm): Likewise.
9376 (altivec_vsubcuq): Likewise.
9377 (altivec_vaddeuqm): Likewise.
9378 (altivec_vaddecuq): Likewise.
9379 (altivec_vsubeuqm): Likewise.
9380 (altivec_vsubecuq): Likewise.
9382 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
9384 (BOOL_128): Likewise.
9385 (BOOL_REGS_OUTPUT): Likewise.
9386 (BOOL_REGS_OP1): Likewise.
9387 (BOOL_REGS_OP2): Likewise.
9388 (BOOL_REGS_UNARY): Likewise.
9389 (BOOL_REGS_AND_CR0): Likewise.
9391 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
9392 128-bit integer builtin support.
9393 (vec_vadduqm): Likewise.
9394 (vec_vaddecuq): Likewise.
9395 (vec_vaddeuqm): Likewise.
9396 (vec_vsubecuq): Likewise.
9397 (vec_vsubeuqm): Likewise.
9398 (vec_vsubcuq): Likewise.
9399 (vec_vsubuqm): Likewise.
9401 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
9402 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
9403 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
9404 128-bit integer add/subtract to ISA 2.07.
9406 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
9408 * config/arc/arc.c (arc_predicate_delay_insns):
9409 Fix third argument passed to conditionalize_nonjump.
9411 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
9413 * config/aarch64/aarch64-builtins.c
9414 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
9415 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
9416 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
9417 instead of __builtin_lfloor.
9418 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
9420 2014-03-12 Jakub Jelinek <jakub@redhat.com>
9422 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
9423 (tree_ssa_ifcombine_bb_1): New function.
9424 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
9425 is an empty forwarder block to then_bb or vice versa and then_bb
9426 and else_bb are effectively swapped.
9428 2014-03-12 Christian Bruel <christian.bruel@st.com>
9431 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
9432 REG_CFA_DEF_CFA note.
9433 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
9434 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
9436 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
9438 PR tree-optimization/60454
9439 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
9441 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9443 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
9444 Do not define target_cpu_default2 to generic.
9445 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
9446 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
9447 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
9449 2014-03-12 Jakub Jelinek <jakub@redhat.com>
9450 Marc Glisse <marc.glisse@inria.fr>
9452 PR tree-optimization/60502
9453 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
9454 instead of build_low_bits_mask.
9456 2014-03-12 Jakub Jelinek <jakub@redhat.com>
9459 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
9460 if there are multiple uses, but op doesn't live on E edge.
9461 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
9462 clobber stmts before __builtin_unreachable.
9464 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
9466 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
9467 hard_frame_pointer_rtx.
9468 * cse.c (cse_insn): Remove volatile check.
9469 * cselib.c (cselib_process_insn): Likewise.
9470 * dse.c (scan_insn): Likewise.
9472 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
9474 * config/arc/arc.c (conditionalize_nonjump): New function,
9476 (arc_ifcvt): ... this.
9477 (arc_predicate_delay_insns): Use it.
9479 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
9481 * config/arc/predicates.md (extend_operand): During/after reload,
9482 allow const_int_operand.
9483 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
9484 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
9485 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
9487 (umulsi3_highpart_i): Likewise.
9489 2014-03-11 Richard Biener <rguenther@suse.de>
9491 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
9492 Add asserts to guard possible wrong-code bugs.
9494 2014-03-11 Richard Biener <rguenther@suse.de>
9496 PR tree-optimization/60429
9497 PR tree-optimization/60485
9498 * tree-ssa-structalias.c (set_union_with_increment): Properly
9499 take into account all fields that overlap the shifted vars.
9500 (do_sd_constraint): Likewise.
9501 (do_ds_constraint): Likewise.
9502 (get_constraint_for_ptr_offset): Likewise.
9504 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
9506 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
9507 (nios2_compute_frame_layout):
9508 Add calculation of cfun->machine->fp_save_offset.
9509 (nios2_expand_prologue): Correct setting of frame pointer register
9511 (nios2_expand_epilogue): Update recovery of stack pointer from
9512 frame pointer accordingly.
9513 (nios2_initial_elimination_offset): Update calculation of offset
9514 for eliminating to HARD_FRAME_POINTER_REGNUM.
9516 2014-03-10 Jakub Jelinek <jakub@redhat.com>
9519 * ipa.c (symtab_remove_unreachable_nodes): Don't call
9520 cgraph_get_create_node on VAR_DECLs.
9522 2014-03-10 Richard Biener <rguenther@suse.de>
9525 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
9527 2014-03-08 Douglas B Rupp <rupp@gnat.com>
9529 * config/vms/vms.opt (vms_float_format): New variable.
9531 2014-03-08 Tobias Burnus <burnus@net-b.de>
9533 * doc/invoke.texi (-fcilkplus): Update implementation status.
9535 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
9536 Richard Biener <rguenther@suse.de>
9538 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
9539 consistently accross all TUs.
9540 (run_gcc): Enable -fshort-double automatically at link at link-time
9541 and disallow override.
9543 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
9546 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
9547 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
9548 if they can't be used.
9550 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9552 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
9553 for Solaris 11/x86 ld.
9554 * configure: Regenerate.
9556 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9558 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
9559 (LIB_TLS_SPEC): Save as ld_tls_libs.
9560 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
9561 (HAVE_AS_IX86_TLSLDM): New test.
9562 * configure, config.in: Regenerate.
9563 * config/i386/i386.c (legitimize_tls_address): Fall back to
9564 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
9565 cannot support TLS_MODEL_LOCAL_DYNAMIC.
9566 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
9567 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
9569 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
9571 * common.opt (fira-loop-pressure): Mark as optimization.
9573 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
9575 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
9576 an OpenMP mappable type.
9578 2014-03-06 Matthias Klose <doko@ubuntu.com>
9580 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
9581 MULTILIB_OSDIRNAMES is not defined.
9583 2014-03-06 Jakub Jelinek <jakub@redhat.com>
9584 Meador Inge <meadori@codesourcery.com>
9587 * config/arm/arm.c (arm_tls_symbol_p): Remove.
9588 (arm_legitimize_address): Call legitimize_tls_address for any
9589 arm_tls_referenced_p expression, handle constant addend. Call it
9590 before testing for !TARGET_ARM.
9591 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
9593 2014-03-06 Richard Biener <rguenther@suse.de>
9599 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
9601 * tree-streamer.c (record_common_node): Assert we don't record
9602 nodes with type double.
9603 (preload_common_node): Skip type double, complex double and double
9604 pointer since it is now frontend dependent due to fshort-double option.
9606 2014-03-06 Richard Biener <rguenther@suse.de>
9608 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
9609 or -fno-lto is specified and the linker has full plugin support.
9610 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
9611 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
9612 * lto-wrapper.c (merge_and_complain): Merge compile-time
9613 optimization levels.
9614 (run_gcc): And pass it through to the link options.
9616 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
9620 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9622 * cselib.c (remove_useless_values): Skip to avoid quadratic
9623 behavior if the condition moved from...
9624 (cselib_process_insn): ... here holds.
9626 2014-03-05 Jakub Jelinek <jakub@redhat.com>
9629 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
9630 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
9633 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
9634 (TM_H): Add x86-tune.def.
9636 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9638 * config/aarch64/aarch64.c (generic_tunings):
9639 Use cortexa57_extra_costs.
9641 2014-03-05 Jakub Jelinek <jakub@redhat.com>
9644 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
9645 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
9646 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
9649 2014-03-04 Heiher <r@hev.cc>
9651 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
9652 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
9654 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
9656 * config/i386/predicates.md (const2356_operand): Change to ...
9657 (const2367_operand): ... this.
9658 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
9660 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9661 (*avx512pf_scatterpf<mode>sf): Ditto.
9662 (avx512pf_scatterpf<mode>df): Ditto.
9663 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9664 (*avx512pf_scatterpf<mode>df): Ditto.
9665 * config/i386/i386.c (ix86_expand_builtin): Update
9666 incorrect hint operand error message.
9668 2014-03-04 Richard Biener <rguenther@suse.de>
9670 * lto-section-in.c (lto_get_section_data): Fix const cast.
9672 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
9674 * tree-streamer.c (record_common_node): Assert we don't record
9675 nodes with type double.
9676 (preload_common_node): Skip type double, complex double and double
9677 pointer since it is now frontend dependent due to fshort-double option.
9679 2014-03-04 Richard Biener <rguenther@suse.de>
9682 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
9683 (lto_input_toplevel_asms): Likewise.
9684 * lto-section-in.c (lto_get_section_data): Instead do it here
9687 2014-03-04 Richard Biener <rguenther@suse.de>
9689 PR tree-optimization/60382
9690 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
9691 dead PHIs a reduction.
9693 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
9695 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
9697 (_mm_prefetch): Move out of GCC target("sse") pragma.
9698 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
9699 GCC target("prfchw") pragma.
9700 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
9702 * config/i386/i386.c (ix86_option_override_internal): Enable
9703 -mprfchw with -mprefetchwt1.
9705 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
9707 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
9710 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
9712 * opts.h (CL_PCH_IGNORE): Define.
9713 * targhooks.c (option_affects_pch_p):
9714 Return false for options that have CL_PCH_IGNORE set.
9715 * opt-functions.awk: Process PchIgnore.
9716 * doc/options.texi: Document PchIgnore.
9718 * config/arc/arc.opt (misize): Add PchIgnore property.
9720 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9722 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
9723 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
9724 constraint on constants to permit them being loaded into
9725 GENERAL_REGS or BASE_REGS.
9727 2014-03-03 Nick Clifton <nickc@redhat.com>
9729 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
9730 anti-cacnonical alternatives.
9731 (negandhi3_real): New pattern.
9732 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
9734 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9736 * config/avr/avr-mcus.def: Remove atxmega16x1.
9737 * config/avr/avr-tables.opt: Regenerate.
9738 * config/avr/t-multilib: Regenerate.
9739 * doc/avr-mmcu.texi: Regenerate.
9741 2014-03-03 Tobias Grosser <tobias@grosser.es>
9742 Mircea Namolaru <mircea.namolaru@inria.fr>
9744 PR tree-optimization/58028
9745 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
9748 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9750 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
9751 not handled by recognizers.
9753 2014-03-03 Jakub Jelinek <jakub@redhat.com>
9756 * function.c (expand_function_end): Don't emit
9757 clobber_return_register sequence if clobber_after is a BARRIER.
9758 * cfgexpand.c (construct_exit_block): Append instructions before
9759 return_label to prev_bb.
9761 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9763 * config/rs6000/constraints.md: Document reserved use of "wc".
9765 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9768 * ipa.c (function_and_variable_visibility): When dissolving comdat
9769 group, also set all symbols to local.
9771 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9776 2013-12-14 Jan Hubicka <jh@suse.cz>
9778 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
9780 2014-03-02 Jon Beniston <jon@beniston.com>
9786 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
9787 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
9788 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
9789 (simple_return, *simple_return): New patterns
9790 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
9791 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
9793 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
9795 * dwarf2out.c (gen_subprogram_die): Tidy.
9797 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
9800 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
9801 (*mov_t_msb_neg_negc): ... this new insn.
9803 2014-02-28 Jason Merrill <jason@redhat.com>
9806 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
9809 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
9812 * dwarf2out.c (decltype_auto_die): New static.
9813 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
9814 (gen_type_die_with_usage): Handle 'decltype(auto)'.
9815 (is_cxx_auto): Likewise.
9817 2014-02-28 Ian Bolton <ian.bolton@arm.com>
9819 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
9820 we are not using general regs only.
9822 2014-02-28 Richard Biener <rguenther@suse.de>
9825 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
9826 previous fix and only allow to remove trivial pre-headers
9827 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
9828 (remove_forwarder_block): Properly update the latch of a loop.
9830 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9833 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
9834 (cselib_preserved_hash_table): New.
9835 (preserve_constants_and_equivs): Move preserved vals to it.
9836 (cselib_find_slot): Look it up first.
9837 (cselib_init): Initialize it.
9838 (cselib_finish): Release it.
9839 (dump_cselib_table): Dump it.
9841 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9844 * cselib.c (remove_useless_values): Skip to avoid quadratic
9845 behavior if the condition moved from...
9846 (cselib_process_insn): ... here holds.
9848 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9851 * var-tracking.c (vt_initialize): Apply the same condition to
9852 preserve the CFA base value.
9854 2014-02-28 Joey Ye <joey.ye@arm.com>
9857 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
9858 if reload in progress or completed.
9860 2014-02-28 Tobias Burnus <burnus@net-b.de>
9863 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
9866 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
9868 * doc/tm.texi.in (Condition Code Status): Update documention for
9869 relative locations of cc0-setter and cc0-user.
9871 2014-02-27 Jeff Law <law@redhat.com>
9873 PR rtl-optimization/52714
9874 * combine.c (try_combine): When splitting an unrecognized PARALLEL
9875 into two independent simple sets, if I3 is a jump, ensure the
9876 pattern we place into I3 is a (set (pc) ...).
9878 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
9879 Jeff Law <law@redhat.com>
9881 PR rtl-optimization/49847
9882 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
9883 are in different blocks.
9884 * doc/tm.texi (Condition Code Status): Update documention for
9885 relative locations of cc0-setter and cc0-user.
9887 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
9890 * lra.c (lra_emit_add): Check SUBREG too.
9892 2014-02-27 Andreas Schwab <schwab@suse.de>
9894 * config/m68k/m68k.c (m68k_option_override): Disable
9895 -flive-range-shrinkage for classic m68k.
9896 (m68k_override_options_after_change): Likewise.
9898 2014-02-27 Marek Polacek <polacek@redhat.com>
9901 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
9902 -Wmaybe-uninitialized.
9904 2014-02-27 Alan Modra <amodra@gmail.com>
9907 * reload1.c (emit_input_reload_insns): When reload_override_in,
9908 set old to rl->in_reg when rl->in_reg is a subreg.
9910 2014-02-26 Richard Biener <rguenther@suse.de>
9913 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
9915 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9917 * common/config/i386/predicates.md (const1256_operand): Remove.
9918 (const2356_operand): New.
9919 (const_1_to_2_operand): Remove.
9920 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9921 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9922 (*avx512pf_gatherpf<mode>sf): Ditto.
9923 (avx512pf_gatherpf<mode>df): Ditto.
9924 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9925 (*avx512pf_gatherpf<mode>df): Ditto.
9926 (avx512pf_scatterpf<mode>sf): Ditto.
9927 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9928 (*avx512pf_scatterpf<mode>sf): Ditto.
9929 (avx512pf_scatterpf<mode>df): Ditto.
9930 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9931 (*avx512pf_scatterpf<mode>df): Ditto.
9932 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
9934 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
9936 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
9937 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
9938 (_mm512_mask_testn_epi64_mask): Move to ...
9939 * config/i386/avx512cdintrin.h: Here.
9940 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
9941 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
9942 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
9943 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
9944 TARGET_AVX512F from TARGET_AVX512CD.
9946 2014-02-26 Richard Biener <rguenther@suse.de>
9949 * ipa.c (walk_polymorphic_call_targets): Properly guard
9950 call to inline_update_overall_summary.
9952 2014-02-26 Bin Cheng <bin.cheng@arm.com>
9955 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
9956 and latches only if requested. Fix latch if it is removed.
9957 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
9958 LOOPS_HAVE_PREHEADERS.
9960 2014-02-25 Andrew Pinski <apinski@cavium.com>
9962 * builtins.c (expand_builtin_thread_pointer): Create a new target
9963 when the target is NULL.
9965 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
9967 PR rtl-optimization/60317
9968 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9969 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9970 * lra-assigns.c: Include params.h.
9971 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
9972 other reload pseudos considerations.
9974 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9976 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
9977 to use canonical form for nor<mode>3.
9979 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9982 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
9985 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9987 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
9988 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
9989 (ix86_handle_option): Handle OPT_mprefetchwt1.
9990 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
9991 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9993 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9994 OPTION_MASK_ISA_PREFETCHWT1.
9995 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
9996 (PTA_PREFETCHWT1): New.
9997 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
9998 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
9999 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
10000 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
10001 (*prefetch_avx512pf_<mode>_: Change into ...
10002 (*prefetch_prefetchwt1_<mode>: This.
10003 * config/i386/i386.opt (mprefetchwt1): New.
10004 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
10005 (_mm_prefetch): Handle intent to write.
10006 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
10008 2014-02-25 Richard Biener <rguenther@suse.de>
10010 PR middle-end/60291
10011 * emit-rtl.c (mem_attrs_htab): Remove.
10012 (mem_attrs_htab_hash): Likewise.
10013 (mem_attrs_htab_eq): Likewise.
10014 (set_mem_attrs): Always allocate new mem-attrs when something changed.
10015 (init_emit_once): Do not allocate mem_attrs_htab.
