1 2014-06-05 Marek Polacek <polacek@redhat.com>
4 * doc/invoke.texi: Document -Wlogical-not-parentheses.
6 2014-06-04 Tom de Vries <tom@codesourcery.com>
8 * config/s390/s390.md ("addptrdi3", "addptrsi3"): Use INTVAL only on
11 2014-06-04 Marc Glisse <marc.glisse@inria.fr>
13 PR tree-optimization/61385
14 * tree-ssa-phiopt.c (value_replacement): Punt if there are PHI nodes.
16 2014-06-04 Bernd Schmidt <bernds@codesourcery.com>
18 * lto-wrapper.c (fatal, fatal_perror): Remove functions. All callers
19 changed to use fatal_error.
20 (main): Ensure lto_wrapper_cleanup is run atexit.
22 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
24 * lra-constraints.c (valid_address_p): Move earlier in file.
25 (address_eliminator): New structure.
26 (satisfies_memory_constraint_p): New function.
27 (satisfies_address_constraint_p): Likewise.
28 (process_alt_operands, process_address, curr_insn_transform): Use them.
30 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
32 * lra-int.h (lra_static_insn_data): Make operand_alternative a
34 (target_lra_int, default_target_lra_int, this_target_lra_int)
35 (op_alt_data): Delete.
36 * lra.h (lra_init): Delete.
37 * lra.c (default_target_lra_int, this_target_lra_int): Delete.
38 (init_insn_code_data_once): Remove op_alt_data handling.
39 (finish_insn_code_data_once): Likewise.
40 (init_op_alt_data): Delete.
41 (get_static_insn_data): Initialize operand_alternative to null.
42 (free_insn_recog_data): Cast operand_alternative before freeing it.
43 (setup_operand_alternative): Take the operand_alternative as
44 parameter and assume it isn't already cached in the static
46 (lra_set_insn_recog_data): Update accordingly.
48 * ira.c (ira_init): Don't call lra_init.
49 * target-globals.h (this_target_lra_int): Declare.
50 (target_globals): Remove lra_int.
51 (restore_target_globals): Update accordingly.
52 * target-globals.c: Don't include lra-int.h.
53 (default_target_globals, save_target_globals): Remove lra_int.
55 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
57 * recog.h (operand_alternative): Convert reg_class, reject,
58 matched and matches into bitfields.
59 (preprocess_constraints): New overload.
60 (preprocess_insn_constraints): New function.
61 (preprocess_constraints): Take the insn as parameter.
62 (recog_op_alt): Change into a pointer.
63 (target_recog): Add x_op_alt.
64 * recog.c (asm_op_alt): New variable.
65 (recog_op_alt): Change into a pointer.
66 (preprocess_constraints): New overload, replacing the old function
67 definition with one that doesn't use global state.
68 (preprocess_insn_constraints): New function.
69 (preprocess_constraints): Use them. Take the insn as parameter.
70 Use asm_op_alt for asms.
71 (recog_init): Free existing x_op_alt entries.
72 * ira-lives.c (check_and_make_def_conflict): Make operand_alternative
74 (make_early_clobber_and_input_conflicts): Likewise.
75 (process_bb_node_lives): Pass the insn to process_constraints.
76 * reg-stack.c (check_asm_stack_operands): Likewise.
77 (subst_asm_stack_regs): Likewise.
78 * regcprop.c (copyprop_hardreg_forward_1): Likewise.
79 * regrename.c (build_def_use): Likewise.
80 * sched-deps.c (sched_analyze_insn): Likewise.
81 * sel-sched.c (get_reg_class, implicit_clobber_conflict_p): Likewise.
82 * config/arm/arm.c (xscale_sched_adjust_cost): Likewise.
83 (note_invalid_constants): Likewise.
84 * config/i386/i386.c (ix86_legitimate_combined_insn): Likewise.
85 (ix86_legitimate_combined_insn): Make operand_alternative pointer
88 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
90 * recog.c (preprocess_constraints): Don't skip disabled alternatives.
91 * ira-lives.c (check_and_make_def_conflict): Check for disabled
93 (make_early_clobber_and_input_conflicts): Likewise.
94 * config/i386/i386.c (ix86_legitimate_combined_insn): Likewise.
96 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
98 * recog.h (alternative_class): New function.
99 (which_op_alt): Return a const recog_op_alt.
100 * reg-stack.c (check_asm_stack_operands): Update type accordingly.
101 (subst_asm_stack_regs): Likewise.
102 * config/arm/arm.c (note_invalid_constants): Likewise.
103 * regcprop.c (copyprop_hardreg_forward_1): Likewise. Don't modify
104 the operand_alternative; use alternative class instead.
105 * sel-sched.c (get_reg_class): Likewise.
106 * regrename.c (build_def_use): Likewise.
107 (hide_operands, restore_operands, record_out_operands): Update type
110 2014-06-04 Richard Sandiford <rdsandiford@googlemail.com>
112 * recog.h (recog_op_alt): Convert to a flat array.
113 (which_op_alt): New function.
114 * recog.c (recog_op_alt): Convert to a flat array.
115 (preprocess_constraints): Update accordingly, grouping all
116 operands of the same alternative together, rather than the
118 * ira-lives.c (check_and_make_def_conflict): Likewise.
119 (make_early_clobber_and_input_conflicts): Likewise.
120 * config/i386/i386.c (ix86_legitimate_combined_insn): Likewise.
121 * reg-stack.c (check_asm_stack_operands): Use which_op_alt.
122 (subst_asm_stack_regs): Likewise.
123 * regcprop.c (copyprop_hardreg_forward_1): Likewise.
124 * regrename.c (hide_operands, record_out_operands): Likewise.
125 (build_def_use): Likewise.
126 * sel-sched.c (get_reg_class): Likewise.
127 * config/arm/arm.c (note_invalid_constants): Likewise.
129 2014-06-04 Jason Merrill <jason@redhat.com>
133 * gimplify.c (gimplify_arg): Non-static.
134 * gimplify.h: Declare it.
136 2014-06-04 Richard Biener <rguenther@suse.de>
138 * tree.h (may_be_aliased): Trust TREE_ADDRESSABLE from
139 TREE_PUBLIC and DECL_EXTERNAL decls.
141 2014-06-04 Matthew Fortune <matthew.fortune@imgtec.com>
143 * regcprop.c (copyprop_hardreg_forward_1): Account for
144 HARD_REGNO_CALL_PART_CLOBBERED.
146 2014-06-04 Richard Biener <rguenther@suse.de>
148 * configure.ac: Check whether the underlying type of int64_t
149 is long or long long.
150 * configure: Regenerate.
151 * config.in: Likewise.
152 * hwint.h (HOST_WIDE_INT): Match the underlying type of int64_t.
153 (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
155 2014-06-04 Richard Biener <rguenther@suse.de>
157 PR tree-optimization/60098
158 * tree-ssa-dse.c (dse_possible_dead_store_p): Walk until
160 (dse_optimize_stmt): Simplify, now that we found a kill
163 2014-06-04 Richard Biener <rguenther@suse.de>
165 * tree-ssa-alias.c (stmt_may_clobber_ref_p): Improve handling
166 of accesses with non-invariant address.
168 2014-06-04 Martin Liska <mliska@suse.cz>
170 * cgraph.h (cgraph_make_wrapper): New function introduced.
171 * cgraphunit.c (cgraph_make_wrapper): The function implementation.
172 * ipa-inline.h (inline_analyze_function): The function is global.
173 * ipa-inline-analysis.c (inline_analyze_function): Likewise.
175 2014-06-04 Martin Liska <mliska@suse.cz>
177 * tree.h (private_lookup_attribute_starting): New function.
178 (lookup_attribute_starting): Likewise.
179 * tree.c (private_lookup_attribute_starting): Likewise.
181 2014-06-04 Martin Liska <mliska@suse.cz>
183 * cgraph.h (expand_thunk): New argument added.
184 (address_taken_from_non_vtable_p): New global function.
185 * ipa-visibility.c (address_taken_from_non_vtable_p): Likewise.
186 * cgraphclones.c (duplicate_thunk_for_node): Argument added to call.
187 * cgraphunit.c (analyze_function): Likewise.
188 (assemble_thunks_and_aliases): Argument added to call.
189 (expand_thunk): New argument forces to produce GIMPLE thunk.
191 2014-06-04 Martin Liska <mliska@suse.cz>
193 * coverage.h (coverage_compute_cfg_checksum): Argument added.
194 * coverage.c (coverage_compute_cfg_checksum): Likewise.
195 * profile.c (branch_prob): Likewise.
197 2014-06-04 Martin Jambor <mjambor@suse.cz>
200 * ipa-pure-const.c (propagate_pure_const): Add unreachable default
201 handler for switch on an ipa_ref_use enum.
202 * ipa-reference.c (analyze_function): Likewise.
204 2014-06-04 Kai Tietz <ktietz@redhat.com>
206 * recog.c (peep2_attempt): Copy SIBLING_CALL_P flag
207 from old call-instruction.
209 2014-06-04 Bin Cheng <bin.cheng@arm.com>
211 * config/aarch64/aarch64.c (aarch64_classify_address)
212 (aarch64_legitimize_reload_address): Support full addressing modes
214 * config/aarch64/aarch64.md (mov<mode>, movmisalign<mode>)
215 (*aarch64_simd_mov<mode>, *aarch64_simd_mov<mode>): Relax predicates.
217 2014-06-03 Andrew Pinski <apinski@cavium.com>
219 * config/aarch64/aarch64.c (aarch64_if_then_else_costs): Allow non comparisons
222 2014-06-03 Andrew Pinski <apinski@cavium.com>
224 * config/aarch64/aarch64.c (aarch64_if_then_else_costs): New function.
225 (aarch64_rtx_costs): Use aarch64_if_then_else_costs.
227 2014-06-03 Kai Tietz <ktietz@redhat.com>
229 * config/i386/i386.c (ix86_function_value_regno_p): Disallow DX_REG
232 2014-06-03 Dehao Chen <dehao@google.com>
234 * tree-cfg.c (gimple_merge_blocks): Only reset count when BBs are in
237 2014-06-03 Marek Polacek <polacek@redhat.com>
240 * doc/invoke.texi: Document -Wswitch-bool.
241 * function.c (stack_protect_epilogue): Cast controlling expression of
243 * gengtype.c (walk_type): Generate switch expression with its
244 controlling expression cast to int.
246 2014-06-03 Vishnu K S <Vishnu.k_s@atmel.com>
248 * config/avr/avr-mcus.def: Add new avr25 devices attiny441, attiny828
250 * config/avr/avr-tables.opt: Regenerate.
251 * config/avr/t-multilib: Regenerate.
252 * doc/avr-mmcu.texi: Regenerate.
254 2014-06-03 Vishnu K S <vishnu.k_s@atmel.com>
255 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
257 * config/avr/avr-mcus.def (ata6616c): Add new avr25 device.
258 (ata6617c, ata664251): Add new avr35 devices.
259 (ata6612c): Add new avr4 device.
260 (ata6613c, ata6614q): Add new avr5 devices.
261 * config/avr/avr-tables.opt: Regenerate.
262 * config/avr/t-multilib: Regenerate.
263 * doc/avr-mmcu.texi: Regenerate.
265 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
267 * gcc/config/aarch64/aarch64-builtins.c
268 (aarch64_types_binop_ssu_qualifiers): New static data.
269 (TYPES_BINOP_SSU): Define.
270 * gcc/config/aarch64/aarch64-simd-builtins.def (suqadd, ushl, urshl,
271 urshr_n, ushll_n): Use appropriate unsigned qualifiers.
272 * gcc/config/aarch64/arm_neon.h (vrshl_u8, vrshl_u16, vrshl_u32,
273 vrshl_u64, vrshlq_u8, vrshlq_u16, vrshlq_u32, vrshlq_u64, vrshld_u64,
274 vrshr_n_u8, vrshr_n_u16, vrshr_n_u32, vrshr_n_u64, vrshrq_n_u8,
275 vrshrq_n_u16, vrshrq_n_u32, vrshrq_n_u64, vrshrd_n_u64, vshll_n_u8,
276 vshll_n_u16, vshll_n_u32, vuqadd_s8, vuqadd_s16, vuqadd_s32,
277 vuqadd_s64, vuqaddq_s8, vuqaddq_s16, vuqaddq_s32, vuqaddq_s64,
278 vuqaddb_s8, vuqaddh_s16, vuqadds_s32, vuqaddd_s64): Add signedness
279 suffix to builtin function name, remove cast.
280 (vshl_s8, vshl_s16, vshl_s32, vshl_s64, vshl_u8, vshl_u16, vshl_u32,
281 vshl_u64, vshlq_s8, vshlq_s16, vshlq_s32, vshlq_s64, vshlq_u8,
282 vshlq_u16, vshlq_u32, vshlq_u64, vshld_s64, vshld_u64): Remove cast.
284 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
286 * gcc/config/aarch64/aarch64-builtins.c
287 (aarch64_types_binop_uus_qualifiers,
288 aarch64_types_shift_to_unsigned_qualifiers,
289 aarch64_types_unsigned_shiftacc_qualifiers): Define.
290 * gcc/config/aarch64/aarch64-simd-builtins.def (uqshl, uqrshl, uqadd,
291 uqsub, usqadd, usra_n, ursra_n, uqshrn_n, uqrshrn_n, usri_n, usli_n,
292 sqshlu_n, uqshl_n): Update qualifiers.
293 * gcc/config/aarch64/arm_neon.h (vqadd_u8, vqadd_u16, vqadd_u32,
294 vqadd_u64, vqaddq_u8, vqaddq_u16, vqaddq_u32, vqaddq_u64, vqsub_u8,
295 vqsub_u16, vqsub_u32, vqsub_u64, vqsubq_u8, vqsubq_u16, vqsubq_u32,
296 vqsubq_u64, vqaddb_u8, vqaddh_u16, vqadds_u32, vqaddd_u64, vqrshl_u8,
297 vqrshl_u16, vqrshl_u32, vqrshl_u64, vqrshlq_u8, vqrshlq_u16,
298 vqrshlq_u32, vqrshlq_u64, vqrshlb_u8, vqrshlh_u16, vqrshls_u32,
299 vqrshld_u64, vqrshrn_n_u16, vqrshrn_n_u32, vqrshrn_n_u64,
300 vqrshrnh_n_u16, vqrshrns_n_u32, vqrshrnd_n_u64, vqshl_u8, vqshl_u16,
301 vqshl_u32, vqshl_u64, vqshlq_u8, vqshlq_u16, vqshlq_u32, vqshlq_u64,
302 vqshlb_u8, vqshlh_u16, vqshls_u32, vqshld_u64, vqshl_n_u8, vqshl_n_u16,
303 vqshl_n_u32, vqshl_n_u64, vqshlq_n_u8, vqshlq_n_u16, vqshlq_n_u32,
304 vqshlq_n_u64, vqshlb_n_u8, vqshlh_n_u16, vqshls_n_u32, vqshld_n_u64,
305 vqshlu_n_s8, vqshlu_n_s16, vqshlu_n_s32, vqshlu_n_s64, vqshluq_n_s8,
306 vqshluq_n_s16, vqshluq_n_s32, vqshluq_n_s64, vqshlub_n_s8,
307 vqshluh_n_s16, vqshlus_n_s32, vqshlud_n_s64, vqshrn_n_u16,
308 vqshrn_n_u32, vqshrn_n_u64, vqshrnh_n_u16, vqshrns_n_u32,
309 vqshrnd_n_u64, vqsubb_u8, vqsubh_u16, vqsubs_u32, vqsubd_u64,
310 vrsra_n_u8, vrsra_n_u16, vrsra_n_u32, vrsra_n_u64, vrsraq_n_u8,
311 vrsraq_n_u16, vrsraq_n_u32, vrsraq_n_u64, vrsrad_n_u64, vsli_n_u8,
312 vsli_n_u16, vsli_n_u32,vsli_n_u64, vsliq_n_u8, vsliq_n_u16,
313 vsliq_n_u32, vsliq_n_u64, vslid_n_u64, vsqadd_u8, vsqadd_u16,
314 vsqadd_u32, vsqadd_u64, vsqaddq_u8, vsqaddq_u16, vsqaddq_u32,
315 vsqaddq_u64, vsqaddb_u8, vsqaddh_u16, vsqadds_u32, vsqaddd_u64,
316 vsra_n_u8, vsra_n_u16, vsra_n_u32, vsra_n_u64, vsraq_n_u8,
317 vsraq_n_u16, vsraq_n_u32, vsraq_n_u64, vsrad_n_u64, vsri_n_u8,
318 vsri_n_u16, vsri_n_u32, vsri_n_u64, vsriq_n_u8, vsriq_n_u16,
319 vsriq_n_u32, vsriq_n_u64, vsrid_n_u64): Remove casts.
321 2014-06-03 Teresa Johnson <tejohnson@google.com>
323 * tree-sra.c (modify_function): Record caller nodes after rebuild.
325 2014-06-02 Jason Merrill <jason@redhat.com>
328 * varpool.c (ctor_for_folding): Handle uninitialized vtables.
330 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
332 * config/aarch64/aarch64.c (aarch64_evpc_ext): allow and handle
335 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
337 * config/aarch64/aarch64-simd.md (aarch64_rev<REVERSE:rev-op><mode>):
339 * config/aarch64/aarch64.c (aarch64_evpc_rev): New function.
340 (aarch64_expand_vec_perm_const_1): Add call to aarch64_evpc_rev.
341 * config/aarch64/iterators.md (REVERSE): New iterator.
342 (UNSPEC_REV64, UNSPEC_REV32, UNSPEC_REV16): New enum elements.
343 (rev_op): New int_attribute.
344 * config/aarch64/arm_neon.h (vrev16_p8, vrev16_s8, vrev16_u8,
345 vrev16q_p8, vrev16q_s8, vrev16q_u8, vrev32_p8, vrev32_p16, vrev32_s8,
346 vrev32_s16, vrev32_u8, vrev32_u16, vrev32q_p8, vrev32q_p16, vrev32q_s8,
347 vrev32q_s16, vrev32q_u8, vrev32q_u16, vrev64_f32, vrev64_p8,
348 vrev64_p16, vrev64_s8, vrev64_s16, vrev64_s32, vrev64_u8, vrev64_u16,
349 vrev64_u32, vrev64q_f32, vrev64q_p8, vrev64q_p16, vrev64q_s8,
350 vrev64q_s16, vrev64q_s32, vrev64q_u8, vrev64q_u16, vrev64q_u32):
351 Replace temporary __asm__ with __builtin_shuffle.
353 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
355 * config/mips/mips-cpus.def: Add mips32r3, mips32r5, mips64r3 and
357 * config/mips/mips-tables.opt: Regenerate.
358 * config/mips/mips.c (mips_compute_frame_info): Changed if statement
359 to use mips_isa_rev rather than ISA_MIPS32R2.
360 * config/mips/mips.h (ISA_MIPS32R3): New define.
361 (ISA_MIPS32R5): New define.
362 (ISA_MIPS64R3): New define.
363 (ISA_MIPS64R5): New define.
364 (TARGET_CPU_CPP_BUILTINS): Added support for ISA_MIPS32R3, ISA_MIPS32R5,
365 ISA_MIPS64R3 and ISA_MIPS64R5.
366 (MIPS_ISA_LEVEL_SPEC): Added support for mips32r3, mips32r5, mips64r3
368 (MIPS_ISA_SYNCI_SPEC): Likewise.
369 (ISA_HAS_64BIT_REGS): Added ISA_MIPS64R3 and ISA_MIPS64R5.
370 (LINK_SPEC): Added mips32r3 and mips32r5.
371 * config/mips/t-isa3264 (MULTILIB_MATCHES): Map mips32r3 and mips32r5
372 to mips32r2; and mips64r3 and mips64r5 to mips64r2.
373 * config/mips/t-mti-elf (MULTILIB_MATCHES): Likewise.
374 * config/mips/t-mti-linux (MULTILIB_MATCHES): Likewise.
375 * config/mips/t-sde (MULTILIB_MATCHES): Likewise.
376 * config/mips/t-sdemtk (MULTILIB_MATCHES): New define.
377 * doc/invoke.texi: Document mips32r3, mips32r5, mips64r3 and mips64r5.
379 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
381 * doc/invoke.texi: Document -mxpa and -mno-xpa MIPS command line
383 * config/mips/mips.opt (mxpa): New option.
384 * config/mips/mips.h (ASM_SPEC): Pass mxpa and mno-xpa to the
387 2014-06-03 Martin Jambor <mjambor@suse.cz>
390 * ipa-cp.c (cgraph_edge_brings_value_p): Handle edges leading to
393 2014-06-03 Thomas Preud'homme <thomas.preudhomme@arm.com>
395 PR tree-optimization/61328
396 * tree-ssa-math-opts.c (init_symbolic_number): Extract symbolic number
397 initialization from find_bswap_or_nop_1.
398 (find_bswap_or_nop_1): Test return value of find_bswap_or_nop_1 stored
399 in source_expr2 before using the size value the function sets. Also
400 make use of init_symbolic_number () in both the old place and
401 find_bswap_or_nop_load () to avoid reading uninitialized memory when
402 doing recursion in the GIMPLE_BINARY_RHS case.
404 2014-06-03 Richard Biener <rguenther@suse.de>
406 PR tree-optimization/61383
407 * tree-ssa-ifcombine.c (bb_no_side_effects_p): Make sure
410 2014-06-03 Richard Sandiford <rdsandiford@googlemail.com>
412 * defaults.h (USE_MD_CONSTRAINTS, EXTRA_MEMORY_CONSTRAINT)
413 (EXTRA_ADDRESS_CONSTRAINT, DEFAULT_CONSTRAINT_LEN, CONSTRAINT_LEN)
414 (CONST_OK_FOR_CONSTRAINT_P, CONST_DOUBLE_OK_FOR_LETTER_P)
415 (REG_CLASS_FROM_CONSTRAINT, EXTRA_CONSTRAINT_STR): Delete definitions
417 (REG_CLASS_FROM_LETTER, CONST_OK_FOR_LETTER_P)
418 (CONST_DOUBLE_OK_FOR_LETTER_P, EXTRA_CONSTRAINT): Move poising to...