10017 2014-02-25 Richard Biener <rguenther@suse.de>
10020 * lto-opts.c (lto_write_options): Output non-explicit conservative
10021 -fwrapv, -fno-trapv and -fno-strict-overflow.
10022 * lto-wrapper.c (merge_and_complain): Handle merging those options.
10023 (run_gcc): And pass them through.
10025 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
10027 * sel-sched.c (calculate_new_fences): New parameter ptime.
10028 Calculate it as a maximum over all fence cycles.
10029 (sel_sched_region_2): Adjust the call to calculate_new_fences.
10030 Print the final schedule timing when sched_verbose.
10032 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
10034 PR rtl-optimization/60292
10035 * sel-sched.c (fill_vec_av_set): Do not reset target availability
10036 bit fot the fence instruction.
10038 2014-02-24 Alangi Derick <alangiderick@gmail.com>
10040 * calls.h: Fix typo in comment.
10042 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
10044 * config/pa/pa.c (pa_output_move_double): Don't valididate when
10045 adjusting offsetable addresses.
10047 2014-02-24 Guozhi Wei <carrot@google.com>
10049 * sparseset.h (sparseset_pop): Fix the wrong index.
10051 2014-02-24 Walter Lee <walt@tilera.com>
10053 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
10054 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
10056 * common/config/tilegx/tilegx-common.c
10057 (TARGET_DEFAULT_TARGET_FLAGS): Define.
10058 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
10059 (LINK_SPEC): Ditto.
10060 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
10061 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
10062 (tilegx_gimplify_va_arg_expr): Handle big endian.
10063 (tilegx_expand_unaligned_load): Ditto.
10064 (tilegx_expand_unaligned_store): Ditto.
10065 (TARGET_RETURN_IN_MSB): New.
10066 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
10067 (TARGET_ENDIAN_DEFAULT): New.
10068 (TARGET_BIG_ENDIAN): Handle big endian.
10069 (BYTES_BIG_ENDIAN): Ditto.
10070 (WORDS_BIG_ENDIAN): Ditto.
10071 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
10072 (ENDIAN_SPEC): New.
10073 (EXTRA_SPECS): New.
10074 * config/tilegx/tilegx.md (extv): Handle big endian.
10076 (insn_st<n>): Ditto.
10077 (insn_st<n>_add<bitsuffix>): Ditto.
10078 (insn_stnt<n>): Ditto.
10079 (insn_stnt<n>_add<bitsuffix>):Ditto.
10080 (vec_interleave_highv8qi): Handle big endian.
10081 (vec_interleave_highv8qi_be): New.
10082 (vec_interleave_highv8qi_le): New.
10083 (insn_v1int_h): Handle big endian.
10084 (vec_interleave_lowv8qi): Handle big endian.
10085 (vec_interleave_lowv8qi_be): New.
10086 (vec_interleave_lowv8qi_le): New.
10087 (insn_v1int_l): Handle big endian.
10088 (vec_interleave_highv4hi): Handle big endian.
10089 (vec_interleave_highv4hi_be): New.
10090 (vec_interleave_highv4hi_le): New.
10091 (insn_v2int_h): Handle big endian.
10092 (vec_interleave_lowv4hi): Handle big endian.
10093 (vec_interleave_lowv4hi_be): New.
10094 (vec_interleave_lowv4hi_le): New.
10095 (insn_v2int_l): Handle big endian.
10096 (vec_interleave_highv2si): Handle big endian.
10097 (vec_interleave_highv2si_be): New.
10098 (vec_interleave_highv2si_le): New.
10099 (insn_v4int_h): Handle big endian.
10100 (vec_interleave_lowv2si): Handle big endian.
10101 (vec_interleave_lowv2si_be): New.
10102 (vec_interleave_lowv2si_le): New.
10103 (insn_v4int_l): Handle big endian.
10104 * config/tilegx/tilegx.opt (mbig-endian): New option.
10105 (mlittle-endian): New option.
10106 * doc/install.texi: Document tilegxbe-linux.
10107 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
10109 2014-02-24 Martin Jambor <mjambor@suse.cz>
10112 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
10113 there are no parameter descriptors.
10115 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
10117 PR rtl-optimization/60268
10118 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
10119 initialization to ...
10120 (sched_rgn_init): ... here.
10121 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
10123 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
10125 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
10128 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
10130 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
10133 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
10135 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
10136 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
10138 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
10140 * config/microblaze/predicates.md: Add cmp_op predicate.
10141 * config/microblaze/microblaze.md: Add branch_compare instruction
10142 which uses cmp_op predicate and emits cmp insn before branch.
10143 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
10144 to microblaze_expand_conditional_branch and consolidate logic.
10145 (microblaze_expand_conditional_branch): emit branch_compare
10146 insn instead of handling cmp op separate from branch insn.
10148 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10150 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
10153 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10155 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
10156 define_insn with define_expand and new define_insn
10157 *altivec_lve<VI_char>x_internal.
10158 (altivec_stve<VI_char>x): Replace define_insn with define_expand
10159 and new define_insn *altivec_stve<VI_char>x_internal.
10160 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
10162 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
10164 (altivec_expand_stvex_be): New function.
10166 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
10168 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
10169 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
10170 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
10171 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
10173 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
10176 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
10177 instead of emit_move_insn.
10179 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10181 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
10182 vspltw with vsldoi.
10183 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
10184 gen_altivec_vsumsws.
10186 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10188 * config/rs6000/altivec.md (altivec_lvxl): Rename as
10189 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
10190 (altivec_lvxl_<mode>): New define_expand incorporating
10191 -maltivec=be semantics where needed.
10192 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
10193 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
10194 semantics where needed.
10195 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
10196 (altivec_stvx_<mode>): New define_expand incorporating
10197 -maltivec=be semantics where needed.
10198 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
10199 VM2 iterator instead of V4SI.
10200 (altivec_stvxl_<mode>): New define_expand incorporating
10201 -maltivec=be semantics where needed.
10202 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
10203 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
10204 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
10205 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
10206 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
10207 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
10208 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
10209 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
10210 ALTIVEC_BUILTIN_STVXL.
10211 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
10212 (altivec_expand_stvx_be): Likewise.
10213 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
10214 (altivec_expand_lvx_be): Likewise.
10215 (altivec_expand_stvx_be): Likewise.
10216 (altivec_expand_builtin): Add cases for
10217 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
10218 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
10219 (altivec_init_builtins): Add definitions for
10220 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
10221 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
10223 2014-02-21 Catherine Moore <clm@codesourcery.com>
10225 * doc/invoke.texi (mvirt, mno-virt): Document.
10226 * config/mips/mips.opt (mvirt): New option.
10227 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
10229 2014-02-21 Richard Biener <rguenther@suse.de>
10231 PR tree-optimization/60276
10232 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
10233 (STMT_VINFO_MIN_NEG_DIST): New macro.
10234 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
10235 STMT_VINFO_MIN_NEG_DIST.
10236 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
10237 made for negative dependence distances still hold.
10239 2014-02-21 Richard Biener <rguenther@suse.de>
10241 PR middle-end/60291
10242 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
10243 DECL_INITIAL for globals not in the current function context.
10245 2014-02-21 Jakub Jelinek <jakub@redhat.com>
10247 PR tree-optimization/56490
10248 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
10249 * tree-ssa-uninit.c: Include params.h.
10250 (compute_control_dep_chain): Add num_calls argument, return false
10251 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
10252 num_calls to recursive call.
10253 (find_predicates): Change dep_chain into normal array,
10254 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
10255 variable and adjust compute_control_dep_chain caller.
10256 (find_def_preds): Likewise.
10258 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
10260 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
10261 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
10263 2014-02-21 Nick Clifton <nickc@redhat.com>
10265 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
10266 (pushhi1): Likewise.
10267 (popqi1): Add mode to pre_dec.
10268 (pophi1): Likewise.
10270 2014-02-21 Jakub Jelinek <jakub@redhat.com>
10272 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
10273 mode for mask of V8SFmode permutation.
10275 2014-02-20 Richard Henderson <rth@redhat.com>
10278 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
10279 a new pseudo for OLDVAL.
10281 2014-02-20 Jakub Jelinek <jakub@redhat.com>
10284 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
10285 gen_reg_rtx if d->testing_p.
10286 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
10287 if d->testing_p and we will certainly return true.
10288 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
10291 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
10293 * emit-rtl.c (gen_reg_rtx): Assert that
10294 crtl->emit.regno_pointer_align_length is non-zero.
10296 2014-02-20 Richard Henderson <rth@redhat.com>
10299 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
10300 on failure the store back into EXPECT.
10302 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
10303 Sandra Loosemore <sandra@codesourcery.com>
10305 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
10306 * config/nios2/nios2.c (nios2_function_profiler): Add
10307 -fPIC (flag_pic == 2) support.
10308 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
10309 (nios2_large_offset_p): New function.
10310 (nios2_unspec_reloc_p): Move up position, update to use
10311 nios2_large_offset_p.
10312 (nios2_unspec_address): Remove function.
10313 (nios2_unspec_offset): New function.
10314 (nios2_large_got_address): New function.
10315 (nios2_got_address): Add large offset support.
10316 (nios2_legitimize_tls_address): Update usage of removed and new
10318 (nios2_symbol_binds_local_p): New function.
10319 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
10320 (nios2_legitimize_address): Update to use nios2_large_offset_p.
10321 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
10322 (nios2_print_operand): Merge H/L processing, add hiadj/lo
10323 processing for (const (unspec ...)).
10324 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
10326 2014-02-20 Richard Biener <rguenther@suse.de>
10328 * tree-cfg.c (replace_uses_by): Mark altered BBs before
10329 doing the substitution.
10330 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
10332 2014-02-20 Martin Jambor <mjambor@suse.cz>
10335 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
10336 info when checking whether lattices are bottom.
10338 2014-02-20 Richard Biener <rguenther@suse.de>
10340 PR middle-end/60221
10341 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
10344 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
10347 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
10348 parameter specifying the scaling.
10349 (inline_call): Update.
10350 (want_inline_recursively): Guard division by zero.
10351 (recursive_inlining): Update.
10352 * ipa-inline.h (clone_inlined_nodes): Update.
10354 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
10357 * config/i386/i386.c (classify_argument): Pass structures of size
10358 64 bytes or less in register.
10360 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
10361 Kirill Yukhin <kirill.yukhin@intel.com>
10363 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
10364 (_mm_rcp28_round_ss): Ditto.
10365 (_mm_rsqrt28_round_sd): Ditto.
10366 (_mm_rsqrt28_round_ss): Ditto.
10367 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
10368 (_mm_rcp14_round_ss): Ditto.
10369 (_mm_rsqrt14_round_sd): Ditto.
10370 (_mm_rsqrt14_round_ss): Ditto.
10371 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
10372 the first input operand, get rid of match_dup.
10373 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
10375 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10377 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
10378 operand as the first input operand, set type attribute.
10379 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10380 Set type attribute.
10381 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
10382 operand as the first input operand, set type attribute.
10384 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10386 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
10389 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
10392 * config/i386/i386.c (construct_container): Remove TFmode check
10393 for X86_64_INTEGER_CLASS.
10395 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
10398 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
10399 only when -Wpsabi is enabled.
10401 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
10404 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
10405 passing arrays in registers are the same as for structs, so remove the
10406 special case for them.
10408 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
10410 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
10411 destination type, extract only the valid bits if the source type is not
10412 integral and has a different mode.
10414 2014-02-19 Richard Biener <rguenther@suse.de>
10417 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
10420 2014-02-19 Richard Biener <rguenther@suse.de>
10423 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
10424 (ipa_modify_call_arguments): Emit an argument load explicitely and
10425 preserve virtual SSA form there and for the replacement call.
10426 Do not update SSA form nor free dominance info.
10428 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
10430 * ipa.c (function_and_variable_visibility): Also clear WEAK
10431 flag when disolving COMDAT_GROUP.
10433 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
10435 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
10436 * ipa-prop.c (ipa_set_jf_known_type): Return early when
10437 not devirtualizing.
10438 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
10439 do more sanity checks.
10440 (detect_type_change): Return true when giving up early.
10441 (compute_complex_assign_jump_func): Fix type parameter of
10442 ipa_set_ancestor_jf.
10443 (compute_complex_ancestor_jump_func): Likewise.
10444 (update_jump_functions_after_inlining): Fix updating of
10446 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
10448 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
10450 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
10451 inline clones when edge disappears.
10453 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
10456 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
10457 Split 64-bit moves into 2 patterns. Do not allow the use of
10458 direct move for TDmode in little endian, since the decimal value
10459 has little endian bytes within a word, but the 64-bit pieces are
10460 ordered in a big endian fashion, and normal subreg's of TDmode are
10462 (mov<mode>_64bit_dm): Likewise.
10463 (movtd_64bit_nodm): Likewise.
10465 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
10467 PR tree-optimization/60174
10468 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
10469 statement of an SSA_NAME that occurs in an abnormal PHI node.
10471 2014-02-18 Jakub Jelinek <jakub@redhat.com>
10474 * final.c (SEEN_BB): Remove.
10475 (SEEN_NOTE, SEEN_EMITTED): Renumber.
10476 (final_scan_insn): Don't force_source_line on second
10477 NOTE_INSN_BASIC_BLOCK.
10479 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
10482 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
10483 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
10484 (type_natural_mode): Warn ABI change when %zmm register is not
10485 available for AVX512F vector value passing.
10487 2014-02-18 Kai Tietz <ktietz@redhat.com>
10490 * config/i386/i386.c (ix86_expand_prologue): Use value in
10491 rax register as displacement when restoring %r10 or %rax.
10492 Fix wrong offset when restoring both registers.
10494 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
10496 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
10497 assertion with conditional return.
10499 2014-02-18 Jakub Jelinek <jakub@redhat.com>
10500 Uros Bizjak <ubizjak@gmail.com>
10503 * config/i386/driver-i386.c (host_detect_local_cpu): If
10504 YMM state is not saved by the OS, also clear has_f16c. Move
10505 CPUID 0x80000001 handling before YMM state saving checking.
10507 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
10509 PR rtl-optimization/58960
10510 * haifa-sched.c (alloc_global_sched_pressure_data): New,
10511 factored out from ...
10512 (sched_init): ... here.
10513 (free_global_sched_pressure_data): New, factored out from ...
10514 (sched_finish): ... here.
10515 * sched-int.h (free_global_sched_pressure_data): Declare.
10516 * sched-rgn.c (nr_regions_initial): New static global.
10517 (haifa_find_rgns): Initialize it.
10518 (schedule_region): Disable sched-pressure for the newly
10521 2014-02-17 Richard Biener <rguenther@suse.de>
10523 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
10524 release SSA defs of pattern stmts.
10526 2014-02-17 Richard Biener <rguenther@suse.de>
10528 * tree-inline.c (expand_call_inline): Release the virtual
10529 operand defined by the call we are about to inline.
10531 2014-02-17 Richard Biener <rguenther@suse.de>
10533 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
10535 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
10536 Ilya Tocar <ilya.tocar@intel.com>
10538 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
10539 arguments order in builtin.
10540 (_mm512_permutexvar_epi64): Ditto.
10541 (_mm512_mask_permutexvar_epi64): Ditto
10542 (_mm512_maskz_permutexvar_epi32): Ditto
10543 (_mm512_permutexvar_epi32): Ditto
10544 (_mm512_mask_permutexvar_epi32): Ditto
10546 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10548 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
10549 (p8_vmrgow): Likewise.
10551 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10553 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
10556 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
10559 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
10560 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
10561 into 64-bit and 32-bit moves. On 64-bit moves, add support for
10562 using direct move instructions on ISA 2.07. Also adjust
10563 instruction length for 64-bit.
10564 (mov<mode>_64bit, TFmode/TDmode): Likewise.
10565 (mov<mode>_32bit, TFmode/TDmode): Likewise.
10567 2014-02-15 Alan Modra <amodra@gmail.com>
10571 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
10572 find_replacement on parts of insn rtl that might be reloaded.
10574 2014-02-15 Richard Biener <rguenther@suse.de>
10576 PR tree-optimization/60183
10577 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
10578 (tree_ssa_phiprop): Calculate and free post-dominators.
10580 2014-02-14 Jeff Law <law@redhat.com>
10582 PR rtl-optimization/60131
10583 * ree.c (get_extended_src_reg): New function.
10584 (combine_reaching_defs): Use it rather than assuming location of REG.
10585 (find_and_remove_re): Verify first operand of extension is
10586 a REG before adding the insns to the copy list.
10588 2014-02-14 Roland McGrath <mcgrathr@google.com>
10590 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
10591 * configure: Regenerated.
10592 * config.in: Regenerated.
10593 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
10594 instead of ASM_SHORT.
10596 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
10597 Richard Earnshaw <rearnsha@arm.com>
10599 PR rtl-optimization/59535
10600 * lra-constraints.c (process_alt_operands): Encourage alternative
10601 when unassigned pseudo class is superset of the alternative class.
10602 (inherit_reload_reg): Don't inherit when optimizing for code size.
10603 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
10604 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
10605 modes not less than 4 for Thumb1.
10607 2014-02-14 Kyle McMartin <kyle@redhat.com>
10610 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
10612 2014-02-14 Richard Biener <rguenther@suse.de>
10614 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
10615 (get_frame_arg): Drop the assert with langhook types_compatible_p.
10616 Do not strip INDIRECT_REFs.
10618 2014-02-14 Richard Biener <rguenther@suse.de>
10621 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
10622 DECL_FUNCTION_SPECIFIC_TARGET.
10623 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
10624 * tree-streamer-out.c (pack_ts_target_option): Remove.
10625 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
10626 (write_ts_function_decl_tree_pointers): Do not stream
10627 DECL_FUNCTION_SPECIFIC_TARGET.
10628 * tree-streamer-in.c (unpack_ts_target_option): Remove.
10629 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
10630 (lto_input_ts_function_decl_tree_pointers): Do not stream
10631 DECL_FUNCTION_SPECIFIC_TARGET.
10633 2014-02-14 Jakub Jelinek <jakub@redhat.com>
10635 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
10636 (get_initial_def_for_induction, vectorizable_induction): Ignore
10637 debug stmts when looking for exit_phi.
10638 (vectorizable_live_operation): Fix up condition.
10640 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10642 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
10643 nreverse() because it changes the content of original tree list.
10645 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10647 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
10648 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
10650 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10652 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
10653 GNU coding standards.
10655 2014-02-13 Jakub Jelinek <jakub@redhat.com>
10658 * dwarf2out.c (gen_subprogram_die): Don't call
10659 add_calling_convention_attribute if subr_die is old_die.
10661 2014-02-13 Sharad Singhai <singhai@google.com>
10663 * doc/optinfo.texi: Fix order of nodes.
10665 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
10667 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
10668 operands[2], not operands[3].
10670 2014-02-13 Richard Biener <rguenther@suse.de>
10673 * doc/install.texi (ISL): Update recommended version to 0.12.2,
10674 mention the possibility of an in-tree build.
10675 (CLooG): Update recommended version to 0.18.1, mention the
10676 possibility of an in-tree build and clarify that the ISL
10677 bundled with CLooG does not work.
10679 2014-02-13 Jakub Jelinek <jakub@redhat.com>
10682 * expr.c (compress_float_constant): If x is a hard register,
10683 extend into a pseudo and then move to x.
10685 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
10687 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
10688 caused by bad second argument to warning_at() with -mhotpatch and
10689 nested functions (e.g. with gfortran).
10691 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
10693 * opts.c (option_name): Remove "enabled by default" rider.
10695 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
10697 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
10699 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
10700 Uros Bizjak <ubizjak@gmail.com>
10703 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
10704 * configure: Regenerated.
10706 2014-02-12 Richard Biener <rguenther@suse.de>
10708 * vec.c (vec_prefix::calculate_allocation): Move as
10709 inline variant to vec.h.
10710 (vec_prefix::calculate_allocation_1): New out-of-line version.
10711 * vec.h (vec_prefix::calculate_allocation_1): Declare.
10712 (vec_prefix::m_has_auto_buf): Rename to ...
10713 (vec_prefix::m_using_auto_storage): ... this.
10714 (vec_prefix::calculate_allocation): Inline the easy cases
10715 and dispatch to calculate_allocation_1 which doesn't need the
10717 (va_heap::reserve): Use gcc_checking_assert.
10718 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
10719 m_using_auto_storage.
10720 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
10722 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
10723 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
10724 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
10726 2014-02-12 Richard Biener <rguenther@suse.de>
10728 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
10729 when we found a dependence.
10731 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
10733 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
10735 (maybe_fold_stmt): ... into this new function.
10736 * omp-low.c (lower_omp): Update comment.
10738 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
10741 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
10744 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
10746 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
10747 identifiers in comments.
10748 (cortexa53_extra_costs): Likewise.
10749 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
10750 (cortexa7_extra_costs): Likewise.
10751 (cortexa12_extra_costs): Likewise.
10752 (cortexa15_extra_costs): Likewise.
10753 (v7m_extra_costs): Likewise.
10755 2014-02-12 Richard Biener <rguenther@suse.de>
10757 PR middle-end/60092
10758 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
10759 of posix_memalign being successful.
10760 (lower_stmt): Restrict lowering of posix_memalign to when
10761 -ftree-bit-ccp is enabled.
10763 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10765 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
10767 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
10769 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
10771 PR rtl-optimization/60116
10772 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
10773 other_insn once the combination has been validated.
10775 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
10778 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
10780 * ipa-devirt.c: Include demangle.h
10781 (odr_violation_reported): New static variable.
10782 (add_type_duplicate): Update odr_violations.
10783 (maybe_record_node): Add completep parameter; update it.
10784 (record_target_from_binfo): Add COMPLETEP parameter;
10785 update it as needed.
10786 (possible_polymorphic_call_targets_1): Likewise.
10787 (struct polymorphic_call_target_d): Add nonconstruction_targets;
10788 rename FINAL to COMPLETE.
10789 (record_targets_from_bases): Sanity check we found the binfo;
10790 fix COMPLETEP updating.
10791 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
10792 parameter, fix computing of COMPLETEP.
10793 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
10794 at LTO time do demangling.
10795 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
10796 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
10798 (gimple_get_virt_method_for_binfo): Likewise.
10799 * gimple-fold.h (gimple_get_virt_method_for_binfo,
10800 gimple_get_virt_method_for_vtable): Update prototypes.
10802 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
10805 * genautomata.c (add_presence_absence): Fix typo with
10806 {final_}presence_list.
10808 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10811 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
10812 for VSX/Altivec vectors that land in GPR registers.
10814 2014-02-11 Richard Henderson <rth@redhat.com>
10815 Jakub Jelinek <jakub@redhat.com>
10818 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
10819 around drhs if type conversion to lacc->type is not useless.
10821 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10823 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
10825 (cortex-a57.cortex-a53): Likewise.
10826 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
10828 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10830 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
10833 2014-02-11 Renlin Li <Renlin.Li@arm.com>
10835 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
10836 add_options_for_arm_vfp3.
10838 2014-02-11 Jeff Law <law@redhat.com>
10840 PR middle-end/54041
10841 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
10842 object with an undesirable mode.
10844 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10847 * config/i386/sol2-9.h: New file.
10848 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
10849 *-*-solaris2.9*): Use it.
10851 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10853 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
10854 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
10856 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10858 * config/microblaze/microblaze.c: Extend mcpu version format
10860 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
10862 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
10864 2014-02-10 Richard Henderson <rth@redhat.com>
10867 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
10868 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
10869 ms-abi vs -mno-accumulate-outgoing-args.
10870 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
10871 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
10874 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
10876 PR middle-end/60080
10877 * cfgexpand.c (expand_asm_operands): Attach source location to
10878 ASM_INPUT rtx objects.
10879 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
10881 2014-02-10 Nick Clifton <nickc@redhat.com>
10883 * config/mn10300/mn10300.c (popcount): New function.
10884 (mn10300_expand_prologue): Include saved registers in stack usage
10887 2014-02-10 Jeff Law <law@redhat.com>
10889 PR middle-end/52306
10890 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
10891 when changing the SET_DEST of a prior insn to avoid an input reload.
10893 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10895 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
10896 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
10897 -mcall-openbsd, or -mcall-linux.
10898 (CC1_ENDIAN_BIG_SPEC): Remove.
10899 (CC1_ENDIAN_LITTLE_SPEC): Remove.
10900 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10901 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
10902 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
10903 and %cc1_endian_default.
10904 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10906 2014-02-10 Richard Biener <rguenther@suse.de>
10908 PR tree-optimization/60115
10909 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
10910 MEM_REF handling. Properly verify that the accesses are not
10911 out of the objects bound.
10913 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10915 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
10918 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
10920 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
10921 proper constants and fix formatting.
10922 (possible_polymorphic_call_targets): Fix formatting.
10924 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
10925 Ilya Tocar <ilya.tocar@intel.com>
10927 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
10928 (_mm512_loadu_epi32): Renamed into...
10929 (_mm512_loadu_si512): This.
10930 (_mm512_storeu_epi32): Renamed into...
10931 (_mm512_storeu_si512): This.
10932 (_mm512_maskz_ceil_ps): Removed.
10933 (_mm512_maskz_ceil_pd): Ditto.
10934 (_mm512_maskz_floor_ps): Ditto.
10935 (_mm512_maskz_floor_pd): Ditto.
10936 (_mm512_floor_round_ps): Ditto.
10937 (_mm512_floor_round_pd): Ditto.
10938 (_mm512_ceil_round_ps): Ditto.
10939 (_mm512_ceil_round_pd): Ditto.
10940 (_mm512_mask_floor_round_ps): Ditto.
10941 (_mm512_mask_floor_round_pd): Ditto.
10942 (_mm512_mask_ceil_round_ps): Ditto.
10943 (_mm512_mask_ceil_round_pd): Ditto.
10944 (_mm512_maskz_floor_round_ps): Ditto.
10945 (_mm512_maskz_floor_round_pd): Ditto.
10946 (_mm512_maskz_ceil_round_ps): Ditto.
10947 (_mm512_maskz_ceil_round_pd): Ditto.
10948 (_mm512_expand_pd): Ditto.
10949 (_mm512_expand_ps): Ditto.
10950 * config/i386/i386.c (ix86_builtins): Remove
10951 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
10952 (bdesc_args): Ditto.
10953 * config/i386/predicates.md (const1256_operand): New.
10954 (const_1_to_2_operand): Ditto.
10955 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
10956 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10957 (*avx512pf_gatherpf<mode>sf): Ditto.
10958 (avx512pf_gatherpf<mode>df): Ditto.
10959 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10960 (*avx512pf_gatherpf<mode>df): Ditto.
10961 (avx512pf_scatterpf<mode>sf): Ditto.
10962 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10963 (*avx512pf_scatterpf<mode>sf): Ditto.
10964 (avx512pf_scatterpf<mode>df): Ditto.
10965 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10966 (*avx512pf_scatterpf<mode>df): Ditto.
10967 (avx512f_expand<mode>): Removed.
10968 (<shift_insn><mode>3<mask_name>): Change predicate type.
10970 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10972 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
10973 not at the end of datarefs vector use ordered_remove to avoid
10974 reordering datarefs vector.
10977 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
10978 mark local addressable non-static vars as GOVD_PRIVATE
10979 instead of GOVD_LOCAL.
10980 * omp-low.c (lower_omp_for): Move gimple_bind_vars
10981 and BLOCK_VARS of gimple_bind_block to new_stmt rather
10984 PR middle-end/60092
10985 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
10986 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
10987 assume_aligned or alloc_align attributes.
10988 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
10989 arguments. Handle also assume_aligned and alloc_align attributes.
10990 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
10991 calls to functions with assume_aligned or alloc_align attributes.
10992 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
10994 2014-02-08 Terry Guo <terry.guo@arm.com>
10996 * doc/invoke.texi: Document ARM -march=armv7e-m.
10998 2014-02-08 Jakub Jelinek <jakub@redhat.com>
11000 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
11001 flag on __cilkrts_rethrow builtin.
11004 * ipa-cp.c (determine_versionability): Fail at -O0
11005 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
11006 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
11009 2014-02-04 Jakub Jelinek <jakub@redhat.com>
11012 * tree-inline.c (copy_forbidden): Fail for
11013 __attribute__((optimize (0))) functions.
11015 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
11017 * varpool.c: Include pointer-set.h.
11018 (varpool_remove_unreferenced_decls): Variables in other partitions
11019 will not be output; be however careful to not lose information
11020 about partitioning.
11022 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
11024 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
11025 lookup in the vtable constructor.
11027 2014-02-07 Jeff Law <law@redhat.com>
11030 * config/m68k/m68k.md (ashldi_extsi): Turn into a
11031 define_insn_and_split.
11033 * ipa-inline.c (inline_small_functions): Fix typos.
11035 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11037 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
11038 (s390_can_use_return_insn): Declare.
11039 * config/s390/s390.h (EPILOGUE_USES): Define.
11040 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
11042 (s390_chunkify_start): Handle return JUMP_LABELs.
11043 (s390_early_mach): Emit a main_pool instruction on the entry edge.
11044 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
11045 (s390_can_use_return_insn): New functions.
11046 (s390_fix_long_loop_prediction): Handle conditional returns.
11047 (TARGET_SET_UP_BY_PROLOGUE): Define.
11048 * config/s390/s390.md (ANY_RETURN): New code iterator.
11049 (*creturn, *csimple_return, return, simple_return): New patterns.
11051 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11053 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
11054 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
11055 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
11056 REG_CFA_RESTORE list when deciding not to restore a register.
11058 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11060 * config/s390/s390.c: Include tree-pass.h and context.h.
11061 (s390_early_mach): New function, split out from...
11062 (s390_emit_prologue): ...here.
11063 (pass_data_s390_early_mach): New pass structure.
11064 (pass_s390_early_mach): New class.
11065 (s390_option_override): Create and register early_mach pass.
11066 Move to end of file.
11068 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11070 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
11071 to match for the exit block.
11073 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11075 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
11076 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
11077 Reject misaligned operands.
11079 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11081 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
11083 2014-02-07 Richard Biener <rguenther@suse.de>
11085 PR middle-end/60092
11086 * gimple-low.c (lower_builtin_posix_memalign): New function.
11087 (lower_stmt): Call it to lower posix_memalign in a way
11088 to make alignment info accessible.
11090 2014-02-07 Jakub Jelinek <jakub@redhat.com>
11093 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
11094 __builtin_setjmp_receiver.
11096 2014-02-07 Richard Biener <rguenther@suse.de>
11098 PR middle-end/60092
11099 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
11100 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
11101 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
11102 Handle BUILT_IN_POSIX_MEMALIGN.
11103 (find_func_clobbers): Likewise.
11104 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
11105 (call_may_clobber_ref_p_1): Likewise.
11107 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
11110 * ipa-devirt.c (record_target_from_binfo): Remove overactive
11113 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
11116 * lto-cgraph.c (lto_output_node): Use
11117 symtab_get_symbol_partitioning_class.
11118 (lto_output_varpool_node): likewise.
11119 (symtab_get_symbol_partitioning_class): Move here from
11120 lto/lto-partition.c
11121 * cgraph.h (symbol_partitioning_class): Likewise.
11122 (symtab_get_symbol_partitioning_class): Declare.
11124 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
11126 * ggc.h (ggc_internal_cleared_alloc): New macro.
11127 * vec.h (vec_safe_copy): Handle memory stats.
11128 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
11129 * target-globals.c (save_target_globals): Likewise.
11131 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
11134 * expr.c (emit_move_resolve_push): Export; be bit more selective
11135 on when to clear alias set.
11136 * expr.h (emit_move_resolve_push): Declare.
11137 * function.h (struct function): Add tail_call_marked.
11138 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
11139 * config/i386/i386-protos.h (ix86_expand_push): Remove.
11140 * config/i386/i386.md (TImode move expander): De not call
11142 (FP push expanders): Preserve memory attributes.
11143 * config/i386/sse.md (push<mode>1): Remove.
11144 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
11145 (ix86_expand_push): Remove.
11146 * config/i386/mmx.md (push<mode>1): Remove.
11148 2014-02-06 Jakub Jelinek <jakub@redhat.com>
11150 PR rtl-optimization/60030
11151 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
11152 lopart with paradoxical subreg before shifting it up by hprec.
11154 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11156 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
11157 Remove extra newline at end of file.
11158 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
11159 (arm_issue_rate): Handle cortexa57.
11160 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
11161 (cortex-a57.cortex-a53): Likewise.
11163 2014-02-06 Jakub Jelinek <jakub@redhat.com>
11166 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
11167 don't record in REG_FRAME_RELATED_EXPR registers not set in that
11169 (arm_expand_prologue): Adjust all callers.
11170 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
11171 info, registers also at the lowest numbered registers side. Use
11172 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
11176 * var-tracking.c (adjust_mems): Before adding a SET to
11177 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
11179 2014-02-06 Alan Modra <amodra@gmail.com>
11182 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
11183 change SDmode to DDmode when lra_in_progress.
11185 2014-02-06 Jakub Jelinek <jakub@redhat.com>
11187 PR middle-end/59150
11188 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
11189 free_data_ref on the dr first, and before goto again also set dr
11190 to the next dr. For simd_lane_access, free old datarefs[i] before
11191 overwriting it. For get_vectype_for_scalar_type failure, don't
11192 free_data_ref if simd_lane_access.