419 * system.h: ...here and make it unconditional.
420 * target.def (conditional_register_usage): Mention
421 define_register_constraint instead of old-style constraint macros.
422 * doc/tm.texi.in: Remove documentation for old-style constraint macros.
423 * doc/tm.texi: Regenerate.
424 * genoutput.c: Remove USE_MD_CONSTRAINTS conditions and all code
425 protected by !USE_MD_CONSTRAINTS.
426 * config/frv/frv.md: Remove quote from old version of documentation.
427 * config/frv/frv.c (frv_conditional_register_usage): Likewise.
428 * config/m32r/m32r.c (easy_di_const, easy_df_const): Avoid mentioning
429 CONST_DOUBLE_OK_FOR_LETTER.
430 * config/sh/constraints.md: Likewise EXTRA_CONSTRAINT.
432 2014-06-02 Andrew Pinski <apinski@cavium.com>
434 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER):
435 /lib/ld-linux32-aarch64.so.1 is used for ILP32.
436 (LINUX_TARGET_LINK_SPEC): Update linker script for ILP32.
437 file whose name depends on -mabi= and -mbig-endian.
438 * config/aarch64/t-aarch64-linux (MULTILIB_OSDIRNAMES):
439 Handle LP64 better and handle ilp32 too.
440 (MULTILIB_OPTIONS): Delete.
441 (MULTILIB_DIRNAMES): Delete.
443 2014-06-02 Andrew MacLeod <amacleod@redhat.com>
445 * expr.h: Remove prototypes of functions defined in builtins.c.
446 * tree.h: (build_call_expr_*, build_string_literal): Add prototypes.
447 Remove prototypes of functions defined in builtins.c.
448 * builtins.h: Update prototype list to include all exported functions.
449 * builtins.c: (default_libc_has_function, gnu_libc_has_function,
450 no_c99_libc_has_function): Move to targhooks.c
451 (build_string_literal, build_call_expr_loc_array,
452 build_call_expr_loc_vec, build_call_expr_loc, build_call_expr): Move
454 (expand_builtin_object_size, fold_builtin_object_size): Make static.
455 * targhooks.c (default_libc_has_function, gnu_libc_has_function,
456 no_c99_libc_has_function): Relocate from builtins.c.
457 * tree.c: Include builtins.h.
458 (build_call_expr_loc_array, build_call_expr_loc_vec,
459 build_call_expr_loc, build_call_expr, build_string_literal): Relocate
461 * fold-const.h (fold_fma): Move prototype to builtins.h.
462 * realmpfr.h (do_mpc_arg2): Move prototype to builtins.h.
463 * asan.c: Include builtins.h.
464 * cfgexpand.c: Likewise.
465 * convert.c: Likewise.
466 * emit-rtl.c: Likewise.
467 * except.c: Likewise.
469 * fold-const.c: Likewise.
470 * gimple-fold.c: Likewise.
471 * gimple-ssa-strength-reduction.c: Likewise.
472 * gimplify.c: Likewise.
473 * ipa-inline.c: Likewise.
474 * ipa-prop.c: Likewise.
475 * lto-streamer-out.c: Likewise.
477 * tree-inline.c: Likewise.
478 * tree-object-size.c: Likewise.
479 * tree-sra.c: Likewise.
480 * tree-ssa-ccp.c: Likewise.
481 * tree-ssa-forwprop.c: Likewise.
482 * tree-ssa-loop-ivcanon.c: Likewise.
483 * tree-ssa-loop-ivopts.c: Likewise.
484 * tree-ssa-math-opts.c: Likewise.
485 * tree-ssa-reassoc.c: Likewise.
486 * tree-ssa-threadedge.c: Likewise.
487 * tree-streamer-in.c: Likewise.
488 * tree-vect-data-refs.c: Likewise.
489 * tree-vect-patterns.c: Likewise.
490 * tree-vect-stmts.c: Likewise.
491 * config/aarch64/aarch64.c: Likewise.
492 * config/alpha/alpha.c: Likewise.
493 * config/arc/arc.c: Likewise.
494 * config/arm/arm.c: Likewise.
495 * config/avr/avr.c: Likewise.
496 * config/bfin/bfin.c: Likewise.
497 * config/c6x/c6x.c: Likewise.
498 * config/cr16/cr16.c: Likewise.
499 * config/cris/cris.c: Likewise.
500 * config/epiphany/epiphany.c: Likewise.
501 * config/fr30/fr30.c: Likewise.
502 * config/frv/frv.c: Likewise.
503 * config/h8300/h8300.c: Likewise.
504 * config/i386/i386.c: Likewise.
505 * config/i386/winnt.c: Likewise.
506 * config/ia64/ia64.c: Likewise.
507 * config/iq2000/iq2000.c: Likewise.
508 * config/lm32/lm32.c: Likewise.
509 * config/m32c/m32c.c: Likewise.
510 * config/m32r/m32r.c: Likewise.
511 * config/m68k/m68k.c: Likewise.
512 * config/mcore/mcore.c: Likewise.
513 * config/mep/mep.c: Likewise.
514 * config/microblaze/microblaze.c: Likewise.
515 * config/mips/mips.c: Likewise.
516 * config/mmix/mmix.c: Likewise.
517 * config/mn10300/mn10300.c: Likewise.
518 * config/moxie/moxie.c: Likewise.
519 * config/msp430/msp430.c: Likewise.
520 * config/nds32/nds32.c: Likewise.
521 * config/pa/pa.c: Likewise.
522 * config/pdp11/pdp11.c: Likewise.
523 * config/picochip/picochip.c: Likewise.
524 * config/rl78/rl78.c: Likewise.
525 * config/rs6000/rs6000.c: Likewise.
526 * config/rx/rx.c: Likewise.
527 * config/s390/s390.c: Likewise.
528 * config/score/score.c: Likewise.
529 * config/sh/sh.c: Likewise.
530 * config/sparc/sparc.c: Likewise.
531 * config/spu/spu.c: Likewise.
532 * config/stormy16/stormy16.c: Likewise.
533 * config/tilegx/tilegx.c: Likewise.
534 * config/tilepro/tilepro.c: Likewise.
535 * config/v850/v850.c: Likewise.
536 * config/vax/vax.c: Likewise.
537 * config/xtensa/xtensa.c: Likewise.
539 2014-06-02 Jeff Law <law@redhat.com>
541 PR rtl-optimization/61094
542 * ree.c (combine_reaching_defs): Do not reextend an insn if it
543 was marked as do_no_reextend. If a copy is needed to eliminate
544 an extension, then mark it as do_not_reextend.
546 2014-06-02 Marcus Shawcroft <marcus.shawcroft@arm.com>
548 * config/aarch64/aarch64.md (set_fpcr): Drop ISB after FPCR write.
550 2014-06-02 Richard Henderson <rth@redhat.com>
553 * config/alpha/alpha.c (print_operand_address): Allow symbolic
554 addresses inside asms. Use output_operand_lossage instead of
557 2014-06-02 Uros Bizjak <ubizjak@gmail.com>
560 * config/i386/i386.c (ix86_expand_vec_perm) [case V32QImode]: Use
561 GEN_INT (-128) instead of GEN_INT (128) to set MSB of QImode constant.
563 2014-06-02 Tom de Vries <tom@codesourcery.com>
565 * config/aarch64/aarch64.c (aarch64_float_const_representable_p): Handle
566 case that x has VOIDmode.
568 2014-06-02 Bernd Schmidt <bernds@codesourcery.com>
570 * varasm.c (copy_constant): Delete function.
571 (build_constant_desc): Don't call it.
573 2014-06-02 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
576 * config/arm/arm.h (TARGET_SUPPORTS_WIDE_INT): Define.
577 * config/arm/arm.md (mov64 splitter): Replace const_double_operand
578 with immediate_operand.
580 2014-06-02 Andreas Schwab <schwab@suse.de>
583 (ia64_first_cycle_multipass_dfa_lookahead_guard): Check
584 pending_data_specs first.
586 2014-06-02 Richard Biener <rguenther@suse.de>
588 PR tree-optimization/61378
589 * tree-ssa-sccvn.c (vn_reference_lookup_3): Initialize
592 2014-06-01 Uros Bizjak <ubizjak@gmail.com>
594 * config/i386/constraints.md (Bw): Rename from 'w'.
595 (Bz): Rename from 'z'.
596 * config/i386/i386.md: Change 'w' to 'Bw' and 'z' to 'Bz' globally.
598 2014-06-01 Kai Tietz <ktietz@redhat.com>
601 * config/i386/constrains.md (define_constrain): New 'Bs' constraint.
602 * config/i386/i386.md (sibcall_insn_operand): Use Bs
603 instead of m constraint.
605 2014-05-31 Andreas Schwab <schwab@linux-m68k.org>
607 * config/m68k/m68k.md (beq0_di, bne0_di): Make the "o" constraint
608 a separate alternative where the scratch operand 2 is marked as
611 2014-05-31 Kugan Vivekanandarajah <kuganv@linaro.org>
613 * config/arm/arm.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New define.
614 (arm_builtins) : Add ARM_BUILTIN_GET_FPSCR and ARM_BUILTIN_SET_FPSCR.
615 (bdesc_2arg) : Add description for builtins __builtins_arm_set_fpscr
616 and __builtins_arm_get_fpscr.
617 (arm_init_builtins) : Initialize builtins __builtins_arm_set_fpscr and
618 __builtins_arm_get_fpscr.
619 (arm_expand_builtin) : Expand builtins __builtins_arm_set_fpscr and
620 __builtins_arm_ldfpscr.
621 (arm_atomic_assign_expand_fenv): New function.
622 * config/arm/vfp.md (set_fpscr): New pattern.
623 (get_fpscr) : Likewise.
624 * config/arm/unspecs.md (unspecv): Add VUNSPEC_GET_FPSCR and
626 * doc/extend.texi (AARCH64 Built-in Functions) : Document
627 __builtins_arm_set_fpscr, __builtins_arm_get_fpscr.
629 2014-05-30 Jakub Jelinek <jakub@redhat.com>
631 * asan.c (report_error_func): Add SLOW_P argument, use
632 BUILT_IN_ASAN_*_N if set.
633 (build_check_stmt): Likewise.
634 (instrument_derefs): If T has insufficient alignment,
635 force same handling as for odd sizes.
637 * sanitizer.def (BUILT_IN_ASAN_REPORT_LOAD_N,
638 BUILT_IN_ASAN_REPORT_STORE_N): New.
639 * asan.c (struct asan_mem_ref): Change access_size type to
641 (asan_mem_ref_init, asan_mem_ref_new, get_mem_refs_of_builtin_call,
642 update_mem_ref_hash_table): Likewise.
643 (asan_mem_ref_hasher::hash): Hash in a HWI.
644 (report_error_func): Change size_in_bytes argument to HWI.
645 Use *_N builtins if size_in_bytes is larger than 16 or not power of
647 (build_shadow_mem_access): New function.
648 (build_check_stmt): Use it. Change size_in_bytes argument to HWI.
649 Handle size_in_bytes not power of two or larger than 16.
650 (instrument_derefs): Don't give up if size_in_bytes is not
651 power of two or is larger than 16.
653 2014-05-30 Kai Tietz <ktietz@redhat.com>
656 * config/i386/i386.c (x86_output_mi_thunk): Add memory case
657 for sibling-tail-calls.
658 * config/i386/i386.md (sibcall_insn_operand): Add memory-constrain
660 * config/i386/predicates.md (sibcall_memory_operand): New predicate.
661 (sibcall_insn_operand): Add check for sibcall_memory_operand.
663 2014-05-30 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
665 * config/avr/avr-mcus.def: Change ATA6289 ISA to AVR4
666 * config/avr/avr-tables.opt: Regenerate.
667 * config/avr/t-multilib: Regenerate.
668 * doc/avr-mmcu.texi: Regenerate.
670 2014-05-30 Ian Lance Taylor <iant@google.com>
672 * config/i386/xmmintrin.h (_mm_pause): Move out of scope of pragma
675 2014-05-30 Tom de Vries <tom@codesourcery.com>
677 * config/i386/i386.c (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS):
680 2014-05-30 Tom de Vries <tom@codesourcery.com>
682 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
683 * lra.c (initialize_lra_reg_info_element): Add init of
684 actual_call_used_reg_set field.
685 (lra): Call lra_create_live_ranges before lra_inheritance for
687 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
689 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
690 instead of call_used_reg_set for -fuse-caller-save.
691 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
693 2014-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
695 * config/arm/thumb2.md (*thumb2_movhi_insn): Set type of movw
697 * config/arm/vfp.md (*thumb2_movsi_vfp): Likewise.
699 2014-05-30 Richard Sandiford <rdsandiford@googlemail.com>
701 * ira.c (ira_get_dup_out_num): Check for output operands at
702 the start of the loop. Handle cases where an included alternative
703 follows an excluded one.
705 2014-05-29 Mike Stump <mikestump@comcast.net>
708 * collect2.c (maybe_run_lto_and_relink): Be sure to always run
709 post ld passes when lto is used.
711 2014-05-29 Vladimir Makarov <vmakarov@redhat.com>
713 PR rtl-optimization/61325
714 * lra-constraints.c (process_address): Rename to process_address_1.
715 (process_address): New function.
717 2014-05-29 Alan Lawrence <alan.lawrence@arm.com>
719 * config/aarch64/aarch64-builtins.c (aarch64_types_binopv_qualifiers,
720 TYPES_BINOPV): New static data.
721 * config/aarch64/aarch64-simd-builtins.def (im_lane_bound):
723 * config/aarch64/aarch64-simd.md (aarch64_ext,
724 aarch64_im_lane_boundsi): New patterns.
725 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_const_1): Match
727 (aarch64_evpc_ext): New function.
729 * config/aarch64/iterators.md (UNSPEC_EXT): New enum element.
731 * config/aarch64/arm_neon.h (vext_f32, vext_f64, vext_p8, vext_p16,
732 vext_s8, vext_s16, vext_s32, vext_s64, vext_u8, vext_u16, vext_u32,
733 vext_u64, vextq_f32, vextq_f64, vextq_p8, vextq_p16, vextq_s8,
734 vextq_s16, vextq_s32, vextq_s64, vextq_u8, vextq_u16, vextq_u32,
735 vextq_u64): Replace __asm with __builtin_shuffle and im_lane_boundsi.
737 2014-05-29 Tom de Vries <tom@codesourcery.com>
739 * rtl.h (BLOCK_SYMBOL_CHECK): Use SYMBOL_REF_FLAGS.
741 2014-05-29 Richard Earnshaw <rearnsha@arm.com>
742 Richard Sandiford <rdsandiford@googlemail.com>
744 * arm/iterators.md (shiftable_ops): New code iterator.
745 (t2_binop0, arith_shift_insn): New code attributes.
746 * arm/predicates.md (shift_nomul_operator): New predicate.
747 * arm/arm.md (insn_enabled): Delete.
748 (enabled): Remove insn_enabled test.
749 (*arith_shiftsi): Delete. Replace with ...
750 (*<arith_shift_insn>_multsi): ... new pattern.
751 (*<arith_shift_insn>_shiftsi): ... new pattern.
752 * config/arm/arm.c (arm_print_operand): Handle operand format 'b'.
754 2014-05-29 Radovan Obradovic <robradovic@mips.com>
755 Tom de Vries <tom@codesourcery.com>
757 * config/mips/mips.h (POST_CALL_TMP_REG): Define.
758 * config/mips/mips.c (mips_emit_call_insn): Add POST_CALL_TMP_REG
760 (mips_split_call): Use POST_CALL_TMP_REG.
761 (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): Redefine to true.
763 2014-05-29 Tom de Vries <tom@codesourcery.com>
765 * final.c (collect_fn_hard_reg_usage): Guard variable declaration
766 with #ifdef STACK_REGS.
768 2014-05-28 Jan Hubicka <hubicka@ucw.cz>
770 * varasm.c (get_variable_section): Walk aliases.
771 (place_block_symbol): Walk aliases.
773 2014-05-28 Tom de Vries <tom@codesourcery.com>
776 2014-05-28 Tom de Vries <tom@codesourcery.com>
778 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
779 * lra.c (initialize_lra_reg_info_element): Add init of
780 actual_call_used_reg_set field.
781 (lra): Call lra_create_live_ranges before lra_inheritance for
783 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
785 * lra-constraints.c (need_for_call_save_p): Use
786 actual_call_used_reg_set instead of call_used_reg_set for
788 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
790 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
792 * doc/md.texi: Document that the % constraint character must
793 be at the beginning of the string.
794 * genoutput.c (validate_insn_alternatives): Check that '=',
795 '+' and '%' only appear at the beginning of a constraint.
796 * ira.c (commutative_constraint_p): Delete.
797 (ira_get_dup_out_num): Expect the '%' commutativity marker to be
798 at the start of the string.
799 * config/alpha/alpha.md (*movmemdi_1, *clrmemdi_1): Remove
801 * config/arm/neon.md (bicdi3_neon): Likewise.
802 * config/iq2000/iq2000.md (addsi3_internal, subsi3_internal, sgt_si)
803 (slt_si, sltu_si): Likewise.
804 * config/vax/vax.md (sbcdi3): Likewise.
805 * config/h8300/h8300.md (*cmpstz): Remove duplicate '+'.
806 * config/arc/arc.md (mulsi_600, mulsidi_600, umulsidi_600)
807 (mul64): Move '%' to beginning of constraint.
808 * config/arm/arm.md (*xordi3_insn): Likewise.
809 * config/nds32/nds32.md (add<mode>3, mulsi3, andsi3, iorsi3)
812 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
814 * doc/md.texi: Document the restrictions on the "enabled" attribute.
816 2014-05-28 Jason Merrill <jason@redhat.com>
819 * cgraph.h (symtab_node::get_comdat_group_id): New.
820 * cgraphunit.c (analyze_functions): Call it.
821 * symtab.c (dump_symtab_node): Likewise.
822 * tree.c (decl_comdat_group_id): New.
823 * tree.h: Declare it.
824 * lto-streamer-out.c (write_symbol): Use it.
825 * trans-mem.c (ipa_tm_create_version_alias): Likewise.
827 2014-05-28 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
830 * wide-int.cc: Do not include longlong.h when compiling with clang.
832 2014-05-28 Richard Biener <rguenther@suse.de>
834 * tree-ssa-propagate.c (add_control_edge): Print less vertical space.
835 * tree-vrp.c (extract_range_from_ssa_name): Also copy VR_UNDEFINED.
836 (vrp_visit_assignment_or_call): Print less vertical space.
837 (vrp_visit_stmt): Likewise.
838 (vrp_visit_phi_node): Likewise. For a PHI argument with
839 VR_VARYING range consider recording it as copy.
841 2014-05-28 Richard Biener <rguenther@suse.de>
844 2014-05-28 Richard Biener <rguenther@suse.de>
846 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
848 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
850 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
851 sufficiently aligned and an offset is used at the same time.
852 (expand_expr_real_1): Likewise.
854 2014-05-28 Richard Biener <rguenther@suse.de>
857 * fold-const.c (fold_comparison): When folding
858 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
859 the sign of the remaining constant operand stays the same.
861 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
863 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
864 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
865 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
867 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES is true.
868 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
869 (m32bit-doubles) Likewise.
870 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
871 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
874 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
876 * configure.ac ($gcc_cv_ld_clearcap): New test.
877 * configure: Regenerate.
878 * config.in: Regenerate.
879 * config/sol2.opt (mclear-hwcap): New option.
880 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
881 * config/sol2-clearcap.map: Moved here from
882 testsuite/gcc.target/i386/clearcap.map.
883 * config/sol2-clearcapv2.map: Move here from
884 gcc.target/i386/clearcapv2.map.
885 * config/t-sol2 (install): Depend on install-clearcap-map.
886 (install-clearcap-map): New target.
887 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
890 2014-05-28 Richard Biener <rguenther@suse.de>
892 * hwint.h (*_HALF_WIDE_INT*): Move to ...
893 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
894 ... here and remove the rest.
895 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
897 2014-05-28 Richard Biener <rguenther@suse.de>
899 PR tree-optimization/61335
900 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
901 new range fails, drop to varying.
903 2014-05-28 Tom de Vries <tom@codesourcery.com>
905 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
906 * lra.c (initialize_lra_reg_info_element): Add init of
907 actual_call_used_reg_set field.
908 (lra): Call lra_create_live_ranges before lra_inheritance for
910 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
912 * lra-constraints.c (need_for_call_save_p): Use
913 actual_call_used_reg_set instead of call_used_reg_set for
915 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
917 2014-05-28 Radovan Obradovic <robradovic@mips.com>
918 Tom de Vries <tom@codesourcery.com>
920 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save
922 (@item -fuse-caller-save): New item.
924 2014-05-28 Radovan Obradovic <robradovic@mips.com>
925 Tom de Vries <tom@codesourcery.com>
927 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
928 OPT_fuse_caller_save.
930 2014-05-28 Radovan Obradovic <robradovic@mips.com>
931 Tom de Vries <tom@codesourcery.com>
933 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
934 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
935 get_call_reg_set_usage.
936 * resource.c (mark_set_resources, mark_target_live_regs): Use
937 get_call_reg_set_usage.