11194 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
11197 * tree.h (opts_for_fn): New inline function.
11198 (opt_for_fn): Define.
11199 * config/i386/i386.c (ix86_function_regparm): Use
11200 opt_for_fn (decl, optimize) instead of optimize.
11202 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
11204 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
11205 for SYMBOL_REF in large memory model.
11207 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11209 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
11210 and crypto support.
11211 (cortex-a57): Likewise.
11212 (cortex-a57.cortex-a53): Likewise.
11214 2014-02-06 Yury Gribov <y.gribov@samsung.com>
11215 Kugan Vivekanandarajah <kuganv@linaro.org>
11217 * config/arm/arm.c (arm_vector_alignment_reachable): Check
11219 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
11221 2014-02-06 Richard Biener <rguenther@suse.de>
11223 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
11224 set_loop_copy and initialize_original_copy_tables.
11226 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
11228 * config/aarch64/aarch64-simd.md
11229 (aarch64_ashr_simddi): Change QI to SI.
11231 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
11232 Jakub Jelinek <jakub@redhat.com>
11234 PR middle-end/60013
11235 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
11238 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11240 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
11241 CODE_FOR_altivec_vpku[hw]um to
11242 CODE_FOR_altivec_vpku[hw]um_direct.
11243 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
11244 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
11245 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
11246 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
11248 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11250 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
11251 generation for -maltivec=be.
11252 (altivec_vsumsws): Simplify redundant test.
11254 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11256 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
11257 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
11258 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
11259 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
11260 gen_altivec_vpkuwum.
11261 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
11263 (altivec_vpks<VI_char>ss): Likewise.
11264 (altivec_vpks<VI_char>us): Likewise.
11265 (altivec_vpku<VI_char>us): Likewise.
11266 (altivec_vpku<VI_char>um): Likewise.
11267 (altivec_vpku<VI_char>um_direct): New (copy of
11268 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
11270 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
11271 target is little endian and -maltivec=be is not specified.
11272 (*altivec_vupkhs<VU_char>_direct): New (copy of
11273 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
11274 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
11275 target is little endian and -maltivec=be is not specified.
11276 (*altivec_vupkls<VU_char>_direct): New (copy of
11277 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
11278 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
11279 little endian and -maltivec=be is not specified.
11280 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
11281 little endian and -maltivec=be is not specified.
11283 2014-02-05 Richard Henderson <rth@redhat.com>
11286 * combine-stack-adj.c: Revert r206943.
11287 * sched-int.h (struct deps_desc): Add last_args_size.
11288 * sched-deps.c (init_deps): Initialize it.
11289 (sched_analyze_insn): Add OUTPUT dependencies between insns that
11290 contain REG_ARGS_SIZE notes.
11292 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
11294 * lto-cgraph.c (asm_nodes_output): Make global.
11295 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
11296 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
11297 (driver_handle_option): Handle OPT_fwpa.
11299 2014-02-05 Jakub Jelinek <jakub@redhat.com>
11302 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
11303 a comment typo and formatting issue. If odr_hash hasn't been
11304 created, return vNULL and set *completep to false.
11306 PR middle-end/57499
11307 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
11308 bb with no successors.
11310 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
11313 * doc/invoke.texi (-march): Clarify documentation for ARM.
11314 (-mtune): Likewise.
11317 2014-02-05 Richard Biener <rguenther@suse.de>
11319 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
11320 when not vectorizing because of too many alias checks.
11321 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
11322 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
11324 2014-02-05 Nick Clifton <nickc@redhat.com>
11326 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
11327 accept extended registers in any mode when compiling for the MN10300.
11329 2014-02-05 Yury Gribov <y.gribov@samsung.com>
11331 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
11332 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
11333 sanitization attributes.
11334 (can_inline_edge_p): Likewise.
11335 (sanitize_attrs_match_for_inline_p): New function.
11337 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
11339 * ipa-prop.c (detect_type_change): Shor circuit testing of
11340 type changes on THIS pointer.
11342 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
11345 * config/pa/pa.c (legitimize_tls_address): Return original address
11346 if not passed a SYMBOL_REF rtx.
11347 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
11349 (pa_emit_move_sequence): Simplify TLS source operands.
11350 (pa_legitimate_constant_p): Reject all TLS constants.
11351 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
11352 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
11354 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
11356 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
11357 groups when we know they are controlled by LTO.
11358 * varasm.c (default_binds_local_p_1): If object is in other partition,
11359 it will be resolved locally.
11361 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
11363 * config/host-linux.c (linux_gt_pch_use_address): Don't
11364 use SSIZE_MAX because it is not always defined.
11366 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
11369 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
11370 threshold for pseudo splitting.
11371 (update_ebb_live_info): Process call argument hard registers and
11372 hard registers from insn definition too.
11373 (max_small_class_regs_num): New constant.
11374 (inherit_in_ebb): Update live hard regs through EBBs. Update
11375 reloads_num only for small register classes. Don't split for
11378 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
11381 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
11384 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
11386 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
11387 visibility is safe.
11389 2014-02-04 Marek Polacek <polacek@redhat.com>
11391 * gdbinit.in (pel): Define.
11393 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
11395 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
11398 2014-02-04 Richard Biener <rguenther@suse.de>
11401 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
11402 in function context local.
11403 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
11404 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
11405 similar to LTO_imported_decl_ref.
11407 2014-02-04 Jakub Jelinek <jakub@redhat.com>
11409 PR tree-optimization/60002
11410 * cgraphclones.c (build_function_decl_skip_args): Clear
11411 DECL_LANG_SPECIFIC.
11413 PR tree-optimization/60023
11414 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
11415 false to gsi_replace.
11416 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
11417 has been in some EH region and vec_stmt could throw, add
11418 vec_stmt into the same EH region.
11419 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
11420 has no lhs, ignore it.
11421 * internal-fn.c (expand_MASK_LOAD): Likewise.
11424 * tree-inline.c (copy_forbidden): Fail for
11425 __attribute__((optimize (0))) functions.
11428 * omp-low.c (simd_clone_struct_copy): If from->inbranch
11429 is set, copy one less argument.
11430 (expand_simd_clones): Don't subtract clone_info->inbranch
11431 from simd_clone_struct_alloc argument.
11433 PR rtl-optimization/57915
11434 * recog.c (simplify_while_replacing): If all unary/binary/relational
11435 operation arguments are constant, attempt to simplify those.
11437 PR middle-end/59261
11438 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
11439 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
11441 2014-02-04 Richard Biener <rguenther@suse.de>
11443 PR tree-optimization/60012
11444 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
11445 TBAA disambiguation to all DDRs.
11447 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11450 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
11451 (LINK_SPEC): Use it for -shared, -shared-libgcc.
11453 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11456 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
11458 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11460 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
11461 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
11463 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11466 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
11467 to figure out targets of polymorphic calls with known decl.
11468 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
11469 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
11470 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
11471 (get_polymorphic_call_info): ... here.
11472 (get_polymorphic_call_info_from_invariant): New function.
11474 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11476 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
11477 lookup via vtable pointer; check for type consistency
11478 and turn inconsitent facts into UNREACHABLE.
11479 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
11480 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
11481 type inconsistent querries; return UNREACHABLE instead.
11483 2014-02-03 Richard Henderson <rth@twiddle.net>
11486 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
11487 already processed this node.
11488 (normalize_one_pred_1): Pass along mark_set.
11489 (normalize_one_pred): Create and destroy a pointer_set_t.
11490 (normalize_one_pred_chain): Likewise.
11492 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
11494 PR gcov-profile/58602
11495 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
11497 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11500 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
11501 -fno-devirtualize; try to devirtualize by the knowledge of
11502 virtual table pointer given by aggregate propagation.
11503 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
11504 (ipa_print_node_jump_functions): Dump also offset that
11505 is relevant for polymorphic calls.
11506 (determine_known_aggregate_parts): Add arg_type parameter; use it
11507 instead of determining the type from pointer type.
11508 (ipa_compute_jump_functions_for_edge): Update call of
11509 determine_known_aggregate_parts.
11510 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
11511 (gimple_get_virt_method_for_binfo): ... here; simplify using
11512 vtable_pointer_value_to_vtable.
11513 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
11514 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
11515 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
11516 (vtable_pointer_value_to_vtable): Break out from ...; handle also
11518 (vtable_pointer_value_to_binfo): ... here.
11519 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
11521 2014-02-03 Teresa Johnson <tejohnson@google.com>
11523 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
11524 redef of outer loop index variable.
11526 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
11530 * doc/extend.texi (Function Attributes): Typo.
11532 2014-02-03 Cong Hou <congh@google.com>
11534 PR tree-optimization/60000
11535 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
11536 if the vectorized statement is a store. A store statement can only
11537 appear at the end of pattern statements.
11539 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
11541 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
11542 (ix86_option_override_internal): Default long double to 64-bit for
11543 32-bit Bionic and to 128-bit for 64-bit Bionic.
11545 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
11546 TARGET_LONG_DOUBLE_128 is true.
11547 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
11549 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
11550 (mlong-double-64): Negate -mlong-double-128.
11551 (mlong-double-128): New option.
11553 * config/i386/i386-c.c (ix86_target_macros): Define
11554 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
11556 * doc/invoke.texi: Document -mlong-double-128.
11558 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
11560 PR rtl-optimization/60024
11561 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
11563 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
11565 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
11567 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
11569 PR rtl-optimization/57662
11570 * sel-sched.c (code_motion_path_driver): Do not mark already not
11571 existing blocks in the visiting bitmap.
11573 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
11575 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
11576 on the insn being emitted.
11578 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
11579 Will Deacon <will.deacon@arm.com>
11581 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
11583 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11585 * config/arm/arm-tables.opt: Regenerate.
11587 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11589 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
11590 for vector types other than V16QImode.
11591 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
11592 define_expand, and call altivec_expand_vec_perm_le when producing
11593 code with little endian element order.
11594 (*altivec_vperm_<mode>_internal): New insn having previous
11595 behavior of altivec_vperm_<mode>.
11596 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
11597 altivec_expand_vec_perm_le when producing code with little endian
11599 (*altivec_vperm_<mode>_uns_internal): New insn having previous
11600 behavior of altivec_vperm_<mode>_uns.
11602 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11604 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
11605 (altivec_vsumsws): Add handling for -maltivec=be with a little
11607 (altivec_vsumsws_direct): New.
11608 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
11609 gen_altivec_vsumsws.
11611 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
11613 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
11614 vtable_pointer_value_to_binfo): New functions.
11615 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
11616 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
11618 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
11620 * config/nios2/nios2.md (load_got_register): Initialize GOT
11621 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
11622 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
11624 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
11626 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
11627 preserverd by passthrough, do not propagate the type.
11629 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11631 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
11632 (mips_atomic_assign_expand_fenv): New function.
11633 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
11635 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11637 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
11638 (__builtin_mips_set_fcsr): Likewise.
11639 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
11640 MIPS_USI_FTYPE_VOID.
11641 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
11642 (mips16_expand_set_fcsr): Likewise.
11643 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
11644 (mips16_set_fcsr_stub): Likewise.
11645 (mips16_get_fcsr_one_only_stub): New class.
11646 (mips16_set_fcsr_one_only_stub): Likewise.
11647 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
11648 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
11649 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
11650 (hard_float): New availability predicate.
11651 (mips_builtins): Add get_fcsr and set_fcsr.
11652 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
11653 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
11654 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
11655 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
11656 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
11659 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11661 * config/mips/mips.c (mips_one_only_stub): New class.
11662 (mips_need_mips16_rdhwr_p): Replace with...
11663 (mips16_rdhwr_stub): ...this new variable.
11664 (mips16_stub_call_address): New function.
11665 (mips16_rdhwr_one_only_stub): New class.
11666 (mips_expand_thread_pointer): Use mips16_stub_call_address.
11667 (mips_output_mips16_rdhwr): Delete.
11668 (mips_finish_stub): New function.
11669 (mips_code_end): Use it to handle rdhwr stubs.
11671 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
11674 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
11675 when calculating size of integer atomic types.
11677 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
11679 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
11681 2014-02-01 Jakub Jelinek <jakub@redhat.com>
11683 PR tree-optimization/60003
11684 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
11685 * profile.c (branch_prob): Use gimple_call_builtin_p
11686 to check for BUILT_IN_SETJMP_RECEIVER.
11687 * tree-inline.c (copy_bb): Call notice_special_calls.
11689 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
11692 * lra-constraints.c (process_alt_operands): Update reload_sum only
11695 2014-01-31 Richard Henderson <rth@redhat.com>
11697 PR middle-end/60004
11698 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
11699 until after else_eh is processed.
11701 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
11703 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
11704 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
11705 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
11706 in smmintrin.h, remove them.
11707 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
11708 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
11709 * config/i386/i386.md (ROUND_SAE): Fix value.
11710 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
11711 (const48_operand): New.
11712 * config/i386/subst.md (round), (round_expand): Use
11713 const_4_or_8_to_11_operand.
11714 (round_saeonly), (round_saeonly_expand): Use const48_operand.
11716 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
11718 * config/i386/constraints.md (Yk): Swap meaning with k.
11719 * config/i386/i386.md (movhi_internal): Change Yk to k.
11720 (movqi_internal): Ditto.
11721 (*k<logic><mode>): Ditto.
11724 (kandn<mode>): Ditto.
11725 (*<code>hi_1): Ditto.
11726 (*<code>qi_1): Ditto.
11727 (kxnor<mode>): Ditto.
11728 (kortestzhi): Ditto.
11729 (kortestchi): Ditto.
11731 (*one_cmplhi2_1): Ditto.
11732 (*one_cmplqi2_1): Ditto.
11733 * config/i386/sse.md (): Change k to Yk.
11734 (avx512f_load<mode>_mask): Ditto.
11735 (avx512f_blendm<mode>): Ditto.
11736 (avx512f_store<mode>_mask): Ditto.
11737 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
11738 (avx512f_storedqu<mode>_mask): Ditto.
11739 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
11741 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
11742 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
11743 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
11744 (avx512f_maskcmp<mode>3): Ditto.
11745 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
11746 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
11747 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
11748 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
11749 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
11750 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
11751 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
11752 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
11753 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
11754 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
11755 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
11756 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
11757 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
11758 (vec_extract_lo_<mode>_maskm): Ditto.
11759 (vec_extract_hi_<mode>_maskm): Ditto.
11760 (avx512f_vternlog<mode>_mask): Ditto.
11761 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
11762 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
11763 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
11764 (avx512f_<code>v8div16qi2_mask): Ditto.
11765 (avx512f_<code>v8div16qi2_mask_store): Ditto.
11766 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
11767 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
11768 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
11769 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
11770 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
11771 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11772 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
11773 (*avx512pf_scatterpf<mode>df_mask): Ditto.
11774 (avx512cd_maskb_vec_dupv8di): Ditto.
11775 (avx512cd_maskw_vec_dupv16si): Ditto.
11776 (avx512f_vpermi2var<mode>3_maskz): Ditto.
11777 (avx512f_vpermi2var<mode>3_mask): Ditto.
11778 (avx512f_vpermi2var<mode>3_mask): Ditto.
11779 (avx512f_vpermt2var<mode>3_maskz): Ditto.
11780 (*avx512f_gathersi<mode>): Ditto.
11781 (*avx512f_gathersi<mode>_2): Ditto.
11782 (*avx512f_gatherdi<mode>): Ditto.
11783 (*avx512f_gatherdi<mode>_2): Ditto.
11784 (*avx512f_scattersi<mode>): Ditto.
11785 (*avx512f_scatterdi<mode>): Ditto.
11786 (avx512f_compress<mode>_mask): Ditto.
11787 (avx512f_compressstore<mode>_mask): Ditto.
11788 (avx512f_expand<mode>_mask): Ditto.
11789 * config/i386/subst.md (mask): Change k to Yk.
11790 (mask_scalar_merge): Ditto.
11793 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
11795 * doc/extend.texi (Vector Extensions): Document ?: in C++.
11797 2014-01-31 Richard Biener <rguenther@suse.de>
11799 PR middle-end/59990
11800 * builtins.c (fold_builtin_memory_op): Make sure to not
11801 use a floating-point mode or a boolean or enumeral type for
11802 the copy operation.
11804 2014-01-30 DJ Delorie <dj@redhat.com>
11806 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
11807 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
11808 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
11809 whenever main() has an epilogue.
11811 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11813 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
11814 unused variable "field".
11815 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
11816 (vsx_mergeh_<mode>): Likewise.
11817 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
11818 (altivec_vmrghh): Likewise.
11819 (altivec_vmrghw): Likewise.
11820 (altivec_vmrglb): Likewise.
11821 (altivec_vmrglh): Likewise.
11822 (altivec_vmrglw): Likewise.