938 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
940 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
941 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
942 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
943 * ira-build.c (ira_create_allocno): Init
944 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
945 (create_cap_allocno, propagate_allocno_info)
946 (propagate_some_info_from_allocno)
947 (copy_info_to_removed_store_destinations): Handle
948 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
949 * ira-costs.c (ira_tune_allocno_costs): Use
950 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
952 2014-05-28 Radovan Obradovic <robradovic@mips.com>
953 Tom de Vries <tom@codesourcery.com>
955 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
956 and function_used_regs_valid fields.
957 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
958 find_all_hard_reg_sets.
959 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
960 (get_call_reg_set_usage): New function.
961 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
962 * regs.h (get_call_reg_set_usage): Declare.
964 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
967 * config/dbx.h (License): Add Runtime Library Exception.
968 * config/newlib-stdint.h (License): Same.
969 * config/rtems.h (License): Same
970 * config/initfini-array.h (License): Same
971 * config/v850/v850.h (License): Same.
972 * config/v850/v850-opts.h (License): Same
973 * config/v850/rtems.h (License): Same.
975 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
978 * doc/extend.texi (Local Labels): Note that label differences are
979 not supported for AVR.
981 2014-05-27 Andrew Pinski <apinski@cavium.com>
983 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
984 Use <w> for the register in assembly template.
985 (stack_protect_test): Use the mode of operands[0] for the result.
986 (stack_protect_test_<mode>): Use <w> for the register
987 in assembly template.
989 2014-05-27 DJ Delorie <dj@redhat.com>
991 * config/rx/rx.c (add_vector_labels): New.
992 (rx_output_function_prologue): Call it.
993 (rx_handle_func_attribute): Don't require empty arguments.
994 (rx_handle_vector_attribute): New.
995 (rx_attribute_table): Add "vector" attribute.
996 * doc/extend.texi (interrupt, vector): Document new/changed
997 RX-specific attributes.
999 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
1001 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
1003 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
1004 predicate to detect a negative quotient.
1006 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
1008 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
1009 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
1010 Add X - Y CMP 0 to X CMP Y transformation.
1011 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
1013 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
1015 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
1018 2014-05-27 Steve Ellcey <sellcey@mips.com>
1020 * config/mips/mips.c: Add include of cgraph.h.
1022 2014-05-27 Richard Biener <rguenther@suse.de>
1024 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
1026 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
1029 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
1030 * config/arm/arm-cores.def (License): Same.
1031 * config/arm/arm-opts.h (License): Same.
1032 * config/arm/aout.h (License): Same.
1033 * config/arm/bpabi.h (License): Same.
1034 * config/arm/elf.h (License): Same.
1035 * config/arm/linux-elf.h (License): Same.
1036 * config/arm/linux-gas.h (License): Same.
1037 * config/arm/netbsd-elf.h (License): Same.
1038 * config/arm/uclinux-eabi.h (License): Same.
1039 * config/arm/uclinux-elf.h (License): Same.
1040 * config/arm/vxworks.h (License): Same.
1042 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1044 * config/arm/neon.md (neon_bswap<mode>): New pattern.
1045 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
1046 (arm_init_neon_builtins): Handle NEON_BSWAP.
1047 Define required type nodes.
1048 (arm_expand_neon_builtin): Handle NEON_BSWAP.
1049 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
1050 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
1051 * config/arm/iterators.md (VDQHSD): New mode iterator.
1053 2014-05-27 Richard Biener <rguenther@suse.de>
1055 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
1056 Try using literal operands when comparing value-ranges failed.
1058 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
1060 * ira.c (commutative_operand): Adjust for change to recog_data.
1061 [Missing from previous commit.]
1063 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
1065 * system.h (TEST_BIT): New macro.
1066 * recog.h (alternative_mask): New type.
1067 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
1068 (recog_data_d): Replace alternative_enabled_p array with
1069 enabled_alternatives.
1070 (target_recog): New structure.
1071 (default_target_recog, this_target_recog): Declare.
1072 (get_enabled_alternatives, recog_init): Likewise.
1073 * recog.c (default_target_recog, this_target_recog): New variables.
1074 (get_enabled_alternatives): New function.
1075 (extract_insn): Use it.
1076 (recog_init): New function.
1077 (preprocess_constraints, constrain_operands): Adjust for change to
1079 * postreload.c (reload_cse_simplify_operands): Likewise.
1080 * reload.c (find_reloads): Likewise.
1081 * ira-costs.c (record_reg_classes): Likewise.
1082 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
1083 all alternatives after a disabled one would be skipped.
1084 (ira_implicitly_set_insn_hard_regs): Likewise.
1085 * ira.c (ira_setup_alts): Adjust for change to recog_data.
1086 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
1087 with enabled_alternatives.
1088 * lra.c (free_insn_recog_data): Update accordingly.
1089 (lra_update_insn_recog_data): Likewise.
1090 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
1091 * lra-constraints.c (process_alt_operands): Likewise. Handle
1092 only_alternative as part of the enabled mask.
1093 * target-globals.h (this_target_recog): Declare.
1094 (target_globals): Add a recog field.
1095 (restore_target_globals): Restore this_target_recog.
1096 * target-globals.c: Include recog.h.
1097 (default_target_globals): Initialize recog field.
1098 (save_target_globals): Likewise.
1099 * reginfo.c (reinit_regs): Call recog_init.
1100 * toplev.c (backend_init_target): Likewise.
1102 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
1104 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
1105 rather than any named insn's code.
1107 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
1110 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
1111 * config/arm/arm-cores.def (License): Same.
1113 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
1115 * tree.h (decl_comdat_group): Declare.
1116 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
1117 * tree.c (decl_comdat_group): Here.
1119 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
1121 PR rtl-optimization/61222
1122 * combine.c (simplify_shift_const_1): When moving a PLUS outside
1123 the shift, truncate the PLUS operand to the result mode.
1125 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
1128 * config/i386/i386.c (ix86_rtx_costs)
1129 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
1132 2014-05-26 Martin Jambor <mjambor@suse.cz>
1134 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
1137 2014-05-26 Richard Biener <rguenther@suse.de>
1139 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
1140 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
1141 Provide specializations.
1142 (wi::int_traits <HOST_WIDE_INT>,
1143 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
1145 2014-05-26 Alan Modra <amodra@gmail.com>
1148 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
1149 params and return a bool. Remove dead code. Update comment.
1150 Assert we have a const_int source. Remove bogus code from
1151 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
1152 handling of constants > 2G and reg_equal note, from..
1153 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
1154 return value. Update comment. If we can, use a new pseudo
1155 for intermediate calculations.
1156 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
1158 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
1159 call to rs6000_emit_set_const in splitter.
1160 (movdi_internal64+2, +3): Likewise.
1162 2014-05-26 Richard Biener <rguenther@suse.de>
1164 * system.h: Define __STDC_FORMAT_MACROS before
1165 including inttypes.h.
1166 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
1167 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
1168 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
1169 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
1170 HOST_WIDEST_INT_C): Remove.
1171 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
1172 if C99 inttypes.h is not available.
1173 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
1174 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
1175 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
1176 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
1177 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
1178 (struct output_info): Likewise.
1179 (print_statistics): Adjust.
1180 (dump_bitmap_statistics): Likewise.
1181 * bt-load.c (migrate_btr_defs): Print with PRId64.
1182 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
1183 (MAX_SAFE_MULTIPLIER): Adjust.
1184 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
1185 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
1186 dump_cgraph_node): Likewise.
1187 * final.c (dump_basic_block_info): Likewise.
1188 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
1189 * gcov.c (format_gcov): Likewise.
1190 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
1192 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
1193 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
1194 (inline_small_functions, dump_overall_stats, dump_inline_stats):
1195 Use PRId64 for dumping.
1196 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
1197 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
1198 (add_allocno_hard_regs): Adjust.
1199 * loop-doloop.c (doloop_modify): Print using PRId64.
1200 * loop-iv.c (inverse): Compute in uint64_t.
1201 (determine_max_iter, iv_number_of_iterations): Likewise.
1202 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
1204 * lto-streamer-out.c (write_symbol): Use uint64_t.
1205 * mcf.c (CAP_INFINITY): Use int64_t maximum.
1206 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
1207 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
1208 * modulo-sched.c (const_iteration_count): Use int64_t.
1209 (sms_schedule): Dump using PRId64.
1210 * predict.c (dump_prediction): Likewise.
1211 * pretty-print.h (pp_widest_integer): Remove.
1212 * profile.c (get_working_sets, is_edge_inconsistent,
1213 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
1214 * tree-pretty-print.c (pp_double_int): Remove case handling
1215 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
1216 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
1218 (pass_optimize_bswap::execute): Remove restriction on hosts.
1219 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
1220 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
1221 * tree.c (widest_int_cst_value): Remove.
1222 * tree.h (widest_int_cst_value): Likewise.
1223 * value-prof.c (dump_histogram_value): Print using PRId64.
1224 * gengtype.c (main): Also inject int64_t.
1225 * ggc-page.c (struct max_alignment): Use int64_t.
1226 * alloc-pool.c (struct allocation_object_def): Likewise.
1227 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
1229 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
1230 * doc/tm.texi: Regenerated.
1231 * gengtype-lex.l (IWORD): Handle [u]int64_t.
1232 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
1233 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
1234 mmix_output_register_setting): Use [u]int64_t in prototypes.
1235 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
1236 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
1237 mmix_output_octa, mmix_output_shifted_value): Adjust.
1238 (mmix_intval): Adjust. Remove unreachable case.
1239 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
1241 2014-05-26 Richard Biener <rguenther@suse.de>
1243 * configure.ac: Drop __int64 type check. Insist that we
1244 found uint64_t and int64_t.
1245 * hwint.h (HOST_BITS_PER___INT64): Remove.
1246 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
1247 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
1248 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
1249 (HOST_WIDEST_FAST_INT): Remove __int64 case.
1250 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
1251 for dst_q_src_df_rms_cdt.
1252 * configure: Regenerate.
1253 * config.in: Likewise.
1255 2014-05-26 Michael Tautschnig <mt@debian.org>
1258 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
1259 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
1261 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1263 PR rtl-optimization/61278
1264 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
1266 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1268 PR rtl-optimization/61220
1269 Part of PR rtl-optimization/61225
1270 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
1271 insn; skip split_edge for a block with only one successor.
1273 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1275 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
1278 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1280 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
1281 (update_vtable_references): New function.
1282 (function_and_variable_visibility): Rewrite also vtable initializers.
1283 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
1285 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1287 * ggc.h (ggc_grow): New function.
1288 * ggc-none.c (ggc_grow): New function.
1289 * ggc-page.c (ggc_grow): Likewise.
1291 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1293 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
1294 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
1295 comdat_can_be_unshared_p, cgraph_externally_visible_p,
1296 varpool_externally_visible_p, can_replace_by_local_alias,
1297 update_visibility_by_resolution_info, function_and_variable_visibility,
1298 pass_data_ipa_function_and_variable_visibility,
1299 make_pass_ipa_function_and_variable_visibility,
1300 whole_program_function_and_variable_visibility,
1301 pass_data_ipa_whole_program_visibility,
1302 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
1303 * cgraph.h (cgraph_local_node_p): Declare.
1304 * ipa-visibility.c: New file.
1305 * Makefile.in (OBJS): Add ipa-visiblity.o
1307 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1309 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
1310 that var decl is available.
1312 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1314 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
1315 symtab_node pointer.
1316 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
1317 (find_decls_types_r): Do not walk COMDAT_GROUP.
1318 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
1319 * varasm.c (make_decl_one_only): Use set_comdat_group;
1320 create node if needed.
1321 * ipa-inline-transform.c (save_inline_function_body): Update
1322 way we decl->symtab mapping.
1323 * symtab.c (symtab_hash, hash_node, eq_node
1324 symtab_insert_node_to_hashtable): Remove.
1325 (symtab_register_node): Update.
1326 (symtab_unregister_node): Update.
1327 (symtab_get_node): Reimplement as inline function.
1328 (symtab_add_to_same_comdat_group): Update.
1329 (symtab_dissolve_same_comdat_group_list): Update.
1330 (dump_symtab_base): Update.
1331 (verify_symtab_base): Update.
1332 (symtab_make_decl_local): Update.
1333 (fixup_same_cpp_alias_visibility): Update.
1334 (symtab_nonoverwritable_alias): Update.
1335 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
1336 * ipa.c (update_visibility_by_resolution_info): UPdate.
1337 * bb-reorder.c: Include cgraph.h
1338 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
1340 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
1341 * cgraph.c (cgraph_get_create_node): Update.
1342 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
1344 (symtab_get_node): Make inline.
1345 (symtab_insert_node_to_hashtable): Remove.
1346 (symtab_can_be_discarded): Update.
1347 (decl_comdat_group): New function.
1348 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
1350 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
1352 (read_comdat_group): New function.
1353 (input_node, input_varpool_node): Use it.
1354 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
1356 * mips.c (mips_start_unique_function): Likewise.
1357 (ix86_code_end): Likewise.
1358 (rs6000_code_end): Likweise.
1359 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
1361 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1363 * gengtype-state.c (fatal_reading_state): Bring offline.
1364 * optabs.c (widening_optab_handler): Bring offline.
1365 * optabs.h (widening_optab_handler): Likewise.
1366 * final.c (get_attr_length_1): Likewise.
1368 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1370 * sched-int.h (sd_iterator_cond): Manually tail recurse.
1372 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1374 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
1375 (ppc440-compare): Include shift with dot.
1376 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
1377 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
1379 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
1381 (e6500_sfx2): Include it.
1382 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
1383 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
1384 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
1385 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
1386 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
1387 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
1388 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
1389 *lshiftrt_internal1le, *lshiftrt_internal1be,
1390 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
1391 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
1392 *rotldi3_internal10le, *rotldi3_internal10be,
1393 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
1394 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
1395 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
1396 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
1397 define_insns): Use type "shift" in the appropriate alternatives.
1399 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1401 * config/rs6000/rs6000.md (type): Add "logical". Delete
1403 (dot): Adjust comment.
1404 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
1405 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
1406 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
1407 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
1408 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
1409 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
1410 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
1411 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1413 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1414 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1415 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1416 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1417 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1418 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1419 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1420 * config/rs6000/8540.md (ppc8540_su): Adjust.
1421 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1422 cell-cmp-microcoded): Adjust.
1423 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1424 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1425 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1426 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1427 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1428 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1429 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1430 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1431 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1433 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1434 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
1435 Adjust. Adjust comment.
1436 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1437 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1439 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1441 * config/rs6000/rs6000.md (type): Add "add".
1442 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
1443 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
1444 define_insns): Use it.
1445 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1447 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1448 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1449 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1450 * config/rs6000/601.md (ppc601-integer): Adjust.
1451 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1452 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1453 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1454 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1455 * config/rs6000/8540.md (ppc8540_su): Adjust.
1456 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1457 cell-cmp-microcoded): Adjust.
1458 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1459 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1460 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1461 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1462 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1463 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1464 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1465 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1466 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1468 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1469 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
1470 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1471 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1473 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1475 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
1476 "delayed_compare", "var_delayed_compare".
1477 (var_shift): New attribute.
1478 (cell_micro): Adjust.
1479 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
1480 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
1481 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
1482 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
1483 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
1484 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
1485 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
1486 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
1487 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
1488 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
1489 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
1490 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
1491 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
1492 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
1493 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
1494 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
1495 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
1496 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
1497 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
1498 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
1499 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
1500 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
1501 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
1502 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1503 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1505 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1506 * config/rs6000/440.md (ppc440-integer): Adjust.
1507 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1508 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
1509 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1510 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1511 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1512 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1513 * config/rs6000/8540.md (ppc8540_su): Adjust.
1514 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1515 cell-cmp-microcoded): Adjust.
1516 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
1517 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1518 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
1519 e500mc64_delayed): Adjust.
1520 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
1521 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
1522 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1523 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
1524 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
1525 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
1526 power6-delayed-compare, power6-var-delayed-compare): Adjust.
1527 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
1528 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
1530 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1531 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1533 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1535 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
1536 (bits): New mode_attr.
1537 (idiv_ldiv): Delete mode_attr.
1538 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
1539 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1540 rs6000_adjust_priority, is_nonpipeline_insn,
1541 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1543 * config/rs6000/40x.md (ppc403-idiv): Adjust.
1544 * config/rs6000/440.md (ppc440-idiv): Adjust.
1545 * config/rs6000/476.md (ppc476-idiv): Adjust.
1546 * config/rs6000/601.md (ppc601-idiv): Adjust.
1547 * config/rs6000/603.md (ppc603-idiv): Adjust.
1548 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
1549 ppc620-ldiv): Adjust.
1550 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
1551 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
1552 * config/rs6000/8540.md (ppc8540_divide): Adjust.
1553 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
1554 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
1555 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
1556 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
1557 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
1558 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
1559 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
1560 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
1561 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
1562 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
1563 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
1564 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
1565 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
1566 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
1567 * config/rs6000/titan.md (titan_fxu_div): Adjust.
1569 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1571 * config/rs6000/rs6000.md (type): Delete "insert_word",
1572 "insert_dword". Add "insert".
1573 (size): Update comment.
1574 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1575 insn_must_be_first_in_group): Adjust.
1576 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
1577 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
1578 *insvsi_internal6, insvdi_internal): Adjust.
1580 * config/rs6000/40x.md (ppc403-integer): Adjust.
1581 * config/rs6000/440.md (ppc440-integer): Adjust.
1582 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
1583 * config/rs6000/601.md (ppc601-integer): Adjust.
1584 * config/rs6000/603.md (ppc603-integer): Adjust.
1585 * config/rs6000/6xx.md (ppc604-integer): Adjust.
1586 * config/rs6000/7450.md (ppc7450-integer): Adjust.
1587 * config/rs6000/7xx.md (ppc750-integer): Adjust.
1588 * config/rs6000/8540.md (ppc8540_su): Adjust.
1589 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
1590 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
1591 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1592 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
1593 * config/rs6000/e5500.md (e5500_sfx): Adjust.
1594 * config/rs6000/e6500.md (e6500_sfx): Adjust.
1595 * config/rs6000/mpc.md (mpccore-integer): Adjust.
1596 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
1597 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
1598 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
1599 * config/rs6000/power7.md (power7-integer): Adjust.
1600 * config/rs6000/power8.md (power8-1cyc): Adjust.
1601 * config/rs6000/rs64.md (rs64a-integer): Adjust.
1602 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1604 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1606 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
1607 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
1608 (size): New attribute.
1609 (dot): New attribute.
1610 (cell_micro): Adjust.
1611 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
1612 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
1613 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
1614 umuldi3_highpart): Adjust.
1615 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1616 rs6000_adjust_priority, is_nonpipeline_insn,
1617 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1619 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
1620 ppc405-imul3): Adjust.
1621 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
1622 * config/rs6000/476.md (ppc476-imul): Adjust.
1623 * config/rs6000/601.md (ppc601-imul): Adjust.
1624 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
1625 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
1626 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
1627 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
1628 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
1630 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
1631 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
1632 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
1634 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
1635 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
1636 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
1637 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
1638 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
1639 * config/rs6000/mpc.md (mpccore-imul): Adjust.
1640 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
1641 power4-lmul, power4-imul, power4-imul3): Adjust.
1642 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
1643 power5-lmul, power5-imul, power5-imul3): Adjust.
1644 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
1645 power6-lmul, power6-imul, power6-imul3): Adjust.
1646 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
1647 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
1649 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
1650 rs64a-lmul): Adjust.
1651 * config/rs6000/titan.md (titan_imul): Adjust.
1653 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1655 * config/rs6000/rs6000.md (type): Add new value "halfmul".
1656 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
1657 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
1658 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
1659 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
1660 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
1661 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
1662 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
1663 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
1664 * config/rs6000/titan.md: Delete nonsensical comment.
1665 (titan_imul): Add type imul3.
1666 (titan_mulhw): Remove type imul3; add type halfmul.
1668 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1670 * config/rs6000/rs6000.md (type): Reorder, reformat.
1672 2014-05-23 Martin Jambor <mjambor@suse.cz>
1674 PR tree-optimization/53787
1675 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
1676 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
1677 analysis_done, update all uses.
1678 * ipa-prop.c: Include domwalk.h
1679 (param_analysis_info): Removed.
1680 (param_aa_status): New type.
1681 (ipa_bb_info): Likewise.
1682 (func_body_info): Likewise.
1683 (ipa_get_bb_info): New function.
1684 (aa_overwalked): Likewise.
1685 (find_dominating_aa_status): Likewise.
1686 (parm_bb_aa_status_for_bb): Likewise.
1687 (parm_preserved_before_stmt_p): Changed to use new param AA info.
1688 (load_from_unmodified_param): Accept func_body_info as a parameter
1689 instead of parms_ainfo.
1690 (parm_ref_data_preserved_p): Changed to use new param AA info.
1691 (parm_ref_data_pass_through_p): Likewise.
1692 (ipa_load_from_parm_agg_1): Likewise. Update callers.
1693 (compute_complex_assign_jump_func): Changed to use new param AA info.
1694 (compute_complex_ancestor_jump_func): Likewise.
1695 (ipa_compute_jump_functions_for_edge): Likewise.
1696 (ipa_compute_jump_functions): Removed.
1697 (ipa_compute_jump_functions_for_bb): New function.
1698 (ipa_analyze_indirect_call_uses): Likewise, moved variable
1700 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
1701 and info, moved variable declarations down.
1702 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
1704 (ipa_analyze_stmt_uses): Likewise.
1705 (ipa_analyze_params_uses): Removed.
1706 (ipa_analyze_params_uses_in_bb): New function.
1707 (ipa_analyze_controlled_uses): Likewise.
1708 (free_ipa_bb_info): Likewise.
1709 (analysis_dom_walker): New class.
1710 (ipa_analyze_node): Handle node-specific forbidden analysis,
1711 initialize and free func_body_info, use dominator walker.