11823 (altivec_vspltb): Add missing uses.
11824 (altivec_vsplth): Likewise.
11825 (altivec_vspltw): Likewise.
11826 (altivec_vspltsf): Likewise.
11828 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11831 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
11832 frame related instructions.
11834 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
11836 PR rtl-optimization/59959
11837 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
11838 any reload of register whose subreg is invalid.
11840 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11842 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
11843 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
11844 Add missing return type - void.
11846 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11848 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
11849 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
11850 remove element index adjustment for endian (now handled in vsx.md
11852 (altivec_expand_vec_perm_const): Use
11853 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
11854 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
11855 (vsx_xxspltw_<mode>): Adjust element index for little endian.
11856 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
11857 define_expand and a new define_insn *altivec_vspltb_internal;
11858 adjust for -maltivec=be on a little endian target.
11859 (altivec_vspltb_direct): New.
11860 (altivec_vsplth): Divide into a define_expand and a new
11861 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
11862 little endian target.
11863 (altivec_vsplth_direct): New.
11864 (altivec_vspltw): Divide into a define_expand and a new
11865 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
11866 little endian target.
11867 (altivec_vspltw_direct): New.
11868 (altivec_vspltsf): Divide into a define_expand and a new
11869 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
11870 a little endian target.
11872 2014-01-30 Richard Biener <rguenther@suse.de>
11874 PR tree-optimization/59993
11875 * tree-ssa-forwprop.c (associate_pointerplus): Check we
11876 can propagate form the earlier stmt and avoid the transform
11877 when the intermediate result is needed.
11879 2014-01-30 Alangi Derick <alangiderick@gmail.com>
11881 * README.Portability: Fix typo.
11883 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
11885 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
11886 comparison_operator with ordered_comparison_operator.
11888 2014-01-30 Nick Clifton <nickc@redhat.com>
11890 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
11891 Rename to mn10300_store_multiple_regs.
11892 * config/mn10300/mn10300.c: Likewise.
11893 * config/mn10300/mn10300.md (store_movm): Fix typo: call
11894 store_multiple_regs.
11895 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
11896 Call mn10300_store_multiple_regs.
11898 2014-01-30 Nick Clifton <nickc@redhat.com>
11899 DJ Delorie <dj@redhat.com>
11901 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
11902 %fp 2 to keep registers after it properly word-aligned.
11903 (rl78_alloc_physical_registers_umul): Handle the case where both
11904 input operands are the same.
11906 2014-01-30 Richard Biener <rguenther@suse.de>
11908 PR tree-optimization/59903
11909 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
11912 2014-01-30 Jason Merrill <jason@redhat.com>
11915 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
11918 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
11920 2014-01-30 Richard Biener <rguenther@suse.de>
11922 PR tree-optimization/59951
11923 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
11925 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
11928 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
11929 SFmode to DFmode case.
11931 2014-01-29 DJ Delorie <dj@redhat.com>
11933 * config/msp430/msp430.opt (-minrt): New.
11934 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
11936 (ENDFILE_SPEC): Likewise.
11938 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
11940 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
11941 (estimate_function_body_sizes): Use it.
11943 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
11946 * dwarf2out.c (is_cxx_auto): New.
11947 (is_base_type): Use it.
11948 (gen_type_die_with_usage): Likewise.
11950 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11952 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
11953 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
11954 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
11955 -maltivec=be with LE targets.
11956 (vsx_mergeh_<mode>): Likewise.
11957 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
11958 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
11959 (altivec_vmrghb): Replace with define_expand and new
11960 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
11961 (altivec_vmrghb_direct): New define_insn.
11962 (altivec_vmrghh): Replace with define_expand and new
11963 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
11964 (altivec_vmrghh_direct): New define_insn.
11965 (altivec_vmrghw): Replace with define_expand and new
11966 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
11967 (altivec_vmrghw_direct): New define_insn.
11968 (*altivec_vmrghsf): Adjust for endianness.
11969 (altivec_vmrglb): Replace with define_expand and new
11970 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
11971 (altivec_vmrglb_direct): New define_insn.
11972 (altivec_vmrglh): Replace with define_expand and new
11973 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
11974 (altivec_vmrglh_direct): New define_insn.
11975 (altivec_vmrglw): Replace with define_expand and new
11976 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
11977 (altivec_vmrglw_direct): New define_insn.
11978 (*altivec_vmrglsf): Adjust for endianness.
11979 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11980 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11981 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11982 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11983 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11984 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11985 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11986 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11988 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
11990 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
11991 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
11994 2014-01-29 Richard Biener <rguenther@suse.de>
11996 PR tree-optimization/58742
11997 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
11998 associate_pointerplus_align.
11999 (associate_pointerplus_diff): New function.
12000 (associate_pointerplus): Likewise. Call associate_pointerplus_align
12001 and associate_pointerplus_diff.
12003 2014-01-29 Richard Biener <rguenther@suse.de>
12005 * lto-streamer.h (LTO_major_version): Bump to 3.
12006 (LTO_minor_version): Reset to 0.
12008 2014-01-29 Renlin Li <Renlin.Li@arm.com>
12010 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
12011 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
12012 (arm_file_start): Generate correct asm header for armv7ve.
12013 * config/arm/bpabi.h: Add multilib support for armv7ve.
12014 * config/arm/driver-arm.c: Change the architectures of cortex-a7
12015 and cortex-a15 to armv7ve.
12016 * config/arm/t-aprofile: Add multilib support for armv7ve.
12017 * doc/invoke.texi: Document -march=armv7ve.
12019 2014-01-29 Richard Biener <rguenther@suse.de>
12021 PR tree-optimization/58742
12022 * tree-ssa-forwprop.c (associate_plusminus): Return true
12023 if we changed sth, defer EH cleanup to ...
12024 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
12025 (simplify_mult): New function.
12027 2014-01-29 Jakub Jelinek <jakub@redhat.com>
12029 PR middle-end/59917
12030 PR tree-optimization/59920
12031 * tree.c (build_common_builtin_nodes): Remove
12032 __builtin_setjmp_dispatcher initialization.
12033 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
12034 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
12035 instead of gsi_after_labels + manually skipping debug stmts.
12036 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
12037 ignore bbs with IFN_ABNORMAL_DISPATCHER.
12038 * tree-inline.c (copy_edges_for_bb): Remove
12039 can_make_abnormal_goto argument, instead add abnormal_goto_dest
12040 argument. Ignore computed_goto_p stmts. Don't call
12041 make_abnormal_goto_edges. If a call might need abnormal edges
12042 for non-local gotos, see if it already has an edge to
12043 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
12044 with true argument, don't do anything then, otherwise add
12045 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
12046 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
12048 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
12049 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
12050 (lower_stmt): Don't set data->calls_builtin_setjmp.
12051 (lower_builtin_setjmp): Adjust comment.
12052 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
12053 * tree-cfg.c (found_computed_goto): Remove.
12054 (factor_computed_gotos): Remove.
12055 (make_goto_expr_edges): Return bool, true for computed gotos.
12056 Don't call make_abnormal_goto_edges.
12057 (build_gimple_cfg): Don't set found_computed_goto, don't call
12058 factor_computed_gotos.
12059 (computed_goto_p): No longer static.
12060 (make_blocks): Don't set found_computed_goto.
12061 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
12062 (make_edges): If make_goto_expr_edges returns true, push bb
12063 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
12064 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
12065 vector. Record mapping between bbs and OpenMP regions if there
12066 are any, adjust make_gimple_omp_edges caller. Call
12067 handle_abnormal_edges.
12068 (make_abnormal_goto_edges): Remove.
12069 * tree-cfg.h (make_abnormal_goto_edges): Remove.
12070 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
12071 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
12072 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
12073 * internal-fn.def (ABNORMAL_DISPATCHER): New.
12074 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
12075 filling *region also set *region_idx to (*region)->entry->index.
12078 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
12079 For REGs set ORIGINAL_REGNO.
12081 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
12083 * doc/md.texi: Mention that a target shouldn't implement
12084 vec_widen_(s|u)mul_even/odd pair if it is less efficient
12087 2014-01-29 Jakub Jelinek <jakub@redhat.com>
12089 PR tree-optimization/59594
12090 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
12091 a copy of the datarefs vector rather than the vector itself.
12093 2014-01-28 Jason Merrill <jason@redhat.com>
12096 * dwarf2out.c (auto_die): New static.
12097 (gen_type_die_with_usage): Handle C++1y 'auto'.
12098 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
12101 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
12104 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
12105 (SPEC_X32): Likewise.
12106 (SPEC_64): Likewise.
12107 * config/i386/i386.c (ix86_option_override_internal): Turn off
12108 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
12110 (x86_file_start): Output .code16gcc for TARGET_16BIT.
12111 * config/i386/i386.h (TARGET_16BIT): New macro.
12112 (TARGET_16BIT_P): Likewise.
12113 * config/i386/i386.opt: Add m16.
12114 * doc/invoke.texi: Document -m16.
12116 2014-01-28 Jakub Jelinek <jakub@redhat.com>
12118 PR preprocessor/59935
12119 * input.c (location_get_source_line): Bail out on when line number
12120 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
12122 2014-01-28 Richard Biener <rguenther@suse.de>
12124 PR tree-optimization/58742
12125 * tree-ssa-forwprop.c (associate_plusminus): Handle
12126 pointer subtraction of the form (T)(P + A) - (T)P.
12128 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12130 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
12133 2014-01-28 Richard Biener <rguenther@suse.de>
12136 2014-01-28 Richard Biener <rguenther@suse.de>
12138 PR rtl-optimization/45364
12139 PR rtl-optimization/59890
12140 * var-tracking.c (local_get_addr_clear_given_value): Handle
12141 already cleared slot.
12142 (val_reset): Handle not allocated local_get_addr_cache.
12143 (vt_find_locations): Use post-order on the inverted CFG.
12145 2014-01-28 Richard Biener <rguenther@suse.de>
12147 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
12149 2014-01-28 Richard Biener <rguenther@suse.de>
12151 PR rtl-optimization/45364
12152 PR rtl-optimization/59890
12153 * var-tracking.c (local_get_addr_clear_given_value): Handle
12154 already cleared slot.
12155 (val_reset): Handle not allocated local_get_addr_cache.
12156 (vt_find_locations): Use post-order on the inverted CFG.
12158 2014-01-28 Alan Modra <amodra@gmail.com>
12160 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
12161 * configure.ac <recursive call for build != host>: Define
12162 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
12163 and LD_FOR_BUILD too.
12164 * configure: Regenerate.
12166 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
12168 * config/i386/i386.c (get_builtin_code_for_version): Separate
12169 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
12170 Broadwell from Haswell.
12172 2014-01-27 Steve Ellcey <sellcey@mips.com>
12174 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
12175 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
12176 * config/mips/mips.c (mips_option_override): Change setting
12178 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
12179 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
12180 Change from Mask to Var.
12182 2014-01-27 Jeff Law <law@redhat.com>
12184 * ipa-inline.c (inline_small_functions): Fix typo.
12186 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
12188 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
12189 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
12190 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
12191 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
12192 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
12193 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
12194 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
12195 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
12196 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
12197 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
12198 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
12199 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
12200 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
12201 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
12202 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
12203 (_mm512_storeu_epi64): Ditto.
12204 (_mm512_cmpge_epi32_mask): Ditto.
12205 (_mm512_cmpge_epu32_mask): Ditto.
12206 (_mm512_cmpge_epi64_mask): Ditto.
12207 (_mm512_cmpge_epu64_mask): Ditto.
12208 (_mm512_cmple_epi32_mask): Ditto.
12209 (_mm512_cmple_epu32_mask): Ditto.
12210 (_mm512_cmple_epi64_mask): Ditto.
12211 (_mm512_cmple_epu64_mask): Ditto.
12212 (_mm512_cmplt_epi32_mask): Ditto.
12213 (_mm512_cmplt_epu32_mask): Ditto.
12214 (_mm512_cmplt_epi64_mask): Ditto.
12215 (_mm512_cmplt_epu64_mask): Ditto.
12216 (_mm512_cmpneq_epi32_mask): Ditto.
12217 (_mm512_cmpneq_epu32_mask): Ditto.
12218 (_mm512_cmpneq_epi64_mask): Ditto.
12219 (_mm512_cmpneq_epu64_mask): Ditto.
12220 (_mm512_expand_pd): Ditto.
12221 (_mm512_expand_ps): Ditto.
12222 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
12223 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
12224 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
12225 * config/i386/i386.c (ix86_builtins): Add
12226 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
12227 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
12228 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
12229 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
12230 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
12231 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
12232 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
12233 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
12234 IX86_BUILTIN_PMOVUSQW512_MEM.
12235 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
12236 __builtin_ia32_pmovsqd512mem_mask,
12237 __builtin_ia32_pmovqd512mem_mask,
12238 __builtin_ia32_pmovusqw512mem_mask,
12239 __builtin_ia32_pmovsqw512mem_mask,
12240 __builtin_ia32_pmovqw512mem_mask,
12241 __builtin_ia32_pmovusdw512mem_mask,
12242 __builtin_ia32_pmovsdw512mem_mask,
12243 __builtin_ia32_pmovdw512mem_mask,
12244 __builtin_ia32_pmovqb512mem_mask,
12245 __builtin_ia32_pmovusqb512mem_mask,
12246 __builtin_ia32_pmovsqb512mem_mask,
12247 __builtin_ia32_pmovusdb512mem_mask,
12248 __builtin_ia32_pmovsdb512mem_mask,
12249 __builtin_ia32_pmovdb512mem_mask.
12250 (bdesc_args): Add __builtin_ia32_expanddf512,
12251 __builtin_ia32_expandsf512.
12252 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
12253 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
12254 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
12255 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
12256 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
12257 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
12258 (avx512f_<code>v8div16qi2_mask_store): This.
12259 (avx512f_expand<mode>): New.
12261 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
12263 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
12265 (_mm512_mask_prefetch_i64gather_pd): Ditto.
12266 (_mm512_prefetch_i32scatter_pd): Ditto.
12267 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
12268 (_mm512_prefetch_i64scatter_pd): Ditto.
12269 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
12270 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
12271 (_mm512_mask_prefetch_i64gather_ps): Ditto.
12272 (_mm512_prefetch_i32scatter_ps): Ditto.
12273 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
12274 (_mm512_prefetch_i64scatter_ps): Ditto.
12275 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
12276 * config/i386/i386-builtin-types.def: Define
12277 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
12278 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
12279 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
12280 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
12281 IX86_BUILTIN_SCATTERPFQPD.
12282 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
12283 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
12284 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
12285 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
12286 __builtin_ia32_scatterpfqps.
12287 (ix86_expand_builtin): Expand new built-ins.
12288 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
12289 fix memory access data type.
12290 (*avx512pf_gatherpf<mode>_mask): Ditto.
12291 (*avx512pf_gatherpf<mode>): Ditto.
12292 (avx512pf_scatterpf<mode>): Ditto.
12293 (*avx512pf_scatterpf<mode>_mask): Ditto.
12294 (*avx512pf_scatterpf<mode>): Ditto.
12295 (GATHER_SCATTER_SF_MEM_MODE): New.
12296 (avx512pf_gatherpf<mode>df): Ditto.
12297 (*avx512pf_gatherpf<mode>df_mask): Ditto.
12298 (*avx512pf_scatterpf<mode>df): Ditto.
12300 2014-01-27 Jakub Jelinek <jakub@redhat.com>
12303 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
12304 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
12307 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
12309 * common/config/arm/arm-common.c
12310 (arm_rewrite_mcpu): Handle multiple names.
12312 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
12314 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
12316 * gimple-builder.h (create_gimple_tmp): Delete.
12318 2014-01-27 Christian Bruel <christian.bruel@st.com>
12320 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
12323 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
12325 * config/pa/pa.md (call): Generate indirect long calls to non-local
12326 functions when outputing 32-bit code.
12327 (call_value): Likewise except for special call to buggy powf function.
12329 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
12330 portable runtime and PIC indirect calls.
12331 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
12332 and PIC call sequences. Use ldo instead of blr to set return register
12333 in PIC call sequence.
12335 2014-01-25 Walter Lee <walt@tilera.com>
12337 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
12338 avoid clobbering a live register.
12340 2014-01-25 Walter Lee <walt@tilera.com>
12342 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
12343 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
12344 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
12345 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
12347 2014-01-25 Walter Lee <walt@tilera.com>
12349 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
12350 arguments on even registers.
12351 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
12353 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
12354 (BIGGEST_ALIGNMENT): Ditto.
12355 (BIGGEST_FIELD_ALIGNMENT): Ditto.
12357 2014-01-25 Walter Lee <walt@tilera.com>
12359 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
12360 insns before bundling.
12361 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
12363 2014-01-25 Walter Lee <walt@tilera.com>
12365 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
12366 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
12367 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
12369 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
12371 * config/mips/constraints.md (kl): Delete.
12372 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
12373 define expands, using...
12374 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
12375 instructions for MIPS16.
12376 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
12377 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
12379 2014-01-25 Walter Lee <walt@tilera.com>
12381 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
12385 2014-01-25 Walter Lee <walt@tilera.com>
12387 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
12388 (TARGET_EXPAND_TO_RTL_HOOK): Define.
12390 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
12392 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
12395 2014-01-25 Jakub Jelinek <jakub@redhat.com>
12397 * print-rtl.c (in_call_function_usage): New var.
12398 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
12399 EXPR_LIST mode as mode and not as reg note name.
12401 PR middle-end/59561
12402 * cfgloopmanip.c (copy_loop_info): If
12403 loop->warned_aggressive_loop_optimizations, make sure
12404 the flag is set in target loop too.
12406 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
12408 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
12410 * builtins.def: Likewise.
12411 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
12412 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
12413 * ira.c (ira_setup_eliminable_regset): Likewise.
12414 * omp-low.c (gate_expand_omp): Likewise.
12415 (execute_lower_omp): Likewise.
12416 (diagnose_sb_0): Likewise.
12417 (gate_diagnose_omp_blocks): Likewise.
12418 (simd_clone_clauses_extract): Likewise.
12421 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12423 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
12424 correction for little endian...
12425 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
12428 2014-01-24 Jeff Law <law@redhat.com>
12430 PR tree-optimization/59919
12431 * tree-vrp.c (find_assert_locations_1): Do not register asserts
12432 for non-returning calls.
12434 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
12436 * common/config/aarch64/aarch64-common.c
12437 (aarch64_rewrite_mcpu): Handle multiple names.
12438 * config/aarch64/aarch64.h
12439 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
12441 2014-01-24 Dodji Seketeli <dodji@redhat.com>
12443 * input.c (add_file_to_cache_tab): Handle the case where fopen
12446 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
12449 * config/i386/i386.md (pushsf splitter): Get stack adjustment
12450 from push operand if code of push isn't PRE_DEC.
12452 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12455 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
12456 -mquad-memory-atomic. Update -mquad-memory documentation to say
12457 it is only used for non-atomic loads/stores.
12459 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
12460 -mquad-memory or -mquad-memory-atomic switches.
12462 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
12463 -mquad-memory-atomic to ISA 2.07 support.
12465 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
12466 to separate support of normal quad word memory operations (ldq, stq)
12467 from the atomic quad word memory operations.
12469 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
12470 support to separate non-atomic quad word operations from atomic
12471 quad word operations. Disable non-atomic quad word operations in
12472 little endian mode so that we don't have to swap words after the
12473 load and before the store.
12474 (quad_load_store_p): Add comment about atomic quad word support.
12475 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
12476 options printed with -mdebug=reg.
12478 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
12479 -mquad-memory-atomic as the test for whether we have quad word
12480 atomic instructions.
12481 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
12482 or -mp8-vector are used, allow byte/half-word atomic operations.
12484 * config/rs6000/sync.md (load_lockedti): Insure that the address
12485 is a proper indexed or indirect address for the lqarx instruction.
12486 On little endian systems, swap the hi/lo registers after the lqarx
12488 (load_lockedpti): Use indexed_or_indirect_operand predicate to
12489 insure the address is valid for the lqarx instruction.
12490 (store_conditionalti): Insure that the address is a proper indexed
12491 or indirect address for the stqcrx. instruction. On little endian
12492 systems, swap the hi/lo registers before doing the stqcrx.
12494 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
12495 insure the address is valid for the stqcrx. instruction.
12497 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12498 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
12499 type of quad memory support is available.
12501 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
12503 PR regression/59915
12504 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
12505 there is a danger of looping.
12507 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
12509 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
12510 force flag_ira_loop_pressure if set via command line.
12512 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
12514 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
12515 (ashr_simd): New builtin handling DI mode.
12516 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
12517 (aarch64_sshr_simddi): New match pattern.
12518 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
12519 (vshrd_n_s64): Likewise.
12520 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
12522 2014-01-23 Nick Clifton <nickc@redhat.com>
12524 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
12525 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
12526 favour of mcu specific scripts.
12527 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
12530 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
12531 Alex Velenko <Alex.Velenko@arm.com>
12533 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
12534 (vaddv_s16): Likewise.
12535 (vaddv_s32): Likewise.
12536 (vaddv_u8): Likewise.
12537 (vaddv_u16): Likewise.
12538 (vaddv_u32): Likewise.
12539 (vaddvq_s8): Likewise.
12540 (vaddvq_s16): Likewise.
12541 (vaddvq_s32): Likewise.
12542 (vaddvq_s64): Likewise.
12543 (vaddvq_u8): Likewise.
12544 (vaddvq_u16): Likewise.
12545 (vaddvq_u32): Likewise.
12546 (vaddvq_u64): Likewise.
12547 (vaddv_f32): Likewise.
12548 (vaddvq_f32): Likewise.
12549 (vaddvq_f64): Likewise.
12550 (vmaxv_f32): Likewise.
12551 (vmaxv_s8): Likewise.
12552 (vmaxv_s16): Likewise.
12553 (vmaxv_s32): Likewise.
12554 (vmaxv_u8): Likewise.
12555 (vmaxv_u16): Likewise.
12556 (vmaxv_u32): Likewise.
12557 (vmaxvq_f32): Likewise.
12558 (vmaxvq_f64): Likewise.
12559 (vmaxvq_s8): Likewise.
12560 (vmaxvq_s16): Likewise.
12561 (vmaxvq_s32): Likewise.
12562 (vmaxvq_u8): Likewise.
12563 (vmaxvq_u16): Likewise.
12564 (vmaxvq_u32): Likewise.
12565 (vmaxnmv_f32): Likewise.
12566 (vmaxnmvq_f32): Likewise.
12567 (vmaxnmvq_f64): Likewise.
12568 (vminv_f32): Likewise.
12569 (vminv_s8): Likewise.
12570 (vminv_s16): Likewise.
12571 (vminv_s32): Likewise.
12572 (vminv_u8): Likewise.
12573 (vminv_u16): Likewise.
12574 (vminv_u32): Likewise.
12575 (vminvq_f32): Likewise.
12576 (vminvq_f64): Likewise.
12577 (vminvq_s8): Likewise.
12578 (vminvq_s16): Likewise.
12579 (vminvq_s32): Likewise.
12580 (vminvq_u8): Likewise.
12581 (vminvq_u16): Likewise.
12582 (vminvq_u32): Likewise.
12583 (vminnmv_f32): Likewise.
12584 (vminnmvq_f32): Likewise.
12585 (vminnmvq_f64): Likewise.
12587 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
12589 * config/aarch64/aarch64-simd.md
12590 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
12591 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
12592 (*aarch64_mul3_elt<mode>): Likewise.
12593 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
12594 (*aarch64_mul3_elt_to_64v2df): Likewise.
12595 (*aarch64_mla_elt<mode>): Likewise.
12596 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
12597 (*aarch64_mls_elt<mode>): Likewise.
12598 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
12599 (*aarch64_fma4_elt<mode>): Likewise.
12600 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
12601 (*aarch64_fma4_elt_to_64v2df): Likewise.
12602 (*aarch64_fnma4_elt<mode>): Likewise.
12603 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
12604 (*aarch64_fnma4_elt_to_64v2df): Likewise.
12605 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
12606 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
12607 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
12608 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
12609 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
12610 (aarch64_sqdmull_lane<mode>_internal): Likewise.
12611 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
12613 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
12615 * config/aarch64/aarch64-simd.md
12616 (aarch64_be_checked_get_lane<mode>): New define_expand.
12617 * config/aarch64/aarch64-simd-builtins.def
12618 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
12619 New builtin definition.
12620 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
12621 Use new safe be builtin.
12623 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
12625 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
12627 (aarch64_be_st1<mode>): Likewise.
12628 (aarch_ld1<VALL:mode>): Define_expand modified.
12629 (aarch_st1<VALL:mode>): Likewise.
12630 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
12631 (UNSPEC_ST1): Likewise.
12633 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
12635 * config/microblaze/microblaze.md: Add trap insn and attribute
12637 2014-01-23 Dodji Seketeli <dodji@redhat.com>
12639 PR preprocessor/58580
12640 * input.h (location_get_source_line): Take an additional line_size
12642 (void diagnostics_file_cache_fini): Declare new function.
12643 * input.c (struct fcache): New type.
12644 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
12645 New static constants.
12646 (diagnostic_file_cache_init, total_lines_num)
12647 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
12648 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
12649 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
12650 (get_next_line, read_next_line, goto_next_line, read_line_num):
12651 New static function definitions.
12652 (diagnostic_file_cache_fini): New function.
12653 (location_get_source_line): Take an additional output line_len
12654 parameter. Re-write using lookup_or_add_file_to_cache_tab and
12656 * diagnostic.c (diagnostic_finish): Call
12657 diagnostic_file_cache_fini.
12658 (adjust_line): Take an additional input parameter for the length
12659 of the line, rather than calculating it with strlen.
12660 (diagnostic_show_locus): Adjust the use of
12661 location_get_source_line and adjust_line with respect to their new
12662 signature. While displaying a line now, do not stop at the first
12663 null byte. Rather, display the zero byte as a space and keep
12664 going until we reach the size of the line.
12665 * Makefile.in: Add vec.o to OBJS-libcommon
12667 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
12668 Ilya Tocar <ilya.tocar@intel.com>
12670 * config/i386/avx512fintrin.h (_mm512_kmov): New.
12671 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
12672 (__builtin_ia32_kmov16): Ditto.
12673 * config/i386/i386.md (UNSPEC_KMOV): New.
12676 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
12678 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
12679 (_mm512_storeu_si512): Ditto.
12681 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
12684 * rtl.h (get_referenced_operands): Declare.
12685 * recog.c (get_referenced_operands): New function.
12686 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
12687 operands have been referenced when recording LO_SUM references.
12689 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
12691 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
12693 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
12695 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
12696 Enable for generic and recent AMD targets.
12698 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
12700 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
12701 ARG_SIZE note when adjustment was eliminated.
12703 2014-01-22 Jeff Law <law@redhat.com>
12705 PR tree-optimization/59597
12706 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
12707 in file. Accept new argument REGISTERING and use it to modify
12708 dump output appropriately.
12709 (register_jump_thread): Corresponding changes.
12710 (mark_threaded_blocks): Reinstate code to cancel unprofitable
12711 thread paths involving joiner blocks. Add code to dump cancelled
12712 jump threading paths.
12714 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
12716 PR rtl-optimization/59477
12717 * lra-constraints.c (inherit_in_ebb): Process call for living hard
12718 regs. Update reloads_num and potential_reload_hard_regs for all insns.
12720 2014-01-22 Tom Tromey <tromey@redhat.com>
12722 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
12724 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
12726 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
12728 PR rtl-optimization/59896
12729 * lra-constraints.c (process_alt_operands): Check unused note for
12730 matched operands of insn with no output reloads.
12732 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
12734 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
12735 (mips_move_from_gpr_cost): Likewise.
12737 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
12739 PR rtl-optimization/59858
12740 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
12741 ira_class_hard_regs_num.
12742 (process_alt_operands): Increase reject for dying matched operand.
12744 2014-01-21 Jakub Jelinek <jakub@redhat.com>
12747 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
12748 smaller than size, perform several stores or loads and stores
12749 at dst + count - size to store or copy all of size bytes, rather
12750 than just last modesize bytes.
12752 2014-01-20 DJ Delorie <dj@redhat.com>
12754 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
12755 that CLOBBERs are REGs before propogating their values.
12757 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
12759 PR middle-end/59789
12760 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
12761 (cgraph_inline_failed_type): New function.
12762 * cgraph.h (DEFCIFCODE): Add type.
12763 (cgraph_inline_failed_type_t): New enum.
12764 (cgraph_inline_failed_type): New prototype.
12765 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
12766 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
12767 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
12768 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
12769 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
12770 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
12771 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
12772 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
12773 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
12774 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
12775 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
12776 OPTIMIZATION_MISMATCH.
12777 * tree-inline.c (expand_call_inline): Emit errors during
12778 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
12780 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12783 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
12784 mode attribute in insn output.
12786 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
12788 * output.h (output_constant): Delete.
12789 * varasm.c (output_constant): Make private.
12791 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
12793 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
12795 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12797 PR middle-end/59860
12798 * tree.h (fold_builtin_strcat): New prototype.
12799 * builtins.c (fold_builtin_strcat): No longer static. Add len
12800 argument, if non-NULL, don't call c_strlen. Optimize
12801 directly into __builtin_memcpy instead of __builtin_strcpy.
12802 (fold_builtin_2): Adjust fold_builtin_strcat caller.
12803 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
12805 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12807 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12808 for SImode_address_operand operands, having only a REG argument.
12810 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
12812 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
12813 loader name using mbig-endian.
12814 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
12816 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12818 * doc/invoke.texi (-march): Clarify documentation for AArch64.
12819 (-mtune): Likewise.
12822 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
12824 * config/aarch64/aarch64-protos.h
12825 (aarch64_cannot_change_mode_class_ptr): Declare.
12826 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
12827 aarch64_cannot_change_mode_class_ptr): New.
12828 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
12829 backend hook aarch64_cannot_change_mode_class.
12831 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12833 * common/config/aarch64/aarch64-common.c
12834 (aarch64_handle_option): Don't handle any option order logic here.
12835 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
12836 selected_cpu, warn on architecture version mismatch.
12837 (aarch64_override_options): Fix parsing order for option strings.
12839 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12840 Iain Sandoe <iain@codesourcery.com>
12843 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
12844 warning. Amend comment to reflect current functionality.
12846 2014-01-20 Richard Biener <rguenther@suse.de>
12848 PR middle-end/59860
12849 * builtins.c (fold_builtin_strcat): Remove case better handled
12850 by tree-ssa-strlen.c.
12852 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
12854 * config/aarch64/aarch64.opt
12855 (mcpu, march, mtune): Make case-insensitive.
12857 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12860 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12861 if operands[1] is a REG or ZERO_EXTEND of a REG.
12863 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
12865 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
12867 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
12869 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
12870 long non-pic millicode calls.
12872 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12874 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
12876 2014-01-19 Kito Cheng <kito@0xlab.org>
12878 * builtins.c (expand_movstr): Check movstr expand done or fail.
12880 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12881 H.J. Lu <hongjiu.lu@intel.com>
12884 * config/i386/i386.md (*lea<mode>): Zero-extend return register
12885 to DImode for zero-extended addresses.
12887 2014-01-19 Jakub Jelinek <jakub@redhat.com>
12889 PR rtl-optimization/57763
12890 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
12891 on the new indirect jump_insn and increment LABEL_NUSES (label).
12893 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
12897 * config.gcc (x86_archs): New variable.
12898 (x86_64_archs): Likewise.
12899 (x86_cpus): Likewise.
12900 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
12901 --with-arch/--with-cpu= options.
12902 Support --with-arch=/--with-cpu={nehalem,westmere,
12903 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
12905 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12907 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
12908 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
12910 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12912 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
12914 2014-01-18 Jakub Jelinek <jakub@redhat.com>
12917 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
12918 clear cpp_get_options (parse_in)->warn_unused_macros for
12919 ix86_target_macros_internal with cpp_define.
12921 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
12923 * jump.c (delete_related_insns): Keep (use (insn))s.
12924 * reorg.c (redundant_insn): Check for barriers too.
12926 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12928 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
12930 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
12932 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
12933 call to $$dyncall when TARGET_LONG_CALLS is true.
12935 2014-01-17 Jeff Law <law@redhat.com>
12937 * ree.c (combine_set_extension): Temporarily disable test for
12938 changing number of hard registers.
12940 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12942 PR middle-end/58125
12943 * ipa-inline-analysis.c (inline_free_summary):
12944 Do not free summary of aliases.
12946 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12948 PR middle-end/59706
12949 * gimplify.c (gimplify_expr): Use create_tmp_var
12950 instead of create_tmp_var_raw. If cond doesn't have
12951 integral type, don't add the IFN_ANNOTATE builtin at all.
12953 2014-01-17 Martin Jambor <mjambor@suse.cz>
12956 * ipa-cp.c (prev_edge_clone): New variable.
12957 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
12958 Also resize prev_edge_clone vector.
12959 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
12960 (ipcp_edge_removal_hook): New function.
12961 (ipcp_driver): Register ipcp_edge_removal_hook.
12963 2014-01-17 Andrew Pinski <apinski@cavium.com>
12964 Steve Ellcey <sellcey@mips.com>
12967 * config/mips/mips.c (mips_print_operand): Check operand mode instead
12970 2014-01-17 Jeff Law <law@redhat.com>
12972 PR middle-end/57904
12973 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
12974 so that pass_ccp runs first.