1712 (ipcp_modif_dom_walker): New class.
1713 (ipcp_transform_function): Create and free func_body_info, use
1714 ipcp_modif_dom_walker, moved a lot of functionality there.
1716 2014-05-23 Marek Polacek <polacek@redhat.com>
1717 Jakub Jelinek <jakub@redhat.com>
1719 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
1720 * gcc.c (sanitize_spec_function): Likewise.
1721 * convert.c (convert_to_integer): Include "ubsan.h". Add
1722 floating-point to integer instrumentation.
1723 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
1724 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
1725 SANITIZE_NONDEFAULT.
1726 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
1727 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
1728 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
1729 * ubsan.c: Include "realmpfr.h" and "dfp.h".
1730 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
1731 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
1732 float/double/long double.
1733 (ubsan_instrument_float_cast): New function.
1734 * ubsan.h (ubsan_instrument_float_cast): Declare.
1736 2014-05-23 Jiong Wang <jiong.wang@arm.com>
1738 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
1740 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
1741 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
1742 Adjust for tailcalling through registers.
1743 * config/aarch64/aarch64.h (enum reg_class): New caller save
1745 (REG_CLASS_NAMES): Likewise.
1746 (REG_CLASS_CONTENTS): Likewise.
1747 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
1748 Allow tailcalling without decls.
1750 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1752 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1753 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
1755 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
1756 gsi, and variables v_* to v*.
1758 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
1760 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
1762 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1764 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
1765 * omp-low.c: Update accordingly.
1767 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
1768 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
1769 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
1770 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
1771 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
1772 GF_OMP_TARGET_KIND_UPDATE.
1774 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1775 Explicitly enumerate the expected region types.
1777 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
1780 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
1781 documentation; the old documentation didn't clearly state the
1782 constraints on the contents of the pointed-to storage.
1784 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1786 Fix bootstrap error on ia64
1787 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
1788 Return default value.
1790 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
1792 PR tree-optimization/54733
1793 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
1795 (find_bswap_or_nop_load): New.
1796 (find_bswap_1): Renamed to ...
1797 (find_bswap_or_nop_1): This. Also add support for memory source.
1798 (find_bswap): Renamed to ...
1799 (find_bswap_or_nop): This. Also add support for memory source and
1800 detection of bitwise operations equivalent to load in host endianness.
1801 (execute_optimize_bswap): Likewise. Also move its leading comment back
1802 in place and split statement transformation into ...
1803 (bswap_replace): This.
1805 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1807 PR rtl-optimization/61215
1808 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
1809 simplify_gen_subreg until final substitution.
1811 2014-05-23 Alan Modra <amodra@gmail.com>
1814 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
1815 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
1816 Use "Y" constraint rather than "m".
1818 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1820 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
1822 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
1823 New function declaration.
1824 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
1825 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
1826 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
1827 (aarch64_init_builtins) : Initialize builtins
1828 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1829 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1830 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
1831 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
1832 and __builtins_aarch64_set_fpsr.
1833 (aarch64_atomic_assign_expand_fenv): New function.
1834 * config/aarch64/aarch64.md (set_fpcr): New pattern.
1835 (get_fpcr) : Likewise.
1836 (set_fpsr) : Likewise.
1837 (get_fpsr) : Likewise.
1838 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
1839 and UNSPECV_SET_FPSR.
1840 * doc/extend.texi (AARCH64 Built-in Functions) : Document
1841 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1842 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1844 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1846 PR rtl-optimization/60969
1847 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
1848 constraints. Set up mem cost for NO_REGS case.
1850 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
1852 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
1854 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
1856 * config/darwin.c: Include "lto-section-names.h".
1857 (LTO_SEGMENT_NAME): Don't define.
1858 * config/i386/winnt.c: Include "lto-section-names.h".
1859 * lto-streamer.c: Include "lto-section-names.h".
1860 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
1861 * lto-wrapper.c: Include "lto-section-names.h".
1862 (LTO_SECTION_NAME_PREFIX): Don't define.
1863 * lto-section-names.h: New file.
1864 * cgraphunit.c: Include "lto-section-names.h".
1866 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
1868 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
1870 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
1873 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
1875 2014-05-22 Nick Clifton <nickc@redhat.com>
1877 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
1879 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
1881 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
1882 -> (T)A transformation to integer types.
1884 2014-05-22 Teresa Johnson <tejohnson@google.com>
1886 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
1887 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
1888 (gcov_rewrite): Use gcov_nonruntime_assert.
1890 (gcov_write_words): Ditto.
1891 (gcov_write_length): Ditto.
1892 (gcov_read_words): Use gcov_nonruntime_assert, and remove
1893 gcc_assert from IN_LIBGCOV code.
1894 (gcov_read_summary): Use gcov_error to flag profile corruption.
1895 (gcov_sync): Use gcov_nonruntime_assert.
1896 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
1897 (gcov_histo_index): Use gcov_nonruntime_assert.
1898 (static void gcov_histogram_merge): Ditto.
1899 (compute_working_sets): Ditto.
1900 * gcov-io.h (gcov_nonruntime_assert): Define.
1901 (gcov_error): Define for !IN_LIBGCOV
1903 2014-05-22 Richard Biener <rguenther@suse.de>
1905 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1906 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
1907 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
1908 and deallocation site.
1909 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1910 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
1911 passing through the incoming points-to set.
1912 (handle_lhs_call): Use flags argument instead of recomputing it.
1913 (find_func_aliases_for_call): Call handle_lhs_call with proper
1916 2014-05-22 Jakub Jelinek <jakub@redhat.com>
1918 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
1919 all padding bits in REAL_VALUE_TYPE are cleared.
1921 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1923 Cleanup and improve multipass_dfa_lookahead_guard
1924 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
1925 (core2i7_first_cycle_multipass_begin,)
1926 (core2i7_first_cycle_multipass_issue,)
1927 (core2i7_first_cycle_multipass_backtrack): Update signature.
1928 * config/ia64/ia64.c
1929 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
1930 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
1931 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
1933 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
1934 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
1936 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
1938 * doc/tm.texi: Regenerate.
1940 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
1941 * haifa-sched.c (ready_try): Make signed to allow negative values.
1942 (rebug_ready_list_1): Update.
1943 (choose_ready): Simplify.
1944 (sched_extend_ready_list): Update.
1946 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1948 Remove IA64 speculation tweaking flags
1949 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
1950 speculation tuning flags.
1951 (msched-prefer-non-data-spec-insns,)
1952 (msched-prefer-non-control-spec-insns): Obsolete options.
1953 * haifa-sched.c (choose_ready): Remove handling of
1954 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1955 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
1956 and PREFER_NON_DATA_SPEC.
1957 * sel-sched.c (process_spec_exprs): Remove handling of
1958 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1960 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1962 Improve scheduling debug output
1963 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
1964 (advance_one_cycle): Update.
1965 (schedule_insn, queue_to_ready): Add debug printouts.
1966 (debug_ready_list_1): New static function.
1967 (debug_ready_list): Update.
1968 (max_issue): Add debug printouts.
1969 (dump_insn_stream): New static function.
1970 (schedule_block): Use it. Also better indent printouts.
1972 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1974 Fix sched_insn debug counter
1975 * haifa-sched.c (schedule_insn): Update.
1976 (struct haifa_saved_data): Add nonscheduled_insns_begin.
1977 (save_backtrack_point, restore_backtrack_point): Update.
1978 (first_nonscheduled_insn): New static function.
1979 (queue_to_ready, choose_ready): Use it.
1980 (schedule_block): Init nonscheduled_insns_begin.
1981 (sched_emit_insn): Update.
1984 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
1986 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
1988 (aarch64_secondary_reload) : LikeWise.
1989 (aarch64_class_max_nregs) : Remove CORE_REGS.
1990 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
1991 (REG_CLASS_NAMES) : Likewise.
1992 (REG_CLASS_CONTENTS) : LikeWise.
1993 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
1995 2014-05-21 Guozhi Wei <carrot@google.com>
1998 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
2000 (vqdmulhq_n_s16): Likewise.
2002 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
2004 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
2006 2014-05-21 Marek Polacek <polacek@redhat.com>
2009 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
2011 2014-05-21 Martin Jambor <mjambor@suse.cz>
2013 * doc/invoke.texi (Optimize Options): Document parameters
2014 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
2015 ipa-cp-array-index-hint-bonus.
2017 2014-05-21 Mark Wielaard <mjw@redhat.com>
2020 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
2021 version >= 3 or not strict DWARF.
2022 * langhooks.h (struct lang_hooks_for_types): Add
2023 enum_underlying_base_type.
2024 * langhooks.c (lhd_enum_underlying_base_type): New function.
2025 * gcc/langhooks.h (struct lang_hooks_for_types): Add
2026 enum_underlying_base_type.
2027 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
2028 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
2029 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
2031 2014-05-21 Richard Biener <rguenther@suse.de>
2033 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
2035 2014-05-21 John Marino <gnugcc@marino.st>
2037 * config.gcc (*-*-dragonfly*): New target.
2038 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
2039 * configure: Regenerate.
2040 * config/dragonfly-stdint.h: New.
2041 * config/dragonfly.h: New.
2042 * config/dragonfly.opt: New.
2043 * config/i386/dragonfly.h: New.
2044 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
2046 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2048 * tree.def (VOID_CST): New.
2049 * tree-core.h (TI_VOID): New.
2050 * tree.h (void_node): New.
2051 * tree.c (tree_node_structure_for_code, tree_code_size)
2052 (iterative_hash_expr): Handle VOID_CST.
2053 (build_common_tree_nodes): Initialize void_node.
2055 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
2057 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
2059 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
2061 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
2064 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
2065 flag_reorder_blocks_and_partition.
2066 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
2068 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
2071 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
2073 (*addc_r_t): Add new insn_and_split.
2075 2014-05-21 Jakub Jelinek <jakub@redhat.com>
2078 * omp-low.c (handle_simd_reference): New function.
2079 (lower_rec_input_clauses): Use it. Defer adding reference
2080 initialization even for reduction without placeholder if in simd,
2081 handle it properly later on.
2083 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2085 PR tree-optimization/60899
2086 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
2087 assume all static symbols will have definition wile parsing and
2088 check the do have definition later in compilation; check that
2089 variable referring symbol will be output before concluding that
2090 reference is safe; be conservative for referring local statics;
2091 be more precise about when comdat is output in other partition.
2093 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2096 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
2098 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
2099 (ipa_inline): Loop inline_to_all_callers until no more aliases
2102 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2104 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
2105 set writeonly flag only for vars actually written to.
2107 2014-05-20 Dehao Chen <dehao@google.com>
2109 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
2110 and callee count to get clone count.
2111 * tree-inline.c (expand_call_inline): Use callee count instead of bb
2114 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
2116 PR rtl-optimization/61243
2117 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
2119 2014-05-20 Xinliang David Li <davidxl@google.com>
2121 * cgraphunit.c (walk_polymorphic_call_targets): Add
2122 dbgcnt and fopt-info support.
2123 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
2124 * ipa-devirt.c (ipa_devirt): Ditto.
2125 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
2126 * ipa.c (walk_polymorphic_call_targets): Ditto.
2127 * gimple-fold.c (fold_gimple_assign): Ditto.
2128 (gimple_fold_call): Ditto.
2129 * dbgcnt.def: New counter.
2131 2014-05-20 DJ Delorie <dj@redhat.com>
2133 * config/msp430/msp430.md (split): Don't allow subregs when
2134 splitting SImode adds.
2135 (andneghi): Fix subtraction logic.
2136 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
2138 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
2140 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
2142 * except.c (switch_to_exception_section, resolve_unique_section,
2143 get_named_text_section, default_function_rodata_section,
2144 align_variable, get_block_for_decl, default_section_type_flags):
2145 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
2146 * symtab.c (symtab_add_to_same_comdat_group,
2147 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
2148 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
2150 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
2151 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
2152 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
2153 (c6x_function_in_section_p): Likewise.
2154 * config/darwin.c (machopic_select_section): Likewise.
2155 * config/arm/arm.c (arm_function_in_section_p): Likewise.
2156 * config/mips/mips.c (mips_function_rodata_section): Likewise.
2157 * config/mep/mep.c (mep_select_section): LIkewise.
2158 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
2160 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
2162 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
2163 EH region of calls to pure functions that can throw an exception.
2164 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
2165 (copy_reference_ops_from_call): Also copy the EH region of the call if
2166 it can throw an exception.
2168 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2170 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
2171 nested VEC_SELECTs that are inverses of each other.
2173 2014-05-20 Richard Biener <rguenther@suse.de>
2175 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
2176 (extract_and_process_scc_for_name): not here.
2177 (cond_dom_walker::before_dom_children): Only process
2178 stmts that end the BB in interesting ways.
2179 (run_scc_vn): Mark param uses as visited.
2181 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2183 * config/arm/arm.md (arith_shiftsi): Do not predicate for
2186 2014-05-20 Nick Clifton <nickc@redhat.com>
2188 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
2189 (msp430_gimplify_va_arg_expr): New function.
2190 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
2192 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
2193 operand 0 in order to prevent confusion about the number of
2196 2014-05-20 Richard Biener <rguenther@suse.de>
2198 PR tree-optimization/61221
2199 * tree-ssa-pre.c (el_to_update): Remove.
2200 (eliminate_dom_walker::before_dom_children): Handle released
2201 VDEFs by value-numbering them to the associated VUSE. Update
2202 stmt immediately for substituted call address.
2203 (eliminate): Remove delayed stmt updating code.
2204 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
2205 possibly late re-numbered vuses.
2206 (vn_reference_lookup_2): Adjust.
2207 (vn_reference_lookup_pieces): Likewise.
2208 (vn_reference_lookup): Likewise.
2210 2014-05-20 Richard Biener <rguenther@suse.de>
2212 * config.gcc: Remove need_64bit_hwint.
2213 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
2214 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
2216 * config.in: Regenerate.
2217 * configure: Likewise.
2219 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
2221 * doc/extend.texi: Create Label Attributes section,
2222 move all label attributes into it and reference it.
2224 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
2226 * arm.c (thumb1_reorg): When scanning backwards skip anything
2227 that's not a proper insn.
2229 2014-05-19 Richard Biener <rguenther@suse.de>
2231 PR tree-optimization/61221
2232 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
2233 Do nothing for unreachable blocks.
2234 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
2235 Improve unreachability detection.
2237 2014-05-19 Richard Biener <rguenther@suse.de>
2239 PR tree-optimization/61209
2240 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
2242 2014-05-19 Nick Clifton <nickc@redhat.com>
2244 * except.c (init_eh): Fix computation of builtin setjmp buffer
2245 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
2247 2014-05-19 Richard Biener <rguenther@suse.de>
2249 PR tree-optimization/61184
2250 * tree-vrp.c (is_negative_overflow_infinity): Use
2251 TREE_OVERFLOW_P and do that check first.
2252 (is_positive_overflow_infinity): Likewise.
2253 (is_overflow_infinity): Likewise.
2254 (vrp_operand_equal_p): Properly treat operands with
2255 differing overflow as not equal.
2257 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
2259 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
2260 shift simplification where it was intended.
2262 2014-05-19 Christian Bruel <christian.bruel@st.com>
2265 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
2267 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
2270 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
2273 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
2275 * reg-notes.def (CROSSING_JUMP): Likewise.
2276 * rtl.h (rtx_def): Update comment for jump flag.
2277 (CROSSING_JUMP_P): Define.
2278 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
2279 of a REG_CROSSING_JUMP note.
2280 * cfghooks.c (tidy_fallthru_edges): Likewise.
2281 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
2282 * emit-rtl.c (try_split): Likewise.
2283 * haifa-sched.c (sched_create_recovery_edges): Likewise.
2284 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
2285 * jump.c (redirect_jump_2): Likewise.
2286 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
2287 (relax_delay_slots): Likewise.
2288 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
2289 (bbit_di): Likewise.
2290 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
2291 * config/sh/sh.md (jump_compact): Likewise.
2292 * bb-reorder.c (rotate_loop): Likewise.
2293 (pass_duplicate_computed_gotos::execute): Likewise.
2294 (add_reg_crossing_jump_notes): Rename to...
2295 (update_crossing_jump_flags): ...this.
2296 (pass_partition_blocks::execute): Update accordingly.
2298 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
2300 * tree.h: Remove extraneous template <>.
2302 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2304 * ipa.c (symtab_remove_unreachable_nodes): Remove
2305 symbol from comdat group if its body was eliminated.
2306 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
2307 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
2308 (symtab_unregister_node): ... this one.
2309 (verify_symtab_base): More strict checking of comdats.
2310 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
2312 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2314 * tree-pass.h (make_pass_ipa_comdats): New pass.
2315 * timevar.def (TV_IPA_COMDATS): New timevar.
2316 * passes.def (pass_ipa_comdats): Add.
2317 * Makefile.in (OBJS): Add ipa-comdats.o
2318 * ipa-comdats.c: New file.
2320 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2322 * ipa.c (update_visibility_by_resolution_info): New function.
2323 (function_and_variable_visibility): Use it.
2325 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2327 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
2329 (FOR_EACH_DEFINED_SYMBOL): New macro.
2330 (varpool_first_static_initializer, varpool_next_static_initializer,
2331 varpool_first_defined_variable, varpool_next_defined_variable):
2333 (symtab_in_same_comdat_p): Correctly deal with inline functions.
2335 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2337 * ggc-page.c (ggc_handle_finalizers): Add comment.
2339 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2341 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
2342 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
2343 (ggc_internal_cleared_alloc): Likewise.
2344 * ggc-page.c (finalizer): New class.
2345 (vec_finalizer): Likewise.
2346 (globals::finalizers): New member.
2347 (globals::vec_finalizers): Likewise.
2348 (ggc_internal_alloc): Record the finalizer if any for the block being
2350 (ggc_handle_finalizers): New function.
2351 (ggc_collect): Call ggc_handle_finalizers.
2352 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
2354 (ggc_internal_cleared_alloc): Likewise.
2355 (finalize): New function.
2356 (need_finalization_p): Likewise.
2357 (ggc_alloc): Install the type's destructor as the finalizer if it
2359 (ggc_cleared_alloc): Likewise.
2360 (ggc_vec_alloc): Likewise.
2361 (ggc_cleared_vec_alloc): Likewise.
2363 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2365 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
2367 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2369 * alias.c (record_alias_subset): Adjust.
2370 * bitmap.c (bitmap_element_allocate): Likewise.
2371 (bitmap_gc_alloc_stat): Likewise.
2372 * cfg.c (init_flow): Likewise.
2373 (alloc_block): Likewise.
2374 (unchecked_make_edge): Likewise.
2375 * cfgloop.c (alloc_loop): Likewise.
2376 (flow_loops_find): Likewise.
2377 (rescan_loop_exit): Likewise.
2378 * cfgrtl.c (init_rtl_bb_info): Likewise.
2379 * cgraph.c (insert_new_cgraph_node_version): Likewise.
2380 (cgraph_allocate_node): Likewise.
2381 (cgraph_create_edge_1): Likewise.
2382 (cgraph_allocate_init_indirect_info): Likewise.
2383 * cgraphclones.c (cgraph_clone_edge): Likewise.
2384 * cgraphunit.c (add_asm_node): Likewise.
2385 (init_lowered_empty_function): Likewise.
2386 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
2387 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
2388 (alpha_use_linkage): Likewise.
2389 * config/arc/arc.c (arc_init_machine_status): Likewise.
2390 * config/arm/arm.c (arm_init_machine_status): Likewise.
2391 * config/avr/avr.c (avr_init_machine_status): Likewise.
2392 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
2393 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
2394 * config/cris/cris.c (cris_init_machine_status): Likewise.
2395 * config/darwin.c (machopic_indirection_name): Likewise.
2396 (darwin_build_constant_cfstring): Likewise.
2397 (darwin_enter_string_into_cfstring_table): Likewise.
2398 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
2399 * config/frv/frv.c (frv_init_machine_status): Likewise.
2400 * config/i386/i386.c (get_dllimport_decl): Likewise.
2401 (ix86_init_machine_status): Likewise.
2402 (assign_386_stack_local): Likewise.
2403 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
2404 (i386_pe_maybe_record_exported_symbol): Likewise.
2405 (i386_pe_record_stub): Likewise.
2406 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
2407 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
2408 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
2409 (m32c_note_pragma_address): Likewise.
2410 * config/mep/mep.c (mep_init_machine_status): Likewise.
2411 (mep_note_pragma_flag): Likewise.
2412 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
2413 (mips16_local_alias): Likewise.
2414 (mips_init_machine_status): Likewise.
2415 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
2416 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
2417 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
2418 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
2419 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
2420 * config/pa/pa.c (pa_init_machine_status): Likewise.
2421 (pa_get_deferred_plabel): Likewise.
2422 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
2423 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
2424 (rs6000_init_machine_status): Likewise.
2425 (output_toc): Likewise.
2426 * config/s390/s390.c (s390_init_machine_status): Likewise.
2427 * config/score/score.c (score_output_external): Likewise.
2428 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
2429 * config/spu/spu.c (spu_init_machine_status): Likewise.
2430 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
2431 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
2432 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
2433 * coverage.c (coverage_end_function): Likewise.
2434 * dbxout.c (dbxout_init): Likewise.
2435 * doc/gty.texi: Don't mention variable_size attribute.
2436 * dwarf2cfi.c (new_cfi): Adjust.
2437 (new_cfi_row): Likewise.
2438 (copy_cfi_row): Likewise.
2439 (create_cie_data): Likewise.
2440 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
2441 (new_loc_descr): Likewise.
2442 (find_AT_string_in_table): Likewise.
2443 (add_addr_table_entry): Likewise.
2444 (new_die): Likewise.
2445 (add_var_loc_to_decl): Likewise.
2446 (clone_die): Likewise.
2447 (clone_as_declaration): Likewise.
2448 (break_out_comdat_types): Likewise.
2449 (new_loc_list): Likewise.
2450 (add_loc_descr_to_each): Likewise.
2451 (add_location_or_const_value_attribute): Likewise.
2452 (add_linkage_name): Likewise.
2453 (lookup_filename): Likewise.
2454 (dwarf2out_var_location): Likewise.
2455 (new_line_info_table): Likewise.
2456 (dwarf2out_init): Likewise.
2457 (mem_loc_descriptor): Likewise.
2458 (loc_descriptor): Likewise.
2459 (add_const_value_attribute): Likewise.
2460 (tree_add_const_value_attribute): Likewise.
2461 (comp_dir_string): Likewise.
2462 (dwarf2out_vms_debug_main_pointer): Likewise.
2463 (string_cst_pool_decl): Likewise.
2464 * emit-rtl.c (set_mem_attrs): Likewise.
2465 (get_reg_attrs): Likewise.
2466 (start_sequence): Likewise.
2467 (init_emit): Likewise.
2468 (init_emit_regs): Likewise.
2469 * except.c (init_eh_for_function): Likewise.
2470 (gen_eh_region): Likewise.
2471 (gen_eh_region_catch): Likewise.
2472 (gen_eh_landing_pad): Likewise.
2473 (add_call_site): Likewise.
2474 * function.c (add_frame_space): Likewise.
2475 (insert_temp_slot_address): Likewise.
2476 (assign_stack_temp_for_type): Likewise.
2477 (get_hard_reg_initial_val): Likewise.
2478 (allocate_struct_function): Likewise.
2479 (prepare_function_start): Likewise.
2480 (types_used_by_var_decl_insert): Likewise.
2481 * gengtype.c (variable_size_p): Remove function.
2482 (enum alloc_quantity): Remove enum.
2483 (write_typed_alloc_def): Remove function.
2484 (write_typed_struct_alloc_def): Likewise.
2485 (write_typed_typedef_alloc_def): Likewise.
2486 (write_typed_alloc_defns): Likewise.
2488 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
2489 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
2490 * ggc.h (ggc_alloc): new function.
2491 (ggc_cleared_alloc): Likewise.
2492 (ggc_vec_alloc): Template on type of vector element, and remove
2493 element size argument.
2494 (ggc_cleared_vec_alloc): Likewise.
2495 * gimple.c (gimple_build_omp_for): Adjust.
2496 (gimple_copy): Likewise.
2497 * ipa-cp.c (get_replacement_map): Likewise.
2498 (find_aggregate_values_for_callers_subset): Likewise.
2499 (known_aggs_to_agg_replacement_list): Likewise.
2500 * ipa-devirt.c (get_odr_type): Likewise.
2501 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
2502 (read_agg_replacement_chain): Likewise.
2503 * loop-iv.c (get_simple_loop_desc): Likewise.
2504 * lto-cgraph.c (input_node_opt_summary): Likewise.
2505 * lto-section-in.c (lto_new_in_decl_state): Likewise.
2506 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
2507 (input_eh_region): Likewise.
2508 (input_eh_lp): Likewise.
2509 (input_cfg): Likewise.
2510 * optabs.c (set_optab_libfunc): Likewise.
2511 (init_tree_optimization_optabs): Likewise.
2512 (set_conv_libfunc): Likewise.
2513 * passes.c (do_per_function_toporder): Likewise.
2514 * rtl.h: Don't use variable_size gty attribute.
2515 * sese.c (if_region_set_false_region): Adjust.
2516 * stringpool.c (gt_pch_save_stringpool): Likewise.
2517 * target-globals.c (save_target_globals): Likewise.
2518 * toplev.c (general_init): Likewise.
2519 * trans-mem.c (record_tm_replacement): Likewise.
2520 (split_bb_make_tm_edge): Likewise.
2521 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2522 * tree-data-ref.h (lambda_vector_new): Likewise.
2523 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
2524 * tree-iterator.c (tsi_link_before): Likewise.
2525 (tsi_link_after): Likewise.
2526 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
2527 * tree-ssa-loop-niter.c (record_estimate): Likewise.
2528 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
2529 * tree-ssa-operands.h: Don't use variable_size gty attribute.
2530 * tree-ssa.c (init_tree_ssa): Adjust.
2531 * tree-ssanames.c (set_range_info): Likewise.
2532 (get_ptr_info): Likewise.
2533 (duplicate_ssa_name_ptr_info): Likewise.
2534 (duplicate_ssa_name_range_info): Likewise.
2535 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
2536 (unpack_ts_fixed_cst_value_fields): Likewise.
2537 * tree.c (build_fixed): Likewise.
2538 (build_real): Likewise.
2539 (build_string): Likewise.
2540 (decl_priority_info): Likewise.
2541 (decl_debug_expr_insert): Likewise.
2542 (decl_value_expr_insert): Likewise.
2543 (decl_debug_args_insert): Likewise.
2544 (type_hash_add): Likewise.
2545 (build_omp_clause): Likewise.
2546 * ubsan.c (decl_for_type_insert): Likewise.
2547 * varasm.c (get_unnamed_section): Likewise.
2548 (get_noswitch_section): Likewise.
2549 (get_section): Likewise.
2550 (get_block_for_section): Likewise.
2551 (create_block_symbol): Likewise.
2552 (build_constant_desc): Likewise.
2553 (create_constant_pool): Likewise.
2554 (force_const_mem): Likewise.
2555 (record_tm_clone_pair): Likewise.
2556 * varpool.c (varpool_create_empty_node): Likewise.
2558 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2560 * dwarf2out.c (tree_add_const_value_attribute): Call
2561 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
2562 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
2563 instead of ggc_internal_<x>alloc_stat.
2564 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
2565 (ggc_realloc): Likewise.
2566 * ggc-none.c (ggc_internal_alloc): Likewise.
2567 (ggc_internal_cleared_alloc): Likewise.
2568 * ggc-page.c: Likewise.
2569 * ggc.h (ggc_internal_alloc_stat): Likewise.
2570 (ggc_internal_alloc): Remove macro.
2571 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
2572 (ggc_internal_cleared_alloc): Remove macro.
2573 (GGC_RESIZEVEC): Adjust.
2574 (ggc_resizevar): Remove macro.
2575 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
2576 (ggc_internal_cleared_vec_alloc_stat): Likewise.
2577 (ggc_internal_vec_cleared_alloc): Remove macro.
2578 (ggc_alloc_atomic_stat): Drop _stat suffix.
2579 (ggc_alloc_atomic): Remove macro.
2580 (ggc_alloc_cleared_atomic): Remove macro.
2581 (ggc_alloc_string_stat): Drop _stat suffix.
2582 (ggc_alloc_string): Remove macro.
2583 (ggc_alloc_rtx_def_stat): Adjust.
2584 (ggc_alloc_tree_node_stat): Likewise.
2585 (ggc_alloc_cleared_tree_node_stat): Likewise.
2586 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
2587 (ggc_alloc_cleared_simd_clone_stat): Likewise.
2588 * gimple.c (gimple_build_omp_for): Likewise.
2589 (gimple_copy): Likewise.
2590 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
2591 * toplev.c (realloc_for_line_map): Adjust.
2592 * tree-data-ref.h (lambda_vector_new): Likewise.
2593 * tree-phinodes.c (allocate_phi_node): Likewise.
2594 * tree.c (grow_tree_vec_stat): Likewise.
2595 * vec.h (va_gc::reserve): Adjust.
2597 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
2599 * config/microblaze/microblaze.c (break_handler): New Declaration.
2600 (microblaze_break_function_p,microblaze_is_break_handler): New.
2601 (compute_frame_size): Use microblaze_break_function_p.
2602 Add the test of break_handler.
2603 (microblaze_function_prologue) : Add the test of variable
2604 break_handler. Check the fnname by BREAK_HANDLER_NAME.
2605 (microblaze_function_epilogue) : Add the test of break_handler.
2606 (microblaze_globalize_label) : Add the test of break_handler.
2607 Check the name by BREAK_HANDLER_NAME.
2609 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
2611 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
2612 microblaze_is_break_handler test.
2613 (call_internal1,call_value_intern): Use microblaze_break_function_p.
2614 Use SYMBOL_REF_DECL.
2616 * config/microblaze/microblaze-protos.h
2617 (microblaze_break_function_p,microblaze_is_break_handler):
2620 * doc/extend.texi (MicroBlaze break_handler Functions): Document
2621 new MicroBlaze break_handler functions.
2623 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2625 * doc/extend.texi (Size of an asm): Move node text according
2626 to its @menu entry position.
2628 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
2630 PR tree-optimization/61140
2631 PR tree-optimization/61150
2632 PR tree-optimization/61197
2633 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
2635 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2637 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
2639 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
2641 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
2642 __SIZEOF_INT128__ is defined.
2644 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2646 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
2647 (rs6000_delegitimize_address): Use it.
2649 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2651 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
2652 inplace argument. Store the new address in the original MEM when true.
2653 * emit-rtl.c (change_address_1): Likewise.
2654 (adjust_address_1, adjust_automodify_address_1, offset_address):
2656 * rtl.h (plus_constant): Add an inplace argument.
2657 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
2658 when true. Avoid generating (plus X (const_int 0)).
2659 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
2660 in-place. Pass true to plus_constant.
2661 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
2663 2014-05-16 Dehao Chen <dehao@google.com>
2665 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
2667 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2670 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
2672 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
2674 2014-05-16 Dehao Chen <dehao@google.com>
2676 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
2677 optimize_function_for_size_p.
2678 * regs.h (REG_FREQ_FROM_BB): Likewise.
2680 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2683 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
2684 negt_reg_operand cases.
2685 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
2687 * config/sh/predicates.md (cbranch_treg_value): Simplify.
2689 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2691 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
2694 2014-05-16 David Malcolm <dmalcolm@redhat.com>
2697 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2699 * tree-cfg.c (dump_function_to_file): Dump the return type of
2700 functions, in a line to itself before the function body, mimicking
2701 the layout of a C function.
2703 2014-05-16 Dehao Chen <dehao@google.com>
2705 * cfghooks.c (make_forwarder_block): Use direct computation to
2706 get fall-through edge's count and frequency.
2708 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
2710 * config/arc/arc.c (arc_init): Fix typo in error message.
2711 * config/i386/i386.c (ix86_expand_builtin): Likewise.
2712 (split_stack_prologue_scratch_regno): Likewise.
2713 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
2714 word from error message.
2716 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
2718 * ira-costs.c: Fix typo in comment.
2720 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
2722 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
2724 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
2726 * varpool.c (dump_varpool_node): Dump write-only flag.
2727 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
2729 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
2730 write-only variables.
2731 * ipa.c (process_references): New function.
2732 (set_readonly_bit): New function.
2733 (set_writeonly_bit): New function.
2734 (clear_addressable_bit): New function.
2735 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
2736 fix handling of aliases.
2737 * cgraph.h (struct varpool_node): Add writeonly flag.
2739 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
2741 PR rtl-optimization/60969
2742 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
2743 Calculate costs for this case.
2745 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
2747 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
2748 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
2750 2014-05-16 Richard Biener <rguenther@suse.de>
2752 PR tree-optimization/61194
2753 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
2754 bool patterns ending in a COND_EXPR.
2756 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2758 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
2760 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2762 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
2763 where we were unable to cost an RTX.
2765 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2767 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
2770 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2771 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2773 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
2775 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2776 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2778 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
2779 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
2781 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2782 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2784 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
2787 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2788 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2790 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2793 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2794 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2796 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
2797 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
2799 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2800 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2802 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2805 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2806 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2808 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
2809 ZERO_EXTEND and SIGN_EXTEND better.
2811 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2812 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2814 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
2817 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2818 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2820 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
2821 costs when costing loads and stores to memory.
2823 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2824 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
2826 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
2829 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2831 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
2833 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2834 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2836 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
2838 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
2840 (aarch64_rtx_mult_cost): New.
2841 (aarch64_rtx_costs): Use it, refactor as appropriate.
2843 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2844 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2846 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
2847 emit instructions, return number of instructions which would
2849 (aarch64_add_constant): Update call to aarch64_build_constant.
2850 (aarch64_output_mi_thunk): Likewise.
2851 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
2854 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2856 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
2857 (TARGET_RTX_COSTS): Call it.
2859 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2861 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
2862 (cortexa57_vector_cost): Likewise.
2863 (cortexa57_tunings): Use them.
2865 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2867 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
2868 (cpu_addrcost_table): Use it.
2869 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
2870 (aarch64_address_cost): Rewrite using aarch64_classify_address,
2873 2014-05-16 Richard Biener <rguenther@suse.de>
2875 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
2876 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
2877 (visit_phi): Ignore edges marked as not executable.
2878 (class cond_dom_walker): New.
2879 (cond_dom_walker::before_dom_children): Value-number
2880 control statements and mark successor edges as not
2881 executable if possible.
2882 (run_scc_vn): First walk all control statements in
2883 dominator order, marking edges as not executable.
2884 * tree-inline.c (copy_edges_for_bb): Be not confused
2885 about random edge flags.
2887 2014-05-16 Richard Biener <rguenther@suse.de>
2889 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
2891 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
2894 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
2895 (__TM_simple_begin): Use it.
2896 (__TM_begin): Likewise.
2898 2014-05-15 Martin Jambor <mjambor@suse.cz>
2901 * ipa-prop.c (update_indirect_edges_after_inlining): Check
2902 type_preserved flag when the indirect edge is polymorphic.
2904 2014-05-15 Martin Jambor <mjambor@suse.cz>
2906 PR tree-optimization/61090
2907 * tree-sra.c (sra_modify_expr): Pass the current gsi to
2908 build_ref_for_model.
2910 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2912 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
2913 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
2915 2014-05-15 Jakub Jelinek <jakub@redhat.com>
2917 PR tree-optimization/61158
2918 * fold-const.c (fold_binary_loc): If X is zero-extended and
2919 shiftc >= prec, make sure zerobits is all ones instead of
2920 invoking undefined behavior.
2922 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2924 * regcprop.h: New file.
2925 * regcprop.c (skip_debug_insn_p): New decl.
2926 (replace_oldest_value_reg): Check skip_debug_insn_p.
2927 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
2928 * shrink-wrap.c: Include regcprop.h.
2929 (prepare_shrink_wrap): Call
2930 copyprop_hardreg_forward_bb_without_debug_insn.
2932 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2934 * shrink-wrap.h: Update comment.
2935 * shrink-wrap.c: Update comment.
2936 (next_block_for_reg): Rename to live_edge_for_reg.
2937 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
2938 (move_insn_for_shrink_wrap): Split live_edge.
2939 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
2941 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2943 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
2945 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
2946 * config/sparc/sparc.md (fptype_ut699): New attribute.
2947 (in_branch_delay): Return false if -mfix-ut699 is specified and
2948 fptype_ut699 is set to single.
2949 (truncdfsf2): Add fptype_ut699 attribute.
2950 (fix_truncdfsi2): Likewise.
2951 (floatsisf2): Change fptype attribute.
2952 (fix_truncsfsi2): Likewise.
2953 (negtf2_notv9): Delete.
2954 (negtf2_v9): Likewise.
2955 (negtf2_hq): New instruction.
2956 (negtf2): New instruction and splitter.
2957 (negdf2_notv9): Rewrite.
2958 (abstf2_notv9): Delete.
2959 (abstf2_hq_v9): Likewise.
2960 (abstf2_v9): Likewise.
2961 (abstf2_hq): New instruction.
2962 (abstf2): New instruction and splitter.
2963 (absdf2_notv9): Rewrite.
2965 2014-05-14 Cary Coutant <ccoutant@google.com>
2968 * opts.c (common_handle_option): Don't special-case "-g".
2969 (set_debug_level): Default to at least level 2 with "-g".
2971 2014-05-14 DJ Delorie <dj@redhat.com>
2973 * config/msp430/msp430.c (msp430_builtin): Add
2974 MSP430_BUILTIN_DELAY_CYCLES.
2975 (msp430_init_builtins): Register void __delay_cycles(long long).
2976 (msp430_builtin_decl): Add it.
2977 (cg_magic_constant): New.
2978 (msp430_expand_delay_cycles): New.
2979 (msp430_expand_builtin): Call it.
2980 (msp430_print_operand_raw): Change integer printing from "int" to
2982 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
2983 (delay_cycles_start): New.
2984 (delay_cycles_end): New.
2985 (delay_cycles_32): New.
2986 (delay_cycles_32x): New.
2987 (delay_cycles_16): New.
2988 (delay_cycles_16x): New.
2989 (delay_cycles_2): New.
2990 (delay_cycles_1): New.
2991 * doc/extend.texi: Document __delay_cycles().
2993 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
2995 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
2996 length attribute computation.
2998 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
3001 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
3003 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
3006 * config/sparc/sparc.md: Fix types of low and high in DI constant
3007 splitter. Use gen_int_mode in some other splitters.
3009 2014-05-14 Martin Jambor <mjambor@suse.cz>
3012 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
3014 2014-05-14 James Norris <jnorris@codesourcery.com>
3016 * omp-low.c (expand_parallel_call): Remove shadow variable.
3017 (expand_omp_taskreg): Likewise.
3019 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
3021 * common/config/i386/i386-common.c
3022 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
3023 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
3024 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
3025 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
3026 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
3027 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
3028 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
3029 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
3030 xsavecintrin.h, xsavesintrin.h.
3031 (x86_64-*-*): Ditto.
3032 * config/i386/clflushoptintrin.h: New.
3033 * config/i386/xsavecintrin.h: Ditto.
3034 * config/i386/xsavesintrin.h: Ditto.
3035 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
3036 (bit_XSAVES): Ditto.
3037 (bit_XSAVES): Ditto.
3038 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
3039 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
3041 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
3042 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
3043 OPTION_MASK_ISA_XSAVES.
3044 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
3046 (PTA_CLFLUSHOPT) Define.
3047 (PTA_XSAVEC): Ditto.
3048 (PTA_XSAVES): Ditto.
3049 (ix86_option_override_internal): Handle new options.
3050 (ix86_valid_target_attribute_inner_p): Ditto.
3051 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
3052 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
3053 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
3054 (bdesc_special_args): Add __builtin_ia32_xsaves,
3055 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
3056 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
3057 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
3058 (ix86_expand_builtin): Handle new builtins.
3059 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
3060 (TARGET_CLFLUSHOPT_P): Ditto.
3061 (TARGET_XSAVEC): Ditto.
3062 (TARGET_XSAVEC_P): Ditto.
3063 (TARGET_XSAVES): Ditto.
3064 (TARGET_XSAVES_P): Ditto.
3065 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
3066 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
3067 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
3069 (ANY_XRSTOR64): Ditto.
3071 (xrstor): Change into <xrstor>.
3072 (xrstor_rex64): Change into <xrstor>_rex64.
3073 (xrstor64): Change into <xrstor>64
3075 * config/i386/i386.opt (mclflushopt): New.
3078 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
3080 * doc/invoke.texi: Document new options.
3082 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
3084 PR rtl-optimization/60866
3085 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
3086 Default it to -1. Pass it down to init_simplejump_data.
3087 (init_simplejump_data): New parameter old_seqno. Pass it down
3088 to get_seqno_for_a_jump.
3089 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
3090 initializing new jump seqno as a last resort. Add comment.
3091 (sel_redirect_edge_and_branch): Save old seqno of the conditional
3092 jump and pass it down to sel_init_new_insn.
3093 (sel_redirect_edge_and_branch_force): Likewise.
3095 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
3097 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
3098 shifted values to avoid build warning.
3100 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
3102 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
3103 * cfgrtl.c (rtl_merge_blocks): Fix comment.
3104 (cfg_layout_merge_blocks): Likewise.
3105 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
3107 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
3109 PR rtl-optimization/60901
3110 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
3111 bb predecessor belongs to the same scheduling region. Adjust comment.
3113 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
3115 * doc/sourcebuild.texi: (dfp_hw): Document.
3116 (p8vector_hw): Likewise.
3117 (powerpc_eabi_ok): Likewise.
3118 (powerpc_elfv2): Likewise.
3119 (powerpc_htm_ok): Likewise.
3120 (ppc_recip_hw): Likewise.
3123 2014-05-13 Cary Coutant <ccoutant@google.com>
3125 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
3127 2014-05-13 David Malcolm <dmalcolm@redhat.com>
3129 * gengtype-parse.c (require3): Eliminate in favor of...
3131 (require_template_declaration): Update to support optional single *
3134 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
3135 (create_user_defined_type): Handle a single level of explicit
3136 pointerness within template arguments.
3137 (struct write_types_data): Add field "kind".
3138 (filter_type_name): Handle "*" character.
3139 (write_user_func_for_structure_ptr): Require a write_types_data
3140 rather than just a prefix string, so that we can look up the kind
3141 of the wtd and use it as an index into wrote_user_func_for_ptr,
3142 ensuring that such functions are written at most once. Support
3143 subclasses by invoking the marking function of the ultimate base class.
3144 (write_user_func_for_structure_body): Require a write_types_data
3145 rather than just a prefix string, so that we can pass this to
3146 write_user_func_for_structure_ptr.
3147 (write_func_for_structure): Likewise.
3148 (ggc_wtd): Add initializer of new "kind" field.
3149 (pch_wtd): Likewise.
3151 * gengtype.h (enum write_types_kinds): New.
3152 (struct type): Add field wrote_user_func_for_ptr to the "s"
3155 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
3157 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
3158 instead of const_binop.
3159 (fold_binary_loc): Likewise.
3161 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
3163 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
3164 calculation to match get_ref_base_and_extent.
3166 2014-05-13 Catherine Moore <clm@codesourcery.com>
3167 Sandra Loosemore <sandra@codesourcery.com>
3169 * configure.ac: Fix assembly for explicit JALR relocation check.