12976 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12978 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
12979 (ix86_adjust_cost): Use !TARGET_XXX.
12980 (do_reorder_for_imul): Likewise.
12981 (swap_top_of_ready_list): Likewise.
12982 (ix86_sched_reorder): Likewise.
12984 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12986 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
12987 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
12988 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
12989 (intel_memset): New. Duplicate slm_memset.
12990 (intel_cost): New. Duplicate slm_cost.
12991 (m_INTEL): New macro.
12992 (processor_target_table): Add "intel".
12993 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
12994 with PROCESSOR_INTEL for "intel".
12995 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
12996 PROCESSOR_SILVERMONT.
12997 (ix86_issue_rate): Likewise.
12998 (ix86_adjust_cost): Likewise.
12999 (ia32_multipass_dfa_lookahead): Likewise.
13000 (swap_top_of_ready_list): Likewise.
13001 (ix86_sched_reorder): Likewise.
13002 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
13003 instead of TARGET_OPT_AGU.
13004 * config/i386/i386.h (TARGET_INTEL): New.
13005 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
13006 (processor_type): Add PROCESSOR_INTEL.
13007 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
13008 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
13010 2014-01-17 Marek Polacek <polacek@redhat.com>
13013 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
13016 2014-01-17 Richard Biener <rguenther@suse.de>
13018 PR tree-optimization/46590
13019 * opts.c (default_options_table): Add entries for
13020 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
13021 all enabled at -O1 but not for -Og.
13022 * common.opt (fbranch-count-reg): Remove Init(1).
13023 (fmove-loop-invariants): Likewise.
13024 (ftree-pta): Likewise.
13026 2014-01-17 Jakub Jelinek <jakub@redhat.com>
13028 * config/i386/i386.c (ix86_data_alignment): For compatibility with
13029 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
13030 decls to at least the GCC 4.8 used alignments.
13033 * tree-nested.c (convert_nonlocal_reference_stmt,
13034 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
13035 of GIMPLE_BIND stmts, adjust associated decls.
13037 2014-01-17 Richard Biener <rguenther@suse.de>
13039 PR tree-optimization/46590
13040 * vec.h (vec<>::bseach): New member function implementing
13041 binary search according to C89 bsearch.
13042 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
13043 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
13044 bitmap pointer again. Make accesses_in_loop a flat array.
13045 (mem_ref_obstack): New global.
13046 (outermost_indep_loop): Adjust for mem_ref->stored changes.
13047 (mark_ref_stored): Likewise.
13048 (ref_indep_loop_p_2): Likewise.
13049 (set_ref_stored_in_loop): New helper function.
13050 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
13051 (memref_free): Adjust.
13052 (record_mem_ref_loc): Simplify.
13053 (gather_mem_refs_stmt): Adjust.
13054 (sort_locs_in_loop_postorder_cmp): New function.
13055 (analyze_memory_references): Sort accesses_in_loop after
13056 loop postorder number.
13057 (find_ref_loc_in_loop_cmp): New function.
13058 (for_all_locs_in_loop): Find relevant cluster of locs in
13059 accesses_in_loop and iterate without recursion.
13060 (execute_sm): Avoid uninit warning.
13061 (struct ref_always_accessed): Simplify.
13062 (ref_always_accessed::operator ()): Likewise.
13063 (ref_always_accessed_p): Likewise.
13064 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
13065 loop postorder numbers here.
13066 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
13069 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
13072 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
13073 on decls for which assemble_alias has been called.
13075 2014-01-17 Nick Clifton <nickc@redhat.com>
13077 * config/msp430/msp430.opt: (mcpu): New option.
13078 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
13079 (msp430_option_override): Parse target_cpu. If the MCU name
13080 matches a generic string, clear target_mcu.
13081 (msp430_attr): Allow numeric interrupt values up to 63.
13082 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
13083 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
13085 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
13087 * config/msp430/msp430.md (popm): Use %J rather than %I.
13088 (addsi3): Use msp430_nonimmediate_operand for operand 2.
13089 (addhi_cy_i): Use immediate_operand for operand 2.
13090 * doc/invoke.texi: Document -mcpu option.
13092 2014-01-17 Richard Biener <rguenther@suse.de>
13094 PR rtl-optimization/38518
13095 * df.h (df_analyze_loop): Declare.
13096 * df-core.c: Include cfgloop.h.
13097 (df_analyze_1): Split out main part of df_analyze.
13098 (df_analyze): Adjust.
13099 (loop_inverted_post_order_compute): New function.
13100 (loop_post_order_compute): Likewise.
13101 (df_analyze_loop): New function avoiding whole-function
13102 postorder computes.
13103 * loop-invariant.c (find_defs): Use df_analyze_loop.
13104 (find_invariants): Adjust.
13105 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
13107 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
13109 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
13110 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
13112 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
13114 * ipa-ref.c (ipa_remove_stmt_references): Fix references
13115 traversal when removing references.
13117 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
13120 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
13122 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
13124 PR middle-end/56791
13125 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
13126 pushing a reload for an autoinc when we had previously reloaded an
13127 inner part of the address.
13129 2014-01-16 Jakub Jelinek <jakub@redhat.com>
13131 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
13133 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
13134 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
13135 when not giving up or versioning for alias only because of
13137 (vect_analyze_data_ref_dependences): Set to true.
13138 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
13140 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
13141 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
13144 PR middle-end/58344
13145 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
13148 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
13149 operand 0 predicate for gathers, use a new pseudo as subtarget.
13151 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
13153 PR middle-end/59609
13154 * lra-constraints.c (process_alt_operands): Add printing debug info.
13155 Check absence of input/output reloads for matched operands too.
13157 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
13159 PR rtl-optimization/59835
13160 * ira.c (ira_init_register_move_cost): Increase cost for
13163 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
13165 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
13167 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
13170 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
13171 non-register objects. Use gen_(high/low)part more consistently.
13174 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
13177 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
13178 endian support, remove tests for WORDS_BIG_ENDIAN.
13179 (p8_mfvsrd_3_<mode>): Likewise.
13180 (reload_gpr_from_vsx<mode>): Likewise.
13181 (reload_gpr_from_vsxsf): Likewise.
13182 (p8_mfvsrd_4_disf): Likewise.
13184 2014-01-16 Richard Biener <rguenther@suse.de>
13186 PR rtl-optimization/46590
13187 * lcm.c (compute_antinout_edge): Use postorder iteration.
13188 (compute_laterin): Use inverted postorder iteration.
13190 2014-01-16 Nick Clifton <nickc@redhat.com>
13192 PR middle-end/28865
13193 * varasm.c (output_constant): Return the number of bytes actually
13195 (output_constructor_array_range): Update the field size with the
13196 number of bytes emitted by output_constant.
13197 (output_constructor_regular_field): Likewise. Also do not
13198 complain if the total number of bytes emitted is now greater
13199 than the expected fieldpos.
13200 * output.h (output_constant): Update prototype and descriptive comment.
13202 2014-01-16 Marek Polacek <polacek@redhat.com>
13204 PR middle-end/59827
13205 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
13206 it is error_mark_node.
13208 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
13210 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
13211 VALID_AVX256_REG_OR_OI_MODE.
13213 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
13215 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
13216 current procedure should be profiled.
13218 2014-01-15 Andrew Pinski <apinski@cavium.com>
13220 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
13221 of moving from/to the STACK_REG register class.
13223 2014-01-15 Richard Henderson <rth@redhat.com>
13226 * reginfo.c (global_regs_decl): Globalize.
13227 * rtl.h (global_regs_decl): Declare.
13228 * ira.c (do_reload): Diagnose frame_pointer_needed and it
13229 reserved via global_regs.
13231 2014-01-15 Teresa Johnson <tejohnson@google.com>
13233 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
13235 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13237 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
13238 and vmulosh rather than call gen_vec_widen_smult_*.
13239 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
13240 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
13241 (vec_widen_smult_even_v16qi): Likewise.
13242 (vec_widen_umult_even_v8hi): Likewise.
13243 (vec_widen_smult_even_v8hi): Likewise.
13244 (vec_widen_umult_odd_v16qi): Likewise.
13245 (vec_widen_smult_odd_v16qi): Likewise.
13246 (vec_widen_umult_odd_v8hi): Likewise.
13247 (vec_widen_smult_odd_v8hi): Likewise.
13248 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
13249 vmuloub rather than call gen_vec_widen_umult_*.
13250 (vec_widen_umult_lo_v16qi): Likewise.
13251 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
13252 vmulosb rather than call gen_vec_widen_smult_*.
13253 (vec_widen_smult_lo_v16qi): Likewise.
13254 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
13255 rather than call gen_vec_widen_umult_*.
13256 (vec_widen_umult_lo_v8hi): Likewise.
13257 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
13258 rather than call gen_vec_widen_smult_*.
13259 (vec_widen_smult_lo_v8hi): Likewise.
13261 2014-01-15 Jeff Law <law@redhat.com>
13263 PR tree-optimization/59747
13264 * ree.c (find_and_remove_re): Properly handle case where a second
13265 eliminated extension requires widening a copy created for elimination
13266 of a prior extension.
13267 (combine_set_extension): Ensure that the number of hard regs needed
13268 for a destination register does not change when we widen it.
13270 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
13272 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
13273 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
13274 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
13275 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
13276 (avr-*-rtems*): Likewise.
13277 (bfin*-rtems*): Likewise.
13278 (moxie-*-rtems*): Likewise.
13279 (h8300-*-rtems*): Likewise.
13280 (i[34567]86-*-rtems*): Likewise.
13281 (lm32-*-rtems*): Likewise.
13282 (m32r-*-rtems*): Likewise.
13283 (m68k-*-rtems*): Likewise.
13284 (microblaze*-*-rtems*): Likewise.
13285 (mips*-*-rtems*): Likewise.
13286 (powerpc-*-rtems*): Likewise.
13287 (sh-*-rtems*): Likewise.
13288 (sparc-*-rtems*): Likewise.
13289 (sparc64-*-rtems*): Likewise.
13290 (v850-*-rtems*): Likewise.
13291 (m32c-*-rtems*): Likewise.
13293 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
13295 PR rtl-optimization/59511
13296 * ira.c (ira_init_register_move_cost): Use memory costs for some
13297 cases of register move cost calculations.
13298 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
13299 instead of BB frequency.
13300 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
13301 * lra-assigns.c (find_hard_regno_for): Ditto.
13303 2014-01-15 Richard Biener <rguenther@suse.de>
13305 PR tree-optimization/59822
13306 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
13307 (vectorizable_load): Use it to hoist defs of uses of invariant
13308 loads out of the loop.
13310 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
13311 Kugan Vivekanandarajah <kuganv@linaro.org>
13314 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
13317 2014-01-15 Richard Biener <rguenther@suse.de>
13319 PR rtl-optimization/59802
13320 * lcm.c (compute_available): Use inverted postorder to seed
13321 the initial worklist.
13323 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13326 * config/s390/s390.c (s390_preferred_reload_class): Don't return
13327 ADDR_REGS for invalid symrefs in non-PIC code.
13329 2014-01-15 Jakub Jelinek <jakub@redhat.com>
13332 * builtins.c (determine_block_size): Initialize *probable_max_size
13333 even if len_rtx is CONST_INT.
13335 2014-01-14 Andrew Pinski <apinski@cavium.com>
13337 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
13338 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
13339 (cortexa53_tunings): Likewise.
13340 (aarch64_sched_issue_rate): New function.
13341 (TARGET_SCHED_ISSUE_RATE): Define.
13343 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
13345 * ira-costs.c (find_costs_and_classes): Add missed
13346 ira_init_register_move_cost_if_necessary.
13348 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
13351 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
13353 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
13356 * config/i386/i386.c (type_natural_mode): Add a bool parameter
13357 to indicate if type is used for function return value. Warn ABI
13358 change if the vector mode isn't available for function return value.
13359 (ix86_function_arg_advance): Pass false to type_natural_mode.
13360 (ix86_function_arg): Likewise.
13361 (ix86_gimplify_va_arg): Likewise.
13362 (function_arg_32): Don't warn ABI change.
13363 (ix86_function_value): Pass true to type_natural_mode.
13364 (ix86_return_in_memory): Likewise.
13365 (ix86_struct_value_rtx): Removed.
13366 (TARGET_STRUCT_VALUE_RTX): Likewise.
13368 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13370 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
13371 converting a conditional jump into a conditional return.
13373 2014-01-14 Richard Biener <rguenther@suse.de>
13375 PR tree-optimization/58921
13376 PR tree-optimization/59006
13377 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
13378 hoisting invariant stmts.
13379 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
13380 invariant loads on the preheader edge if possible.
13382 2014-01-14 Joey Ye <joey.ye@arm.com>
13384 * doc/plugin.texi (Building GCC plugins): Update to C++.
13386 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
13388 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
13389 (_mm_rcp28_round_ss): Ditto.
13390 (_mm_rsqrt28_round_sd): Ditto.
13391 (_mm_rsqrt28_round_ss): Ditto.
13392 (_mm_rcp28_sd): Ditto.
13393 (_mm_rcp28_ss): Ditto.
13394 (_mm_rsqrt28_sd): Ditto.
13395 (_mm_rsqrt28_ss): Ditto.
13396 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
13397 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
13398 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
13399 (IX86_BUILTIN_RCP28SD): Ditto.
13400 (IX86_BUILTIN_RCP28SS): Ditto.
13401 (IX86_BUILTIN_RSQRT28SD): Ditto.
13402 (IX86_BUILTIN_RSQRT28SS): Ditto.
13403 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
13404 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
13405 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
13406 (ix86_expand_special_args_builtin): Expand new FTYPE.
13407 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
13408 (srcp14<mode>): Make insn unary.
13409 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
13410 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
13411 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
13412 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
13413 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
13414 Fix rounding: make it SAE only.
13415 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
13417 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
13419 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
13420 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
13421 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
13422 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
13423 (round_saeonly_mask_scalar_operand4): Ditto.
13424 (round_saeonly_mask_scalar_op3): Ditto.
13425 (round_saeonly_mask_scalar_op4): Ditto.
13427 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13429 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
13430 Implement -maltivec=be for vec_insert and vec_extract.
13432 2014-01-10 DJ Delorie <dj@redhat.com>
13434 * config/msp430/msp430.md (call_internal): Don't allow memory
13435 references with SP as the base register.
13436 (call_value_internal): Likewise.
13437 * config/msp430/constraints.md (Yc): New. For memory references
13438 that don't use SP as a base register.
13440 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
13441 "an integer without a # prefix"
13442 * config/msp430/msp430.md (epilogue_helper): Use it.
13444 2014-01-13 Jakub Jelinek <jakub@redhat.com>
13447 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
13448 AVX512F gather builtins.
13449 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
13450 on gather decls with INTEGER_TYPE masktype.
13451 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
13452 directly into the builtin rather than hoisting it before loop.
13454 PR tree-optimization/59387
13455 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
13456 (scev_const_prop): If folded_casts and type has undefined overflow,
13457 use force_gimple_operand instead of force_gimple_operand_gsi and
13458 for each added stmt if it is assign with
13459 arith_code_with_undefined_signed_overflow, call
13460 rewrite_to_defined_overflow.
13461 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
13462 gimple-fold.h instead.
13463 (arith_code_with_undefined_signed_overflow,
13464 rewrite_to_defined_overflow): Moved to ...
13465 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
13466 rewrite_to_defined_overflow): ... here. No longer static.
13467 Include gimplify-me.h.
13468 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
13469 rewrite_to_defined_overflow): New prototypes.
13471 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13473 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
13475 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
13477 * builtins.c (get_object_alignment_2): Minor tweak.
13478 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
13480 2014-01-13 Christian Bruel <christian.bruel@st.com>
13482 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
13483 optimized non constant lengths.
13485 2014-01-13 Jakub Jelinek <jakub@redhat.com>
13488 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
13489 load as __atomic_load_N if possible.
13491 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
13493 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
13495 (rs6000_expand_builtin): Adjust call.
13497 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
13500 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
13501 * config/rs6000/rs6000.c: Include target-globals.h.
13502 (rs6000_set_current_function): Instead of doing target_reinit
13503 unconditionally, use save_target_globals_default_opts and
13504 restore_target_globals.
13506 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
13508 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
13509 (rs6000_expand_builtin): Handle mffs and mtfsf.
13510 (rs6000_init_builtins): Define mffs and mtfsf.
13511 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
13512 (rs6000_mffs): New pattern.
13513 (rs6000_mtfsf): New pattern.
13515 2014-01-11 Bin Cheng <bin.cheng@arm.com>
13517 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
13518 Start narrowing with START. Apply candidate-use pair
13519 and check overall cost in narrowing.