3170 * configure: Regenerate.
3172 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3174 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
3175 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
3176 Remove associated type declarations and initialisations.
3177 (arm_expand_neon_builtin): Likewise.
3178 (neon_emit_pair_result_insn): Delete.
3179 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
3180 * config/arm/neon.md (neon_vtrn<mode>): Delete.
3181 (neon_vzip<mode>): Likewise.
3182 (neon_vuzp<mode>): Likewise.
3184 2014-05-13 Richard Biener <rguenther@suse.de>
3187 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
3188 it needs revisiting whether the call still may be tail-called.
3190 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3192 * rtl.def (SYMBOL_REF): Remove middle "0" field.
3193 * rtl.h (block_symbol): Reduce number of fields to 2.
3194 (rtx_def): Add u2.symbol_ref_flags.
3195 (SYMBOL_REF_FLAGS): Use it.
3196 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
3197 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
3198 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
3199 Lower index of SYMBOL_REF_DATA.
3200 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
3201 Print SYMBOL_REF_FLAGS at the same time.
3202 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
3204 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3206 * rtl.def (VAR_LOCATION): Remove "i" field.
3207 * rtl.h (rtx_def): Add u2.var_location_status.
3208 (PAT_VAR_LOCATION_STATUS): Use it.
3209 (gen_rtx_VAR_LOCATION): Declare.
3210 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
3211 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
3212 * var-tracking.c (emit_note_insn_var_location): Remove casts.
3214 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3216 * rtl.def (scratch): Fix outdated comment and remove "0" field.
3217 * gengtype.c (adjust_field_rtx_def): Update accordingly.
3219 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3221 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
3222 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
3223 * rtl.h (rtx_def): Add insn_uid to u2 field.
3224 (RTX_FLAG_CHECK8): Delete in favor of...
3225 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
3226 (INSN_DELETED_P): Update accordingly.
3227 (INSN_UID): Use u2.insn_uid.
3228 (INSN_CHAIN_CODE_P): Define.
3229 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
3230 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
3231 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
3232 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
3233 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
3234 indices accordingly.
3235 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
3236 Update indices for insn-chain rtxes.
3237 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
3238 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
3239 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
3240 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
3241 * combine.c (try_combine): Likewise.
3242 * ira.c (setup_prohibited_mode_move_regs): Likewise.
3244 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3246 * rtl.def (REG): Remove middle field.
3247 * rtl.h (rtx_def): Add orignal_regno to u2.
3248 (ORIGINAL_REGNO): Use it instead of field 1.
3249 (REG_ATTRS): Lower field index accordingly.
3250 * gengtype.c (adjust_field_rtx_def): Remove handling of
3251 ORIGINAL_REGNO. Move REG_ATTRS index down.
3252 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
3253 code that prints the REGNO.
3255 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3257 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
3260 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3262 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
3264 2014-05-13 Bin Cheng <bin.cheng@arm.com>
3266 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
3267 (alloc_iv): Lower base expressions containing ADDR_EXPR.
3269 2014-05-13 Ian Bolton <ian.bolton@arm.com>
3271 * config/aarch64/aarch64-protos.h
3272 (aarch64_hard_regno_caller_save_mode): New prototype.
3273 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
3275 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
3277 2014-05-13 Christian Bruel <christian.bruel@st.com>
3279 * target.def (mode_switching): New hook vector.
3280 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
3281 (mode_exit, modepriority_to_mode): Likewise.
3282 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
3283 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3284 * target.h: Include tm.h and hard-reg-set.h.
3285 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
3286 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
3287 * doc/tm.texi Regenerate.
3288 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
3289 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3290 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
3291 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
3292 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
3293 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3294 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
3295 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
3296 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
3297 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
3298 (ix86_emit_mode_set): Hookify.
3299 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
3301 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3302 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
3303 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
3304 (epiphany_mode_priority_to_mode): Remove declaration.
3305 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
3306 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
3307 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
3309 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
3310 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
3311 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
3313 2014-05-13 Jakub Jelinek <jakub@redhat.com>
3316 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
3317 is const0_rtx, return immediately. Don't test count == 0 when
3320 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3322 * Makefile.in: add shrink-wrap.o.
3323 * config/i386/i386.c: include "shrink-wrap.h"
3324 * function.c: Likewise.
3325 (requires_stack_frame_p, next_block_for_reg,
3326 move_insn_for_shrink_wrap, prepare_shrink_wrap,
3327 dup_block_and_redirect): Move to shrink-wrap.c
3328 (thread_prologue_and_epilogue_insns): Extract three code segments
3329 as functions in shrink-wrap.c
3330 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
3332 * shrink-wrap.c: New file.
3333 * shrink-wrap.h: New file.
3335 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3337 * doc/extend.texi: Reflect current numbers of pragmas. Remove
3338 reference to Solaris.
3340 2014-05-12 Mike Stump <mikestump@comcast.net>
3343 * genattrtab.c (filename): Add.
3344 (convert_set_attr_alternative): Improve error message.
3345 (check_defs): Restore read_md_filename for error messages.
3346 (gen_insn): Save filename.
3348 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
3350 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
3351 -fno-local-ivars and -fivar-visibility.
3352 * c-family/c.opt: Make -Wshadow also implicitly enable
3355 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3357 * doc/tm.texi: Remove reference to deleted macro.
3358 * doc/tm.texi.in: Likewise.
3360 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
3363 * config/avr/avr.c (avr_out_store_psi): Use correct constant
3366 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
3369 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
3370 * config/arm/aout.h (License): Same.
3371 * config/arm/bpabi.h (License): Same.
3372 * config/arm/elf.h (License): Same.
3373 * config/arm/linux-elf.h (License): Same.
3374 * config/arm/linux-gas.h (License): Same.
3375 * config/arm/netbsd-elf.h (License): Same.
3376 * config/arm/uclinux-eabi.h (License): Same.
3377 * config/arm/uclinux-elf.h (License): Same.
3378 * config/arm/vxworks.h (License): Same.
3380 2014-05-11 Jakub Jelinek <jakub@redhat.com>
3382 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
3383 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
3384 number of operands to 3.
3385 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
3386 * tree-nested.c (convert_nonlocal_omp_clauses,
3387 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
3388 * gimplify.c (gimplify_scan_omp_clauses): Handle
3389 OMP_CLAUSE_LINEAR_STMT.
3390 * omp-low.c (lower_rec_input_clauses): Fix typo.
3391 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
3392 cast between Fortran boolean_type_node and C _Bool if
3395 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
3397 PR tree-optimization/61136
3398 * wide-int.h (multiple_of_p): Define a version that doesn't return
3400 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
3401 integer_zerop/const_binop pair.
3402 (multiple_of_p): Likewise, converting both operands to widest_int
3405 2014-05-09 Teresa Johnson <tejohnson@google.com>
3407 * cgraphunit.c (analyze_functions): Use correct dump file.
3409 2014-05-09 Florian Weimer <fweimer@redhat.com>
3411 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
3413 (stack_protect_return_slot_p): New function.
3414 (expand_used_vars): Call stack_protect_decl_p and
3415 stack_protect_return_slot_p for -fstack-protector-strong.
3417 2014-05-09 David Wohlferd <LimeGreenSocks@yahoo.com>
3418 Andrew Haley <aph@redhat.com>
3419 Richard Sandiford <rdsandiford@googlemail.com>
3421 * doc/extend.texi: Rewrite inline asm page / re-org asm-related
3424 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
3427 * fold-const.c (fold_binary_loc): Changed width of mask.
3429 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3431 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
3432 unsigned int initializers for regno_in, regno_out.
3434 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3437 * config/avr/avr.md (cc): Add new attribute set_vzn.
3438 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
3439 Set cc insn attribute to set_vzn instead of set_zn for alternatives
3440 with INC, DEC or NEG.
3441 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
3442 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
3443 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
3445 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3448 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3450 * wide-int.cc (UTItype): Define.
3451 (UDWtype): Define for appropriate W_TYPE_SIZE.
3453 2014-05-09 Richard Biener <rguenther@suse.de>
3455 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
3456 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
3457 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
3458 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
3459 ssa_propagate): Adjust.
3461 2014-05-08 Jeff Law <law@redhat.com>
3463 PR tree-optimization/61009
3464 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
3465 tri-state rather than a boolean. When a block is too big to
3466 thread through, inform caller via negative return value.
3467 (thread_across_edge): If a block was too big for normal threading,
3468 then it's too big for a joiner too, so remove temporary equivalences
3469 and return immediately.
3471 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
3472 Matthias Klose <doko@ubuntu.com>
3475 * optc-gen.awk: Fix option handling for -Wunused-parameter.
3477 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3480 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
3482 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3485 * config/alpha/alpha.c: Include gimple-iterator.h.
3486 (alpha_gimple_fold_builtin): New function. Move
3487 ALPHA_BUILTIN_UMULH folding from ...
3488 (alpha_fold_builtin): ... here.
3489 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
3491 2014-05-08 Wei Mi <wmi@google.com>
3494 * config/i386/i386.c (ix86_compute_frame_layout): Update
3495 preferred_stack_boundary for call, expanded from tls descriptor.
3496 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
3497 to depend on SP register.
3498 (*tls_local_dynamic_base_32_gnu): Ditto.
3499 (*tls_local_dynamic_32_once): Ditto.
3500 (tls_global_dynamic_64_<mode>): Set
3501 ix86_tls_descriptor_calls_expanded_in_cfun.
3502 (tls_local_dynamic_base_64_<mode>): Ditto.
3503 (tls_global_dynamic_32): Set
3504 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
3505 to depend on SP register.
3506 (tls_local_dynamic_base_32): Ditto.
3508 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3510 * config/arm/arm_neon.h: Update comment.
3511 * config/arm/neon-docgen.ml: Delete.
3512 * config/arm/neon-gen.ml: Delete.
3513 * doc/arm-neon-intrinsics.texi: Update comment.
3515 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3517 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
3519 (vand, vorr, veor, vorn, vbic): Remove.
3520 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
3522 (neon_vsub_unspec): Likewise.
3523 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
3525 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3527 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
3528 (vadd_s16): Likewise.
3529 (vadd_s32): Likewise.
3530 (vadd_f32): Likewise.
3531 (vadd_u8): Likewise.
3532 (vadd_u16): Likewise.
3533 (vadd_u32): Likewise.
3534 (vadd_s64): Likewise.
3535 (vadd_u64): Likewise.
3536 (vaddq_s8): Likewise.
3537 (vaddq_s16): Likewise.
3538 (vaddq_s32): Likewise.
3539 (vaddq_s64): Likewise.
3540 (vaddq_f32): Likewise.
3541 (vaddq_u8): Likewise.
3542 (vaddq_u16): Likewise.
3543 (vaddq_u32): Likewise.
3544 (vaddq_u64): Likewise.
3545 (vmul_s8): Likewise.
3546 (vmul_s16): Likewise.
3547 (vmul_s32): Likewise.
3548 (vmul_f32): Likewise.
3549 (vmul_u8): Likewise.
3550 (vmul_u16): Likewise.
3551 (vmul_u32): Likewise.
3552 (vmul_p8): Likewise.
3553 (vmulq_s8): Likewise.
3554 (vmulq_s16): Likewise.
3555 (vmulq_s32): Likewise.
3556 (vmulq_f32): Likewise.
3557 (vmulq_u8): Likewise.
3558 (vmulq_u16): Likewise.
3559 (vmulq_u32): Likewise.
3560 (vsub_s8): Likewise.
3561 (vsub_s16): Likewise.
3562 (vsub_s32): Likewise.
3563 (vsub_f32): Likewise.
3564 (vsub_u8): Likewise.
3565 (vsub_u16): Likewise.
3566 (vsub_u32): Likewise.
3567 (vsub_s64): Likewise.
3568 (vsub_u64): Likewise.
3569 (vsubq_s8): Likewise.
3570 (vsubq_s16): Likewise.
3571 (vsubq_s32): Likewise.
3572 (vsubq_s64): Likewise.
3573 (vsubq_f32): Likewise.
3574 (vsubq_u8): Likewise.
3575 (vsubq_u16): Likewise.
3576 (vsubq_u32): Likewise.
3577 (vsubq_u64): Likewise.
3578 (vand_s8): Likewise.
3579 (vand_s16): Likewise.
3580 (vand_s32): Likewise.
3581 (vand_u8): Likewise.
3582 (vand_u16): Likewise.
3583 (vand_u32): Likewise.
3584 (vand_s64): Likewise.
3585 (vand_u64): Likewise.
3586 (vandq_s8): Likewise.
3587 (vandq_s16): Likewise.
3588 (vandq_s32): Likewise.
3589 (vandq_s64): Likewise.
3590 (vandq_u8): Likewise.
3591 (vandq_u16): Likewise.
3592 (vandq_u32): Likewise.
3593 (vandq_u64): Likewise.
3594 (vorr_s8): Likewise.
3595 (vorr_s16): Likewise.
3596 (vorr_s32): Likewise.
3597 (vorr_u8): Likewise.
3598 (vorr_u16): Likewise.
3599 (vorr_u32): Likewise.
3600 (vorr_s64): Likewise.
3601 (vorr_u64): Likewise.
3602 (vorrq_s8): Likewise.
3603 (vorrq_s16): Likewise.
3604 (vorrq_s32): Likewise.
3605 (vorrq_s64): Likewise.
3606 (vorrq_u8): Likewise.
3607 (vorrq_u16): Likewise.
3608 (vorrq_u32): Likewise.
3609 (vorrq_u64): Likewise.
3610 (veor_s8): Likewise.
3611 (veor_s16): Likewise.
3612 (veor_s32): Likewise.
3613 (veor_u8): Likewise.
3614 (veor_u16): Likewise.
3615 (veor_u32): Likewise.
3616 (veor_s64): Likewise.
3617 (veor_u64): Likewise.
3618 (veorq_s8): Likewise.
3619 (veorq_s16): Likewise.
3620 (veorq_s32): Likewise.
3621 (veorq_s64): Likewise.
3622 (veorq_u8): Likewise.
3623 (veorq_u16): Likewise.
3624 (veorq_u32): Likewise.
3625 (veorq_u64): Likewise.
3626 (vbic_s8): Likewise.
3627 (vbic_s16): Likewise.
3628 (vbic_s32): Likewise.
3629 (vbic_u8): Likewise.
3630 (vbic_u16): Likewise.
3631 (vbic_u32): Likewise.
3632 (vbic_s64): Likewise.
3633 (vbic_u64): Likewise.
3634 (vbicq_s8): Likewise.
3635 (vbicq_s16): Likewise.
3636 (vbicq_s32): Likewise.
3637 (vbicq_s64): Likewise.
3638 (vbicq_u8): Likewise.
3639 (vbicq_u16): Likewise.
3640 (vbicq_u32): Likewise.
3641 (vbicq_u64): Likewise.
3642 (vorn_s8): Likewise.
3643 (vorn_s16): Likewise.
3644 (vorn_s32): Likewise.
3645 (vorn_u8): Likewise.
3646 (vorn_u16): Likewise.
3647 (vorn_u32): Likewise.
3648 (vorn_s64): Likewise.
3649 (vorn_u64): Likewise.
3650 (vornq_s8): Likewise.
3651 (vornq_s16): Likewise.
3652 (vornq_s32): Likewise.
3653 (vornq_s64): Likewise.
3654 (vornq_u8): Likewise.
3655 (vornq_u16): Likewise.
3656 (vornq_u32): Likewise.
3657 (vornq_u64): Likewise.
3659 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3661 * wide-int.cc (UTItype): Define.
3662 (UDWtype): Define for appropriate W_TYPE_SIZE.
3664 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
3666 PR tree-optimization/59100
3667 * tree-ssa-phiopt.c: Include tree-inline.h.
3668 (neutral_element_p, absorbing_element_p): New functions.
3669 (value_replacement): Handle conditional binary operations with a
3670 neutral or absorbing element.
3672 2014-05-08 Richard Biener <rguenther@suse.de>
3674 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
3675 folding the expression.
3676 (valueize_expr): Remove.
3677 (visit_reference_op_load): Do not valueize the result of
3679 (simplify_binary_expression): Likewise.
3680 (simplify_unary_expression): Likewise.
3682 2014-05-08 Richard Biener <rguenther@suse.de>
3684 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
3685 looking at TYPE_ARG_TYPES.
3687 2014-05-08 Richard Biener <rguenther@suse.de>
3689 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
3690 pointer propagation special-case.
3692 2014-05-08 Bin Cheng <bin.cheng@arm.com>
3694 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
3695 core part of address expressions.
3697 2014-05-08 Alan Modra <amodra@gmail.com>
3700 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
3701 loads and stores when -mno-strict-align at any alignment.
3702 (expand_block_clear): Similarly. Also correct calculation of
3705 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
3708 * tree-complex.c (expand_complex_move): Keep line info when expanding
3710 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
3711 of complex expression. Use new argument to display correct location
3712 for values coming from phi statement.
3713 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
3714 (warn_uninitialized_phi): Pass location of phi argument to
3716 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
3717 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
3719 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
3721 * config/rs6000/predicates.md (indexed_address_mem): New.
3722 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
3723 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
3724 fpstore_ux, fpstore_u.
3725 (sign_extend, indexed, update): New.
3726 (cell_micro): Adjust.
3727 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
3728 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
3729 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
3730 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
3731 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
3732 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
3733 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
3734 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
3735 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
3736 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
3737 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
3738 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
3739 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
3740 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
3741 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
3743 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
3744 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
3745 *vsx_extract_<mode>_store): Adjust.
3746 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
3747 is_cracked_insn, insn_must_be_first_in_group,
3748 insn_must_be_last_in_group): Adjust.
3750 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
3752 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
3753 ppc440-fpstore): Adjust.
3754 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
3755 ppc476-fpstore): Adjust.
3756 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
3757 ppc601-fpstore): Adjust.
3758 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
3760 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
3762 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
3763 ppc7450-fpstore): Adjust.
3764 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
3765 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
3766 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
3768 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
3769 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
3770 cell-fpstore, cell-fpstore-update): Adjust.
3771 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
3772 ppce300c3_store, ppce300c3_fpstore): Adjust.
3773 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
3774 e500mc_fpstore): Adjust.
3775 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
3776 e500mc64_store, e500mc64_fpstore): Adjust.
3777 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
3778 e5500_fpstore): Adjust.
3779 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
3780 e6500_fpstore): Adjust.
3781 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
3783 * config/rs6000/power4.md (power4-load, power4-load-ext,
3784 power4-load-ext-update, power4-load-ext-update-indexed,
3785 power4-load-update-indexed, power4-load-update, power4-fpload,
3786 power4-fpload-update, power4-store, power4-store-update,
3787 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
3789 * config/rs6000/power5.md (power5-load, power5-load-ext,
3790 power5-load-ext-update, power5-load-ext-update-indexed,
3791 power5-load-update-indexed, power5-load-update, power5-fpload,
3792 power5-fpload-update, power5-store, power5-store-update,
3793 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
3795 * config/rs6000/power6.md (power6-load, power6-load-ext,
3796 power6-load-update, power6-load-update-indexed,
3797 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
3798 power6-fpload-update, power6-store, power6-store-update,
3799 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
3801 * config/rs6000/power7.md (power7-load, power7-load-ext,
3802 power7-load-update, power7-load-update-indexed,
3803 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
3804 power7-fpload-update, power7-store, power7-store-update,
3805 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
3807 * config/rs6000/power8.md (power8-load, power8-load-update,
3808 power8-load-ext, power8-load-ext-update, power8-fpload,
3809 power8-fpload-update, power8-store, power8-store-update-indexed,
3810 power8-fpstore, power8-fpstore-update): Adjust.
3811 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
3813 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
3814 titan_lsu_store, titan_lsu_fpstore): Adjust.
3815 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
3817 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
3820 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
3821 unrolled byte insns. Emit address increments after move insns.
3823 2014-05-07 David Malcolm <dmalcolm@redhat.com>
3825 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
3826 const_gimple, rather than a gimple.
3827 (gimple_call_builtin_p): Likewise, for the three variants.
3829 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
3830 (gimple_call_builtin_p): Likewise, for the three variants.
3832 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3834 PR tree-optimization/61095
3835 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
3837 2014-05-07 Richard Biener <rguenther@suse.de>
3839 PR tree-optimization/61034
3840 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
3841 (maybe_skip_until): Use translate to take into account
3842 lattices when trying to do disambiguations.
3843 (get_continuation_for_phi_1): Likewise.
3844 (get_continuation_for_phi): Adjust for added translate arguments.
3845 (walk_non_aliased_vuses): Likewise.
3846 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
3847 (walk_non_aliased_vuses): Likewise.
3848 (call_may_clobber_ref_p_1): Declare.
3849 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
3850 calls. Stop early if we are only supposed to disambiguate.
3851 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
3853 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
3855 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
3856 Emit an error when the function has arguments.
3858 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3860 * cfgloop.h (unswitch_loops): Remove.
3861 * doc/passes.texi: Remove references to loop-unswitch.c
3862 * timevar.def (TV_LOOP_UNSWITCH): Remove.
3864 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
3866 * tree-vect-data-refs.c (vect_grouped_load_supported): New
3867 check for loads group of length 3.
3868 (vect_permute_load_chain): New permutations for loads group of
3870 * tree-vect-stmts.c (vect_model_load_cost): Change cost
3871 of vec_perm_shuffle for the new permutations.
3873 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
3875 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
3876 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
3877 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
3878 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
3879 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
3880 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
3881 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
3882 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
3884 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3886 * loop-unswitch.c: Delete.
3888 2014-05-07 Richard Biener <rguenther@suse.de>
3890 * config.gcc: Always set need_64bit_hwint to yes.
3892 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
3894 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
3895 of using optimize_size.
3897 2014-05-06 Mike Stump <mikestump@comcast.net>
3899 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
3901 2014-05-06 Joseph Myers <joseph@codesourcery.com>
3903 * config/i386/sse.md (*mov<mode>_internal)
3904 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
3905 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
3906 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
3907 (*<code><mode>3, *andnot<mode>3<mask_name>)
3908 (<mask_codefor><code><mode>3<mask_name>): Only consider
3909 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
3911 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3914 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3916 * lra-constraints.c (valid_address_p): Move earlier in file.
3917 Add a constraint argument to the address_info version.
3918 (satisfies_memory_constraint_p): New function.
3919 (satisfies_address_constraint_p): Likewise.
3920 (process_alt_operands, curr_insn_transform): Use them.
3921 (process_address): Pass the constraint to valid_address_p when
3922 checking address operands.
3924 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
3926 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
3927 to their respective blocks. Fix inadvertent use of "node".
3929 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3931 * emit-rtl.c (init_derived_machine_modes): New functionm, split
3933 (init_emit_once): ...here.
3934 * rtl.h (init_derived_machine_modes): Declare.
3935 * toplev.c (do_compile): Call it even if no_backend.
3937 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
3938 Mike Stump <mikestump@comcast.net>
3939 Richard Sandiford <rdsandiford@googlemail.com>
3940 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3942 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
3943 (rtx_equal_for_memref_p): Update comment.
3944 (adjust_offset_for_component_ref): Use wide-int interfaces.
3945 * builtins.c (get_object_alignment_2): Likewise.
3946 (c_readstr): Likewise.
3947 (target_char_cast): Add comment.
3948 (determine_block_size): Use wide-int interfaces.
3949 (expand_builtin_signbit): Likewise.
3950 (fold_builtin_int_roundingfn): Likewise.
3951 (fold_builtin_bitop): Likewise.
3952 (fold_builtin_bswap): Likewise.
3953 (fold_builtin_logarithm): Use signop.
3954 (fold_builtin_pow): Likewise.
3955 (fold_builtin_memory_op): Use wide-int interfaces.
3956 (fold_builtin_object_size): Likewise.
3957 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
3958 nb_iterations_estimate.
3959 (record_niter_bound): Use wide-int interfaces.
3960 (get_estimated_loop_iterations_int): Likewise.
3961 (get_estimated_loop_iterations): Likewise.
3962 (get_max_loop_iterations): Likewise.
3963 * cfgloop.h: Include wide-int.h.
3964 (struct nb_iter_bound): Change bound to widest_int.
3965 (struct loop): Change nb_iterations_upper_bound and
3966 nb_iterations_estimate to widest_int.
3967 (record_niter_bound): Switch to use widest_int.
3968 (get_estimated_loop_iterations): Likewise.
3969 (get_max_loop_iterations): Likewise.
3970 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
3971 update for wide-int.
3972 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
3973 * combine.c (try_combine): Likewise.
3974 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
3975 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
3977 (aarch64_float_const_representable_p): Likewise.
3978 * config/arc/arc.c: Include wide-int.h.
3979 (arc_can_use_doloop_p): Use wide-int interfaces.
3980 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
3981 (vfp3_const_double_index): Likewise.
3982 * config/avr/avr.c (avr_out_round): Likewise.
3983 (avr_fold_builtin): Likewise.
3984 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
3985 (bfin_can_use_doloop_p): Likewise.
3986 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
3987 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
3988 * config/i386/i386.c: Include wide-int.h.
3989 (ix86_data_alignment): Use wide-int interfaces.
3990 (ix86_local_alignment): Likewise.
3991 (ix86_emit_swsqrtsf): Update real_from_integer.
3992 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
3993 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
3994 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
3995 (zero_constant): Likewise.
3996 (input_operand): Likewise.
3997 (splat_input_operand): Likewise.
3998 (non_logical_cint_operand): Change const_double to const_wide_int.
3999 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
4000 (easy_altivec_constant): Remove comment.
4001 (paired_expand_vector_init): Use CONSTANT_P.
4002 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
4003 (rs6000_emit_move): Update checks.
4004 (rs6000_aggregate_candidate): Use wide-int interfaces.
4005 (rs6000_expand_ternop_builtin): Likewise.
4006 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
4007 (rs6000_assemble_integer): Likewise.
4008 (rs6000_hash_constant): Likewise.
4009 (output_toc): Likewise.
4010 (rs6000_rtx_costs): Likewise.
4011 (rs6000_emit_swrsqrt); Update call to real_from_integer.
4012 * config/rs6000/rs6000-c.c: Include wide-int.h.
4013 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
4014 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
4015 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
4016 Handle CONST_WIDE_INT.
4017 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
4018 Use tree_fits_uhwi_p.
4019 * config/sparc/sparc.c: Include wide-int.h.
4020 (sparc_fold_builtin): Use wide-int interfaces.
4021 * config/vax/vax.c: Include wide-int.h.
4022 (vax_float_literal): Use real_from_integer.
4023 * coretypes.h (struct hwivec_def): New.
4025 (const_hwivec): New.
4026 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
4027 (equiv_constant): Handle CONST_WIDE_INT.
4028 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
4029 (cselib_hash_rtx): Handle CONST_WIDE_INT.
4030 * dbxout.c (stabstr_U): Use wide-int interfaces.
4031 (dbxout_type): Update to use cst_fits_shwi_p.
4032 * defaults.h (LOG2_BITS_PER_UNIT): Define.
4033 (TARGET_SUPPORTS_WIDE_INT): Add default.
4034 * dfp.c: Include wide-int.h.
4035 (decimal_real_to_integer2): Use wide-int interfaces and rename to
4036 decimal_real_to_integer.
4037 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
4038 decimal_real_to_integer.
4039 * doc/generic.texi (Constant expressions): Update for wide_int.
4040 * doc/rtl.texi (const_double): Likewise.
4041 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
4042 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
4043 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
4044 (REAL_VALUE_FROM_INT): Remove.
4045 (TARGET_SUPPORTS_WIDE_INT): New.
4046 * doc/tm.texi: Regenerate.
4047 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
4048 * double-int.h: Include wide-int.h.
4049 (struct wi::int_traits): New.
4050 * dwarf2out.c (get_full_len): New.
4051 (dw_val_equal_p): Add case dw_val_class_wide_int.
4052 (size_of_loc_descr): Likewise.
4053 (output_loc_operands): Likewise.
4054 (insert_double): Remove.
4055 (insert_wide_int): New.
4057 (print_die): Add case dw_val_class_wide_int.
4058 (attr_checksum): Likewise.
4059 (attr_checksum_ordered): Likewise.
4060 (same_dw_val_p): Likewise.
4061 (size_of_die): Likewise.
4062 (value_format): Likewise.
4063 (output_die): Likewise.
4064 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
4066 (clz_loc_descriptor): Use wide-int interfaces.
4067 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
4068 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
4069 (round_up_to_align): Use wide-int interfaces.
4070 (field_byte_offset): Likewise.
4071 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
4072 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
4073 CONST_DOUBLE handling. Use wide-int interfaces.
4074 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
4075 (gen_enumeration_type_die): Use add_AT_wide.
4076 (hash_loc_operands): Add case dw_val_class_wide_int.
4077 (compare_loc_operands): Likewise.
4078 * dwarf2out.h: Include wide-int.h.
4079 (wide_int_ptr): New.
4080 (enum dw_val_class): Add dw_val_class_wide_int.
4081 (struct dw_val_struct): Add val_wide.
4082 * emit-rtl.c (const_wide_int_htab): New.
4083 (const_wide_int_htab_hash): New.
4084 (const_wide_int_htab_eq): New.
4085 (lookup_const_wide_int): New.
4086 (const_double_htab_hash): Use wide-int interfaces.
4087 (const_double_htab_eq): Likewise.
4088 (rtx_to_double_int): Conditionally compile for wide-int.
4089 (immed_double_int_const): Rename to immed_wide_int_const and
4090 update for wide-int.
4091 (immed_double_const): Conditionally compile for wide-int.
4092 (init_emit_once): Use wide-int interfaces.
4093 * explow.c (plus_constant): Likewise.
4094 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
4095 (lshift_value): Use wide-int interfaces.
4096 (expand_mult): Likewise.
4097 (choose_multiplier): Likewise.
4098 (expand_smod_pow2): Likewise.
4099 (make_tree): Likewise.
4100 * expr.c (convert_modes): Consolidate handling of constants.
4101 Use wide-int interfaces.
4102 (emit_group_load_1): Add note.
4103 (store_expr): Update comment.
4104 (get_inner_reference): Use wide-int interfaces.
4105 (expand_constructor): Update comment.
4106 (expand_expr_real_2): Use wide-int interfaces.
4107 (expand_expr_real_1): Likewise.
4108 (reduce_to_bit_field_precision): Likewise.
4109 (const_vector_from_tree): Likewise.
4110 * final.c: Include wide-int-print.h.
4111 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
4112 * fixed-value.c: Include wide-int.h.
4113 (fixed_from_string): Use wide-int interfaces.
4114 (fixed_to_decimal): Likewise.
4115 (fixed_convert_from_real): Likewise.
4116 (real_convert_from_fixed): Likewise.
4117 * fold-const.h (mem_ref_offset): Return an offset_int.
4118 (div_if_zero_remainder): Remove code parameter.
4119 * fold-const.c (div_if_zero_remainder): Remove code parameter.
4120 Use wide-int interfaces.
4121 (may_negate_without_overflow_p): Use wide-int interfaces.
4122 (negate_expr_p): Likewise.
4123 (fold_negate_expr): Likewise.
4124 (int_const_binop_1): Likewise.
4125 (const_binop): Likewise.
4126 (fold_convert_const_int_from_int): Likewise.
4127 (fold_convert_const_int_from_real): Likewise.
4128 (fold_convert_const_int_from_fixed): Likewise.
4129 (fold_convert_const_fixed_from_int): Likewise.
4130 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
4131 (sign_bit_p): Use wide-int interfaces.
4132 (make_range_step): Likewise.
4133 (build_range_check): Likewise. Pass an integer of the correct type
4134 instead of using integer_one_node.
4135 (range_predecessor): Pass an integer of the correct type instead
4136 of using integer_one_node.
4137 (range_successor): Likewise.
4138 (merge_ranges): Likewise.
4139 (unextend): Use wide-int interfaces.
4140 (extract_muldiv_1): Likewise.
4141 (fold_div_compare): Likewise.
4142 (fold_single_bit_test): Likewise.
4143 (fold_sign_changed_comparison): Likewise.
4144 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
4145 (fold_plusminus_mult_expr): Use wide-int interfaces.
4146 (native_encode_int): Likewise.
4147 (native_interpret_int): Likewise.
4148 (fold_unary_loc): Likewise.
4149 (pointer_may_wrap_p): Likewise.
4150 (size_low_cst): Likewise.
4151 (mask_with_tz): Likewise.
4152 (fold_binary_loc): Likewise.
4153 (fold_ternary_loc): Likewise.
4154 (multiple_of_p): Likewise.
4155 (tree_call_nonnegative_warnv_p): Update calls to
4156 tree_int_cst_min_precision and real_from_integer.
4157 (fold_negate_const): Use wide-int interfaces.
4158 (fold_abs_const): Likewise.
4159 (fold_relational_const): Use tree_int_cst_lt.
4160 (round_up_loc): Use wide-int interfaces.
4161 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
4162 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
4163 * gengtype.c: Remove include of double-int.h.
4164 (do_typedef): Use wide-int interfaces.
4165 (open_base_files): Add wide-int.h.
4166 (main): Add offset_int and widest_int typedefs.
4167 * gengtype-lex.l: Handle "^".
4168 (CXX_KEYWORD): Add "static".
4169 * gengtype-parse.c (require3): New.
4170 (require_template_declaration): Handle constant template arguments
4171 and nested templates.
4172 * gengtype-state.c: Don't include "double-int.h".
4173 * genpreds.c (write_one_predicate_function): Update comment.
4174 (write_tm_constrs_h): Add check for hval and lval use in
4176 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
4177 (add_to_sequence): Likewise.
4178 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
4179 and const_double_operand.
4180 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
4182 * gimple-fold.c (get_base_constructor): Likewise.
4183 (fold_array_ctor_reference): Likewise.
4184 (fold_nonarray_ctor_reference): Likewise.
4185 (fold_const_aggregate_ref_1): Likewise.
4186 (gimple_val_nonnegative_real_p): Likewise.
4187 (gimple_fold_indirect_ref): Likewise.
4188 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
4189 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
4190 (struct slsr_cand_d): Change index to be widest_int.
4191 (struct incr_info_d): Change incr to be widest_int.
4192 (alloc_cand_and_find_basis): Use wide-int interfaces.
4193 (slsr_process_phi): Likewise.
4194 (backtrace_base_for_ref): Likewise. Return a widest_int.
4195 (restructure_reference): Take a widest_int instead of a double_int.
4196 (slsr_process_ref): Use wide-int interfaces.
4197 (create_mul_ssa_cand): Likewise.
4198 (create_mul_imm_cand): Likewise.
4199 (create_add_ssa_cand): Likewise.
4200 (create_add_imm_cand): Take a widest_int instead of a double_int.
4201 (slsr_process_add): Use wide-int interfaces.
4202 (slsr_process_cast): Likewise.
4203 (slsr_process_copy): Likewise.
4204 (dump_candidate): Likewise.
4205 (dump_incr_vec): Likewise.
4206 (replace_ref): Likewise.
4207 (cand_increment): Likewise. Return a widest_int.
4208 (cand_abs_increment): Likewise.
4209 (replace_mult_candidate): Take a widest_int instead of a double_int.
4210 (replace_unconditional_candidate): Use wide-int interfaces.
4211 (incr_vec_index): Take a widest_int instead of a double_int.
4212 (create_add_on_incoming_edge): Likewise.
4213 (create_phi_basis): Use wide-int interfaces.
4214 (replace_conditional_candidate): Likewise.
4215 (record_increment): Take a widest_int instead of a double_int.
4216 (record_phi_increments): Use wide-int interfaces.
4217 (phi_incr_cost): Take a widest_int instead of a double_int.
4218 (lowest_cost_path): Likewise.
4219 (total_savings): Likewise.
4220 (analyze_increments): Use wide-int interfaces.
4221 (ncd_with_phi): Take a widest_int instead of a double_int.
4222 (ncd_of_cand_and_phis): Likewise.
4223 (nearest_common_dominator_for_cands): Likewise.
4224 (insert_initializers): Use wide-int interfaces.
4225 (all_phi_incrs_profitable): Likewise.
4226 (replace_one_candidate): Likewise.
4227 (replace_profitable_candidates): Likewise.
4228 * godump.c: Include wide-int-print.h.
4229 (go_output_typedef): Use wide-int interfaces.
4230 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
4231 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
4232 (build_loop_iteration_domains): Likewise.
4233 * hooks.h: Include wide-int.h rather than double-int.h.
4234 (hook_bool_dint_dint_uint_bool_true): Delete.
4235 (hook_bool_wint_wint_uint_bool_true): Declare.
4236 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
4237 (hook_bool_wint_wint_uint_bool_true): New.
4238 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
4240 (ubsan_expand_si_overflow_mul_check): Likewise.
4241 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
4242 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
4243 (get_ancestor_addr_info): Likewise.
4244 (ipa_modify_call_arguments): Likewise.
4245 * loop-doloop.c (doloop_modify): Likewise.
4246 (doloop_optimize): Likewise.
4247 * loop-iv.c (iv_number_of_iterations): Likewise.
4248 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
4249 (unroll_loop_constant_iterations): Likewise.
4250 (decide_unroll_runtime_iterations): Likewise.
4251 (unroll_loop_runtime_iterations): Likewise.
4252 (decide_peel_simple): Likewise.
4253 (decide_unroll_stupid): Likewise.
4254 * lto-streamer-in.c (streamer_read_wi): Add.
4255 (input_cfg): Use wide-int interfaces.
4256 (lto_input_tree_1): Likewise.
4257 * lto-streamer-out.c (streamer_write_wi): Add.
4258 (hash_tree): Use wide-int interfaces.
4259 (output_cfg): Likewise.
4260 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
4261 (GTFILES): Add wide-int.h and signop.h.
4262 (TAGS): Look for .cc files too.
4263 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
4264 * optabs.c (expand_subword_shift): Likewise.
4265 (expand_doubleword_shift): Likewise.
4266 (expand_absneg_bit): Likewise.
4267 (expand_copysign_absneg): Likewise.
4268 (expand_copysign_bit): Likewise.
4269 * postreload.c (reload_cse_simplify_set): Likewise.
4270 * predict.c (predict_iv_comparison): Likewise.
4271 * pretty-print.h: Include wide-int-print.h.
4273 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
4274 * print-tree.c: Include wide-int-print.h.
4275 (print_node_brief): Use wide-int interfaces.
4276 (print_node): Likewise.
4277 * read-rtl.c (validate_const_wide_int): New.
4278 (read_rtx_code): Add CONST_WIDE_INT case.
4279 * real.c: Include wide-int.h.
4280 (real_to_integer2): Delete.
4281 (real_to_integer): New function, returning a wide_int.
4282 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
4283 (ten_to_ptwo): Update call to real_from_integer.
4284 (real_digit): Likewise.
4285 * real.h: Include signop.h, wide-int.h and insn-modes.h.
4286 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
4287 (REAL_VALUE_TO_INT): Delete.
4288 (real_to_integer): Declare a wide-int form.
4289 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
4290 * recog.c (const_int_operand): Improve comment.
4291 (const_scalar_int_operand): New.
4292 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
4293 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
4294 (split_double): Likewise.
4295 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
4296 (rtx_size): Likewise.
4297 (rtx_alloc_stat_v): New.
4298 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
4299 (cwi_output_hex): New.
4300 (iterative_hash_rtx): Handle CONST_WIDE_INT.
4301 (cwi_check_failed_bounds): New.
4302 * rtl.def (CONST_WIDE_INT): New.
4303 * rtl.h: Include <utility> and wide-int.h.
4304 (struct hwivec_def): New.
4305 (CWI_GET_NUM_ELEM): New.
4306 (CWI_PUT_NUM_ELEM): New.
4307 (struct rtx_def): Add num_elem and hwiv.
4308 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
4309 (CASE_CONST_UNIQUE): Likewise.
4310 (CASE_CONST_ANY): Likewise.
4311 (CONST_SCALAR_INT_P): Likewise.
4312 (CONST_WIDE_INT_P): New.
4314 (HWIVEC_CHECK): New.
4315 (cwi_check_failed_bounds): New.
4317 (HWIVEC_CHECK): New.
4318 (CONST_WIDE_INT_VEC) New.
4319 (CONST_WIDE_INT_NUNITS) New.
4320 (CONST_WIDE_INT_ELT) New.
4321 (rtx_mode_t): New type.
4322 (wi::int_traits <rtx_mode_t>): New.
4324 (wi::min_value): New.
4325 (wi::max_value): New.
4327 (const_wide_int_alloc): New.
4328 (immed_wide_int_const): New.
4329 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
4330 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
4331 * signop.h: New file.
4332 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
4333 (simplify_const_unary_operation): Use wide-int interfaces.
4334 (simplify_binary_operation_1): Likewise.
4335 (simplify_const_binary_operation): Likewise.
4336 (simplify_const_relational_operation): Likewise.
4337 (simplify_immed_subreg): Likewise.
4338 * stmt.c (expand_case): Likewise.
4339 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
4340 signop rather than a bool.
4341 * stor-layout.c (layout_type): Use wide-int interfaces.
4342 (initialize_sizetypes): Update calls to
4343 set_min_and_max_values_for_integral_type.
4344 (set_min_and_max_values_for_integral_type): Take a signop rather
4345 than a bool. Use wide-int interfaces.
4346 (fixup_signed_type): Update accordingly. Remove
4347 HOST_BITS_PER_DOUBLE_INT limit.
4348 (fixup_unsigned_type): Likewise.
4349 * system.h (STATIC_CONSTANT_P): New.
4350 (STATIC_ASSERT): New.
4351 * target.def (can_use_doloop_p): Take widest_ints rather than
4353 * target.h: Include wide-int.h rather than double-int.h.
4354 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
4356 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
4357 rather than INT_CST_LT_UNSIGNED.
4358 (can_use_doloop_if_innermost): Take widest_ints rather than
4360 * tree-affine.c: Include wide-int-print.h.
4361 (double_int_ext_for_comb): Delete.
4362 (wide_int_ext_for_comb): New.
4363 (aff_combination_zero): Use wide-int interfaces.
4364 (aff_combination_const): Take a widest_int instead of a double_int.
4365 (aff_combination_elt): Use wide-int interfaces.
4366 (aff_combination_scale): Take a widest_int instead of a double_int.
4367 (aff_combination_add_elt): Likewise.
4368 (aff_combination_add_cst): Likewise.
4369 (aff_combination_add): Use wide-int interfaces.
4370 (aff_combination_convert): Likewise.
4371 (tree_to_aff_combination): Likewise.
4372 (add_elt_to_tree): Take a widest_int instead of a double_int.
4373 (aff_combination_to_tree): Use wide-int interfaces.
4374 (aff_combination_remove_elt): Likewise.
4375 (aff_combination_add_product): Take a widest_int instead of
4377 (aff_combination_mult): Use wide-int interfaces.
4378 (aff_combination_expand): Likewise.
4379 (double_int_constant_multiple_p): Delete.
4380 (wide_int_constant_multiple_p): New.
4381 (aff_combination_constant_multiple_p): Take a widest_int pointer
4382 instead of a double_int pointer.
4383 (print_aff): Use wide-int interfaces.
4384 (get_inner_reference_aff): Take a widest_int pointer
4385 instead of a double_int pointer.
4386 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
4387 * tree-affine.h: Include wide-int.h.