13520 (iv_ca_prune): Pass new argument.
13522 2014-01-10 Jeff Law <law@redhat.com>
13524 PR middle-end/59743
13525 * ree.c (combine_reaching_defs): Ensure the defining statement
13526 occurs before the extension when optimizing extensions with
13527 different source and destination hard registers.
13529 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
13532 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
13533 vtables into the type inheritance graph.
13535 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13537 PR rtl-optimization/59754
13538 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
13539 modes in the REGNO != REGNO case.
13541 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13543 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
13545 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13547 PR tree-optimization/59745
13548 * tree-predcom.c (tree_predictive_commoning_loop): Call
13549 free_affine_expand_cache if giving up because components is NULL.
13551 * target-globals.c (save_target_globals): Allocate < 4KB structs using
13552 GC in payload of target_globals struct instead of allocating them on
13553 the heap and the larger structs separately using GC.
13554 * target-globals.h (struct target_globals): Make regs, hard_regs,
13555 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
13556 of GTY((skip)) and change type to void *.
13557 (reset_target_globals): Cast loads from those fields to corresponding
13560 2014-01-10 Steve Ellcey <sellcey@mips.com>
13563 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
13564 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
13565 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
13567 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
13570 * aarch64-modes.def (CC_Zmode): New flags mode.
13571 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
13572 represents an equality.
13573 (aarch64_get_condition_code): Handle CC_Zmode.
13574 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
13576 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13578 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
13579 extraction in good case.
13581 2014-01-10 Richard Biener <rguenther@suse.de>
13583 PR tree-optimization/59374
13584 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
13585 checking after SLP discovery. Mark stmts not participating
13586 in any SLP instance properly.
13588 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13590 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
13591 when handling a SET rtx.
13593 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13595 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
13596 (cortex-a57): Likewise.
13597 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
13599 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13601 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
13602 non-iwmmxt builtins.
13604 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
13608 * ipa-devirt.c record_target_from_binfo): Take as argument
13609 stack of binfos and lookup matching one for virtual inheritance.
13610 (possible_polymorphic_call_targets_1): Update.
13612 2014-01-10 Huacai Chen <chenhc@lemote.com>
13614 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
13615 kernel strings for Loongson-2E/2F/3A.
13617 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13619 PR middle-end/59670
13620 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
13621 is_gimple_call before calling gimple_call_internal_p.
13623 2014-01-09 Steve Ellcey <sellcey@mips.com>
13625 * Makefile.in (TREE_FLOW_H): Remove.
13626 (TREE_SSA_H): Add file names from tree-flow.h.
13627 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
13628 * tree.h: Remove tree-flow.h reference.
13629 * hash-table.h: Remove tree-flow.h reference.
13630 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
13631 reference with tree-ssa-loop.h.
13633 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13635 * doc/invoke.texi: Add -maltivec={be,le} options, and document
13636 default element-order behavior for -maltivec.
13637 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
13638 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
13639 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
13640 when targeting big endian, at least for now.
13641 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
13643 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13645 PR middle-end/47735
13646 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
13647 var satisfies use_register_for_decl, just take into account type
13648 alignment, rather than decl alignment.
13650 PR tree-optimization/59622
13651 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
13652 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
13653 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
13654 Don't devirtualize for inplace at all. For targets.length () == 1,
13655 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
13657 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
13659 * config/i386/i386.md (cpu): Remove the unused btver1.
13661 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
13663 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
13665 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13668 * tree-core.h (struct target_globals): New forward declaration.
13669 (struct tree_target_option): Add globals field.
13670 * tree.h (TREE_TARGET_GLOBALS): Define.
13671 (prepare_target_option_nodes_for_pch): New prototype.
13672 * target-globals.h (struct target_globals): Define even if
13673 !SWITCHABLE_TARGET.
13674 * tree.c (prepare_target_option_node_for_pch,
13675 prepare_target_option_nodes_for_pch): New functions.
13676 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
13677 * config/i386/i386.c: Include target-globals.h.
13678 (ix86_set_current_function): Instead of doing target_reinit
13679 unconditionally, use save_target_globals_default_opts and
13680 restore_target_globals.
13682 2014-01-09 Richard Biener <rguenther@suse.de>
13684 PR tree-optimization/59715
13685 * tree-cfg.h (split_critical_edges): Declare.
13686 * tree-cfg.c (split_critical_edges): Export.
13687 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
13689 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
13691 * cfgexpand.c (expand_stack_vars): Optionally disable
13692 asan stack protection.
13693 (expand_used_vars): Likewise.
13694 (partition_stack_vars): Likewise.
13695 * asan.c (asan_emit_stack_protection): Optionally disable
13696 after return stack usage.
13697 (instrument_derefs): Optionally disable memory access instrumentation.
13698 (instrument_builtin_call): Likewise.
13699 (instrument_strlen_call): Likewise.
13700 (asan_protect_global): Optionally disable global variables protection.
13701 * doc/invoke.texi: Added doc for new options.
13702 * params.def: Added new options.
13703 * params.h: Likewise.
13705 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13707 PR rtl-optimization/59724
13708 * ifcvt.c (cond_exec_process_if_block): Don't call
13709 flow_find_head_matching_sequence with 0 longest_match.
13710 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
13711 non-active insns if !stop_after.
13712 (try_head_merge_bb): Revert 2014-01-07 changes.
13714 2014-01-08 Jeff Law <law@redhat.com>
13716 * ree.c (get_sub_rtx): New function, extracted from...
13717 (merge_def_and_ext): Here.
13718 (combine_reaching_defs): Use get_sub_rtx.
13720 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
13722 * cgraph.h (varpool_variable_node): Do not choke on null node.
13724 2014-01-08 Catherine Moore <clm@codesourcery.com>
13726 * config/mips/mips.md (simple_return): Attempt to use JRC
13728 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
13730 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13732 PR rtl-optimization/59137
13733 * reorg.c (steal_delay_list_from_target): Call update_block for
13735 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
13737 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13739 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
13740 two duplicate entries.
13742 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13745 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
13747 * config/mips/mips.c (mips_truncated_op_cost): New function.
13748 (mips_rtx_costs): Adjust test for BADDU.
13749 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
13751 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
13753 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
13754 (*baddu_si): ...this new pattern.
13756 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13759 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
13761 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
13763 PR middle-end/57748
13764 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
13766 (expand_expr, expand_normal): Adjust.
13767 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
13768 inner_reference_p. Use inner_reference_p to expand inner references.
13769 (store_expr): Adjust.
13770 * cfgexpand.c (expand_call_stmt): Adjust.
13772 2014-01-08 Rong Xu <xur@google.com>
13774 * gcov-io.c (gcov_var): Move from gcov-io.h.
13775 (gcov_position): Ditto.
13776 (gcov_is_error): Ditto.
13777 (gcov_rewrite): Ditto.
13778 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
13779 only part to libgcc/libgcov.h.
13781 2014-01-08 Marek Polacek <polacek@redhat.com>
13783 PR middle-end/59669
13784 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
13786 2014-01-08 Marek Polacek <polacek@redhat.com>
13789 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
13791 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13793 PR rtl-optimization/59649
13794 * stor-layout.c (get_mode_bounds): For BImode return
13795 0 and STORE_FLAG_VALUE.
13797 2014-01-08 Richard Biener <rguenther@suse.de>
13799 PR middle-end/59630
13800 * gimple.h (is_gimple_builtin_call): Remove.
13801 (gimple_builtin_call_types_compatible_p): New.
13802 (gimple_call_builtin_p): New overload.
13803 * gimple.c (is_gimple_builtin_call): Remove.
13804 (validate_call): Rename to ...
13805 (gimple_builtin_call_types_compatible_p): ... this and export. Also
13806 check return types.
13807 (validate_type): New static function.
13808 (gimple_call_builtin_p): New overload and adjust.
13809 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
13810 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
13811 (gimple_fold_stmt_to_constant_1): Likewise.
13812 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
13814 2014-01-08 Richard Biener <rguenther@suse.de>
13816 PR middle-end/59471
13817 * gimplify.c (gimplify_expr): Gimplify register-register type
13818 VIEW_CONVERT_EXPRs to separate stmts.
13820 2014-01-07 Jeff Law <law@redhat.com>
13822 PR middle-end/53623
13823 * ree.c (combine_set_extension): Handle case where source
13824 and destination registers in an extension insn are different.
13825 (combine_reaching_defs): Allow source and destination registers
13826 in extension to be different under limited circumstances.
13827 (add_removable_extension): Remove restriction that the
13828 source and destination registers in the extension are the same.
13829 (find_and_remove_re): Emit a copy from the extension's
13830 destination to its source after the defining insn if
13831 the source and destination registers are different.
13833 PR middle-end/59285
13834 * ifcvt.c (merge_if_block): If we are merging a block with more than
13835 one successor with a block with no successors, remove any BARRIER
13836 after the second block.
13838 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
13840 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
13842 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
13845 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
13846 for 14-bit register offsets when INT14_OK_STRICT is false.
13848 2014-01-07 Roland Stigge <stigge@antcom.de>
13849 Michael Meissner <meissner@linux.vnet.ibm.com>
13852 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
13853 Only check TFmode for SPE constants. Don't check TImode or TDmode.
13855 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
13857 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
13860 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
13862 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
13863 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
13864 rtx is const0_rtx or not.
13866 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
13869 * target-globals.c (save_target_globals): Remove this_fn_optab
13871 * toplev.c: Include optabs.h.
13872 (target_reinit): Temporarily restore the global options if another
13873 set of options are in force.
13875 2014-01-07 Jakub Jelinek <jakub@redhat.com>
13877 PR rtl-optimization/58668
13878 * cfgcleanup.c (flow_find_cross_jump): Don't count
13879 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
13880 to determine what is counted.
13881 (flow_find_head_matching_sequence): Use active_insn_p to determine
13883 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
13885 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
13886 determine what is counted.
13888 PR tree-optimization/59643
13889 * tree-predcom.c (split_data_refs_to_components): If one dr is
13890 read and one write, determine_offset fails and the write isn't
13891 in the bad component, just put the read into the bad component.
13893 2014-01-07 Mike Stump <mikestump@comcast.net>
13894 Jakub Jelinek <jakub@redhat.com>
13897 * tree-core.h (struct tree_optimization_option): Change optabs
13898 type from unsigned char * to void *.
13899 * optabs.c (init_tree_optimization_optabs): Adjust
13900 TREE_OPTIMIZATION_OPTABS initialization.
13902 2014-01-06 Jakub Jelinek <jakub@redhat.com>
13905 * config/i386/i386.h (struct machine_function): Add
13906 no_drap_save_restore field.
13907 * config/i386/i386.c (ix86_save_reg): Use
13908 !cfun->machine->no_drap_save_restore instead of
13909 crtl->stack_realign_needed.
13910 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
13911 this function clears frame_pointer_needed. Set
13912 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
13913 and DRAP reg is needed.
13915 2014-01-06 Marek Polacek <polacek@redhat.com>
13918 * doc/implement-c.texi: Mention that other integer types are
13919 permitted as bit-field types in strictly conforming mode.
13921 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
13923 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
13924 is newly allocated.
13926 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
13928 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
13930 2014-01-06 Martin Jambor <mjambor@suse.cz>
13933 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
13935 * ipa-prop.c (ipa_print_node_params): Fix indentation.
13937 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
13941 * var-tracking.c (add_stores): Preserve the value of the source even if
13942 we don't record the store.
13944 2014-01-06 Terry Guo <terry.guo@arm.com>
13946 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
13948 2014-01-05 Iain Sandoe <iain@codesourcery.com>
13951 * config/darwin.c (darwin_function_section): Adjust return values to
13952 correspond to optimisation changes made in r206070.
13954 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
13956 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
13957 from prefetch_block tune setting.
13958 (nocona_cost): Correct size of prefetch block to 64.
13960 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
13962 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
13963 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
13964 used to save the static chain register in the computation of the offset
13965 from which the FP registers need to be restored.
13967 2014-01-04 Jakub Jelinek <jakub@redhat.com>
13969 PR tree-optimization/59519
13970 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
13971 ICE if get_current_def (current_new_name) is already non-NULL, as long
13972 as it is a phi result of some other phi in *new_exit_bb that has
13975 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
13976 or vmovdqu* for misaligned_operand.
13977 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
13978 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
13979 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
13980 aligned_mem for AVX512F masked aligned load and store builtins and for
13981 non-temporal moves.
13983 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
13985 PR tree-optimization/59651
13986 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13987 Address range for negative step should be added by TYPE_SIZE_UNIT.
13989 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
13991 * config/m68k/m68k.c (handle_move_double): Handle pushes with
13992 overlapping registers also for registers other than the stack pointer.
13994 2014-01-03 Marek Polacek <polacek@redhat.com>
13997 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
14000 2014-01-03 Jakub Jelinek <jakub@redhat.com>
14003 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
14006 * config/i386/i386.md (MODE_SIZE): New mode attribute.
14007 (push splitter): Use <P:MODE_SIZE> instead of
14008 GET_MODE_SIZE (<P:MODE>mode).
14009 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
14010 (mov -1, reg peephole2): Likewise.
14011 * config/i386/sse.md (*mov<mode>_internal,
14012 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
14013 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
14014 *<code><mode>3, *andnot<mode>3<mask_name>,
14015 <mask_codefor><code><mode>3<mask_name>): Likewise.
14016 * config/i386/subst.md (mask_mode512bit_condition,
14017 sd_mask_mode512bit_condition): Likewise.
14019 2014-01-02 Xinliang David Li <davidxl@google.com>
14021 PR tree-optimization/59303
14022 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
14023 (dump_predicates): Better output format.
14024 (pred_equal_p): New function.
14025 (is_neq_relop_p): Ditto.
14026 (is_neq_zero_form_p): Ditto.
14027 (pred_expr_equal_p): Ditto.
14028 (pred_neg_p): Ditto.
14029 (simplify_pred): Ditto.
14030 (simplify_preds_2): Ditto.
14031 (simplify_preds_3): Ditto.
14032 (simplify_preds_4): Ditto.
14033 (simplify_preds): Ditto.
14034 (push_pred): Ditto.
14035 (push_to_worklist): Ditto.
14036 (get_pred_info_from_cmp): Ditto.
14037 (is_degenerated_phi): Ditto.
14038 (normalize_one_pred_1): Ditto.
14039 (normalize_one_pred): Ditto.
14040 (normalize_one_pred_chain): Ditto.
14041 (normalize_preds): Ditto.
14042 (normalize_cond_1): Remove function.
14043 (normalize_cond): Ditto.
14044 (is_gcond_subset_of): Ditto.
14045 (is_subset_of_any): Ditto.
14046 (is_or_set_subset_of): Ditto.
14047 (is_and_set_subset_of): Ditto.
14048 (is_norm_cond_subset_of): Ditto.
14049 (pred_chain_length_cmp): Ditto.
14050 (convert_control_dep_chain_into_preds): Type change.
14051 (find_predicates): Ditto.
14052 (find_def_preds): Ditto.
14053 (destroy_predicates_vecs): Ditto.
14054 (find_matching_predicates_in_rest_chains): Ditto.
14055 (use_pred_not_overlap_with_undef_path_pred): Ditto.
14056 (is_pred_expr_subset): Ditto.
14057 (is_pred_chain_subset_of): Ditto.
14058 (is_included_in): Ditto.
14059 (is_superset_of): Ditto.
14061 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
14063 Update copyright years.
14065 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
14067 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
14068 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
14069 config/arc/arc.md, config/arc/arc.opt,
14070 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
14071 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
14072 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
14073 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
14074 config/linux-protos.h, config/linux.c, config/winnt-c.c,
14075 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
14076 vtable-verify.c, vtable-verify.h: Use the standard form for the
14079 2014-01-02 Tobias Burnus <burnus@net-b.de>
14081 * gcc.c (process_command): Update copyright notice dates.
14082 * gcov-dump.c: Ditto.
14084 * doc/cpp.texi: Bump @copying's copyright year.
14085 * doc/cppinternals.texi: Ditto.
14086 * doc/gcc.texi: Ditto.
14087 * doc/gccint.texi: Ditto.
14088 * doc/gcov.texi: Ditto.
14089 * doc/install.texi: Ditto.
14090 * doc/invoke.texi: Ditto.
14092 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14094 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
14096 2014-01-01 Jakub Jelinek <jakub@redhat.com>
14098 * config/i386/sse.md (*mov<mode>_internal): Guard
14099 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
14101 PR rtl-optimization/59647
14102 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
14103 new_rtx into UNSIGNED_FLOAT rtxes.
14105 Copyright (C) 2014 Free Software Foundation, Inc.
14107 Copying and distribution of this file, with or without modification,
14108 are permitted in any medium without royalty provided the copyright
14109 notice and this notice are preserved.