4388 (struct aff_comb_elt): Change type of coef to widest_int.
4389 (struct affine_tree_combination): Change type of offset to widest_int.
4390 (double_int_ext_for_comb): Delete.
4391 (wide_int_ext_for_comb): New.
4392 (aff_combination_const): Use widest_int instead of double_int.
4393 (aff_combination_scale): Likewise.
4394 (aff_combination_add_elt): Likewise.
4395 (aff_combination_constant_multiple_p): Likewise.
4396 (get_inner_reference_aff): Likewise.
4397 (aff_comb_cannot_overlap_p): Likewise.
4398 (aff_combination_zero_p): Use wide-int interfaces.
4399 * tree.c: Include tree.h.
4400 (init_ttree): Use make_int_cst.
4401 (tree_code_size): Removed code for INTEGER_CST case.
4402 (tree_size): Add INTEGER_CST case.
4403 (make_node_stat): Update comment.
4404 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
4405 (build_int_cst_type): Use wide-int interfaces.
4406 (double_int_to_tree): Likewise.
4407 (double_int_fits_to_tree_p): Delete.
4408 (force_fit_type_double): Delete.
4409 (force_fit_type): New.
4410 (int_cst_hash_hash): Use wide-int interfaces.
4411 (int_cst_hash_eq): Likewise.
4412 (build_int_cst_wide): Delete.
4413 (wide_int_to_tree): New.
4414 (cache_integer_cst): Use wide-int interfaces.
4415 (build_low_bits_mask): Likewise.
4416 (cst_and_fits_in_hwi): Likewise.
4417 (real_value_from_int_cst): Likewise.
4418 (make_int_cst_stat): New.
4419 (integer_zerop): Use wide_int interfaces.
4420 (integer_onep): Likewise.
4421 (integer_all_onesp): Likewise.
4422 (integer_pow2p): Likewise.
4423 (integer_nonzerop): Likewise.
4424 (tree_log2): Likewise.
4425 (tree_floor_log2): Likewise.
4426 (tree_ctz): Likewise.
4427 (int_size_in_bytes): Likewise.
4428 (mem_ref_offset): Return an offset_int rather than a double_int.
4429 (build_type_attribute_qual_variant): Use wide_int interfaces.
4430 (type_hash_eq): Likewise
4431 (tree_int_cst_equal): Likewise.
4432 (tree_int_cst_lt): Delete.
4433 (tree_int_cst_compare): Likewise.
4434 (tree_fits_shwi_p): Use wide_int interfaces.
4435 (tree_fits_uhwi_p): Likewise.
4436 (tree_int_cst_sign_bit): Likewise.
4437 (tree_int_cst_sgn): Likewise.
4438 (tree_int_cst_min_precision): Take a signop rather than a bool.
4439 (simple_cst_equal): Use wide_int interfaces.
4440 (compare_tree_int): Likewise.
4441 (iterative_hash_expr): Likewise.
4442 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
4444 (get_type_static_bounds): Use wide_int interfaces.
4445 (tree_int_cst_elt_check_failed): New.
4446 (build_common_tree_nodes): Reordered to set prec before filling in
4448 (int_cst_value): Check cst_and_fits_in_hwi.
4449 (widest_int_cst_value): Use wide_int interfaces.
4450 (upper_bound_in_type): Likewise.
4451 (lower_bound_in_type): Likewise.
4452 (num_ending_zeros): Likewise.
4453 (drop_tree_overflow): Likewise.
4454 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
4455 (gen_conditions_for_pow_cst_base): Likewise.
4456 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
4457 (group_case_labels_stmt): Use wide-int interfaces.
4458 (verify_gimple_assign_binary): Likewise.
4459 (print_loop): Likewise.
4460 * tree-chrec.c (tree_fold_binomial): Likewise.
4461 * tree-core.h (struct tree_base): Add int_length.
4462 (struct tree_int_cst): Change rep of value.
4463 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
4464 (dr_may_alias_p): Likewise.
4465 (max_stmt_executions_tree): Likewise.
4466 * tree.def (INTEGER_CST): Update comment.
4467 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
4468 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
4469 * tree-dump.c: Include wide-int.h and wide-int-print.h.
4470 (dequeue_and_dump): Use wide-int interfaces.
4471 * tree.h: Include wide-int.h.
4472 (NULL_TREE): Moved to earlier loc in file.
4473 (TREE_INT_CST_ELT_CHECK): New.
4474 (tree_int_cst_elt_check_failed): New.
4476 (TREE_INT_CST): Delete.
4477 (TREE_INT_CST_LOW): Use wide-int interfaces.
4478 (TREE_INT_CST_HIGH): Delete.
4479 (TREE_INT_CST_NUNITS): New.
4480 (TREE_INT_CST_EXT_NUNITS): Likewise.
4481 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
4482 (TREE_INT_CST_ELT): Likewise.
4483 (INT_CST_LT): Delete.
4484 (tree_int_cst_elt_check): New (two forms).
4485 (type_code_size): Update comment.
4486 (make_int_cst_stat, make_int_cst): New.
4487 (tree_to_double_int): Delete.
4488 (double_int_fits_to_tree_p): Delete.
4489 (force_fit_type_double): Delete.
4490 (build_int_cstu): Replace with out-of-line function.
4491 (build_int_cst_wide): Delete.
4492 (tree_int_cst_lt): Define inline.
4493 (tree_int_cst_le): New.
4494 (tree_int_cst_compare): Define inline.
4495 (tree_int_cst_min_precision): Take a signop rather than a bool.
4496 (wi::int_traits <const_tree>): New.
4497 (wi::int_traits <tree>): New.
4498 (wi::extended_tree): New.
4499 (wi::int_traits <wi::extended_tree>): New.
4500 (wi::to_widest): New.
4501 (wi::to_offset): New.
4502 (wi::fits_to_tree_p): New.
4503 (wi::min_value): New.
4504 (wi::max_value): New.
4505 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
4506 (copy_tree_body_r): Likewise.
4507 * tree-object-size.c (compute_object_offset): Likewise.
4508 (addr_object_size): Likewise.
4509 * tree-predcom.c: Include wide-int-print.h.
4510 (struct dref_d): Change type of offset to widest_int.
4511 (dump_dref): Call wide-int printer.
4512 (aff_combination_dr_offset): Use wide-int interfaces.
4513 (determine_offset): Take a widest_int pointer rather than a
4515 (split_data_refs_to_components): Use wide-int interfaces.
4516 (suitable_component_p): Likewise.
4517 (order_drefs): Likewise.
4518 (add_ref_to_chain): Likewise.
4519 (valid_initializer_p): Likewise.
4520 (determine_roots_comp): Likewise.
4521 * tree-pretty-print.c: Include wide-int-print.h.
4522 (dump_generic_node): Use wide-int interfaces.
4523 * tree-sra.c (sra_ipa_modify_expr): Likewise.
4524 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
4525 (move_fixed_address_to_symbol): Likewise.
4526 (move_hint_to_base): Likewise.
4527 (move_pointer_to_base): Likewise.
4528 (move_variant_to_index): Likewise.
4529 (most_expensive_mult_to_index): Likewise.
4530 (addr_to_parts): Likewise.
4531 (copy_ref_info): Likewise.
4532 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
4533 (indirect_refs_may_alias_p): Likewise.
4534 (stmt_kills_ref_p_1): Likewise.
4535 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
4536 * tree-ssa-ccp.c: Update comment at top of file. Include
4538 (struct prop_value_d): Change type of mask to widest_int.
4539 (extend_mask): New function.
4540 (dump_lattice_value): Use wide-int interfaces.
4541 (get_default_value): Likewise.
4542 (set_constant_value): Likewise.
4543 (set_value_varying): Likewise.
4544 (valid_lattice_transition): Likewise.
4545 (set_lattice_value): Likewise.
4546 (value_to_double_int): Delete.
4547 (value_to_wide_int): New.
4548 (get_value_from_alignment): Use wide-int interfaces.
4549 (get_value_for_expr): Likewise.
4550 (do_dbg_cnt): Likewise.
4551 (ccp_finalize): Likewise.
4552 (ccp_lattice_meet): Likewise.
4553 (bit_value_unop_1): Use widest_ints rather than double_ints.
4554 (bit_value_binop_1): Likewise.
4555 (bit_value_unop): Use wide-int interfaces.
4556 (bit_value_binop): Likewise.
4557 (bit_value_assume_aligned): Likewise.
4558 (evaluate_stmt): Likewise.
4559 (ccp_fold_stmt): Likewise.
4560 (visit_cond_stmt): Likewise.
4561 (ccp_visit_stmt): Likewise.
4562 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
4563 (constant_pointer_difference): Likewise.
4564 (associate_pointerplus): Likewise.
4565 (combine_conversions): Likewise.
4566 * tree-ssa-loop.h: Include wide-int.h.
4567 (struct tree_niter_desc): Change type of max to widest_int.
4568 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
4569 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
4570 (remove_redundant_iv_tests): Likewise.
4571 (canonicalize_loop_induction_variables): Likewise.
4572 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
4573 (constant_multiple_of): Take a widest_int pointer instead of
4574 a double_int pointer.
4575 (get_computation_aff): Use wide-int interfaces.
4576 (ptr_difference_cost): Likewise.
4577 (difference_cost): Likewise.
4578 (get_loop_invariant_expr_id): Likewise.
4579 (get_computation_cost_at): Likewise.
4580 (iv_elimination_compare_lt): Likewise.
4581 (may_eliminate_iv): Likewise.
4582 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
4583 instead of double_int.
4584 (max_loop_iterations): Likewise.
4585 (max_stmt_executions): Likewise.
4586 (estimated_stmt_executions): Likewise.
4587 * tree-ssa-loop-niter.c: Include wide-int-print.h.
4588 (split_to_var_and_offset): Use wide-int interfaces.
4589 (determine_value_range): Likewise.
4590 (bound_difference_of_offsetted_base): Likewise.
4591 (bounds_add): Take a widest_int instead of a double_int.
4592 (number_of_iterations_ne_max): Use wide-int interfaces.
4593 (number_of_iterations_ne): Likewise.
4594 (number_of_iterations_lt_to_ne): Likewise.
4595 (assert_loop_rolls_lt): Likewise.
4596 (number_of_iterations_lt): Likewise.
4597 (number_of_iterations_le): Likewise.
4598 (number_of_iterations_cond): Likewise.
4599 (number_of_iterations_exit): Likewise.
4600 (finite_loop_p): Likewise.
4601 (derive_constant_upper_bound_assign): Likewise.
4602 (derive_constant_upper_bound): Return a widest_int.
4603 (derive_constant_upper_bound_ops): Likewise.
4604 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
4605 (record_estimate): Take a widest_int rather than a double_int.
4606 (record_nonwrapping_iv): Use wide-int interfaces.
4607 (double_int_cmp): Delete.
4608 (wide_int_cmp): New.
4609 (bound_index): Take a widest_int rather than a double_int.
4610 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
4611 (maybe_lower_iteration_bound): Likewise.
4612 (estimate_numbers_of_iterations_loop): Likewise.
4613 (estimated_loop_iterations): Take a widest_int pointer than than
4614 a double_int pointer.
4615 (estimated_loop_iterations_int): Use wide-int interfaces.
4616 (max_loop_iterations): Take a widest_int pointer than than
4617 a double_int pointer.
4618 (max_loop_iterations_int): Use wide-int interfaces.
4619 (max_stmt_executions): Take a widest_int pointer than than
4620 a double_int pointer.
4621 (estimated_stmt_executions): Likewise.
4622 (n_of_executions_at_most): Use wide-int interfaces.
4623 (scev_probably_wraps_p): Likewise.
4624 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
4626 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
4628 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
4629 double_ints. Adjust for trailing_wide_ints <3> representation.
4630 (set_nonzero_bits): Likewise.
4631 (get_range_info): Return wide_ints rather than double_ints.
4632 Adjust for trailing_wide_ints <3> representation.
4633 (get_nonzero_bits): Likewise.
4634 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
4636 * tree-ssanames.h (struct range_info_def): Replace min, max and
4637 nonzero_bits with a trailing_wide_ints <3>.
4638 (set_range_info): Use wide_int_refs rather than double_ints.
4639 (set_nonzero_bits): Likewise.
4640 (get_range_info): Return wide_ints rather than double_ints.
4641 (get_nonzero_bits): Likewise.
4642 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
4643 * tree-ssa-pre.c (phi_translate_1): Likewise.
4644 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
4645 (acceptable_pow_call): Likewise.
4646 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
4648 (vn_reference_fold_indirect): Likewise.
4649 (vn_reference_maybe_forwprop_address): Likewise.
4650 (valueize_refs_1): Likewise.
4651 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
4652 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
4653 tree_int_cst_lt and tree_int_cst_le.
4654 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
4656 (streamer_alloc_tree): Likewise.
4657 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
4658 (streamer_write_tree_header): Likewise.
4659 (streamer_write_integer_cst): Likewise.
4660 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
4661 (build_constructors): Likewise.
4662 (array_value_type): Likewise.
4663 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
4664 (vect_check_gather): Likewise.
4665 * tree-vect-generic.c (build_replicated_const): Likewise.
4666 (expand_vector_divmod): Likewise.
4667 * tree-vect-loop.c (vect_transform_loop): Likewise.
4668 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
4669 (vect_do_peeling_for_alignment): Likewise.
4670 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
4671 * tree-vrp.c: Include wide-int.h.
4672 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
4673 (extract_range_from_assert): Use wide-int interfaces.
4674 (vrp_int_const_binop): Likewise.
4675 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
4676 double_int pointers.
4677 (ranges_from_anti_range): Use wide-int interfaces.
4678 (quad_int_cmp): Delete.
4679 (quad_int_pair_sort): Likewise.
4680 (extract_range_from_binary_expr_1): Use wide-int interfaces.
4681 (extract_range_from_unary_expr_1): Likewise.
4682 (adjust_range_with_scev): Likewise.
4683 (masked_increment): Take and return wide_ints rather than double_ints.
4684 (register_edge_assert_for_2): Use wide-int interfaces.
4685 (check_array_ref): Likewise.
4686 (search_for_addr_array): Likewise.
4687 (maybe_set_nonzero_bits): Likewise.
4688 (union_ranges): Pass an integer of the correct type instead of
4689 using integer_one_node.
4690 (intersect_ranges): Likewise.
4691 (simplify_truth_ops_using_ranges): Likewise.
4692 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
4693 (range_fits_type_p): Likewise.
4694 (simplify_cond_using_ranges): Likewise. Take a signop rather than
4696 (simplify_conversion_using_ranges): Use wide-int interfaces.
4697 (simplify_float_conversion_using_ranges): Likewise.
4698 (vrp_finalize): Likewise.
4699 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
4700 (gimple_stringops_transform): Likewise.
4701 * varasm.c (decode_addr_const): Likewise.
4702 (const_hash_1): Likewise.
4703 (const_rtx_hash_1): Likewise
4704 (output_constant): Likewise.
4705 (array_size_for_constructor): Likewise.
4706 (output_constructor_regular_field): Likewise.
4707 (output_constructor_bitfield): Likewise.
4708 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
4709 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
4711 * gencheck.c: Define BITS_PER_UNIT.
4714 * wide-int-print.cc: New.
4715 * wide-int-print.h: New.
4717 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4719 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
4721 2014-05-06 Richard Biener <rguenther@suse.de>
4723 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
4724 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
4725 (TODO_verify_all): Adjust.
4726 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
4727 TODO_verify_stmts and TODO_verify_rtl_sharing.
4728 * bb-reorder.c: Likewise.
4729 * cfgexpand.c: Likewise.
4730 * cprop.c: Likewise.
4732 * function.c: Likewise.
4733 * fwprop.c: Likewise.
4735 * gimple-ssa-isolate-paths.c: Likewise.
4736 * gimple-ssa-strength-reduction.c: Likewise.
4737 * ipa-split.c: Likewise.
4738 * loop-init.c: Likewise.
4739 * loop-unroll.c: Likewise.
4740 * lower-subreg.c: Likewise.
4741 * modulo-sched.c: Likewise.
4742 * postreload-gcse.c: Likewise.
4743 * predict.c: Likewise.
4744 * recog.c: Likewise.
4745 * sched-rgn.c: Likewise.
4746 * store-motion.c: Likewise.
4747 * tracer.c: Likewise.
4748 * trans-mem.c: Likewise.
4749 * tree-call-cdce.c: Likewise.
4750 * tree-cfg.c: Likewise.
4751 * tree-cfgcleanup.c: Likewise.
4752 * tree-complex.c: Likewise.
4753 * tree-eh.c: Likewise.
4754 * tree-emutls.c: Likewise.
4755 * tree-if-conv.c: Likewise.
4756 * tree-into-ssa.c: Likewise.
4757 * tree-loop-distribution.c: Likewise.
4758 * tree-object-size.c: Likewise.
4759 * tree-parloops.c: Likewise.
4760 * tree-pass.h: Likewise.
4761 * tree-sra.c: Likewise.
4762 * tree-ssa-ccp.c: Likewise.
4763 * tree-ssa-copy.c: Likewise.
4764 * tree-ssa-copyrename.c: Likewise.
4765 * tree-ssa-dce.c: Likewise.
4766 * tree-ssa-dom.c: Likewise.
4767 * tree-ssa-dse.c: Likewise.
4768 * tree-ssa-forwprop.c: Likewise.
4769 * tree-ssa-ifcombine.c: Likewise.
4770 * tree-ssa-loop-ch.c: Likewise.
4771 * tree-ssa-loop-ivcanon.c: Likewise.
4772 * tree-ssa-loop.c: Likewise.
4773 * tree-ssa-math-opts.c: Likewise.
4774 * tree-ssa-phiopt.c: Likewise.
4775 * tree-ssa-phiprop.c: Likewise.
4776 * tree-ssa-pre.c: Likewise.
4777 * tree-ssa-reassoc.c: Likewise.
4778 * tree-ssa-sink.c: Likewise.
4779 * tree-ssa-strlen.c: Likewise.
4780 * tree-ssa-tail-merge.c: Likewise.
4781 * tree-ssa-uncprop.c: Likewise.
4782 * tree-switch-conversion.c: Likewise.
4783 * tree-tailcall.c: Likewise.
4784 * tree-vect-generic.c: Likewise.
4785 * tree-vectorizer.c: Likewise.
4786 * tree-vrp.c: Likewise.
4788 * var-tracking.c: Likewise.
4789 * bt-load.c: Likewise.
4790 * cfgcleanup.c: Likewise.
4791 * combine-stack-adj.c: Likewise.
4792 * combine.c: Likewise.
4793 * compare-elim.c: Likewise.
4794 * config/epiphany/resolve-sw-modes.c: Likewise.
4795 * config/i386/i386.c: Likewise.
4796 * config/mips/mips.c: Likewise.
4797 * config/s390/s390.c: Likewise.
4798 * config/sh/sh_treg_combine.cc: Likewise.
4799 * config/sparc/sparc.c: Likewise.
4802 * final.c: Likewise.
4803 * ifcvt.c: Likewise.
4804 * mode-switching.c: Likewise.
4805 * passes.c: Likewise.
4806 * postreload.c: Likewise.
4808 * reg-stack.c: Likewise.
4809 * regcprop.c: Likewise.
4810 * regrename.c: Likewise.
4813 2014-05-06 Richard Biener <rguenther@suse.de>
4816 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
4817 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
4819 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
4822 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
4824 2014-05-05 Radovan Obradovic <robradovic@mips.com>
4825 Tom de Vries <tom@codesourcery.com>
4827 * target.def (call_fusage_contains_non_callee_clobbers): New
4829 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
4831 (@node Miscellaneous Register Hooks): New node.
4832 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
4833 * doc/tm.texi: Regenerate.
4835 2014-05-05 Marek Polacek <polacek@redhat.com>
4838 * opts.c (common_handle_option): Call error_at instead of warning_at.
4840 2014-05-05 Richard Biener <rguenther@suse.de>
4842 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
4843 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
4844 under the TODO_verify_il umbrella.
4846 2014-05-05 Richard Biener <rguenther@suse.de>
4848 * passes.c (execute_function_todo): Move TODO_verify_flow under
4849 the TODO_verify_ul umbrella.
4851 2014-05-05 Richard Biener <rguenther@suse.de>
4854 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
4855 X & CST away from a CST that is the mask of a mode.
4857 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4859 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
4860 int argument to enum machine_mode.
4861 (picochip_class_max_nregs): Ditto.
4862 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
4863 (picochip_class_max_nregs): Ditto.
4865 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4867 * target.def: Add new target hook.
4868 * doc/tm.texi: Regenerate.
4869 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
4870 * targhooks.c (default_keep_leaf_when_profiled): New function.
4872 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
4873 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
4875 2014-05-05 Bin Cheng <bin.cheng@arm.com>
4877 PR tree-optimization/60363
4878 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
4879 (copy_phi_args): New parameters. Call get_value_locus_in_path.
4880 (update_destination_phis): New parameter.
4881 (create_edge_and_update_destination_phis): Ditto.
4882 (ssa_fix_duplicate_block_edges): Pass new arguments.
4883 (thread_single_edge): Ditto.
4885 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
4887 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
4888 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
4889 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
4890 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
4891 Use RS6000_BTM_HARD_FLOAT.
4892 (BU_MISC_2): Likewise.
4893 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
4894 RS6000_BTM_HARD_FLOAT.
4895 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
4897 (rs6000_invalid_builtin): Add hard floating builtin support.
4898 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
4899 hard float builtins.
4900 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
4902 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4904 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
4905 Add missing function* argument.
4907 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4909 * lra-constraints.c (valid_address_p): Move earlier in file.
4910 Add a constraint argument to the address_info version.
4911 (satisfies_memory_constraint_p): New function.
4912 (satisfies_address_constraint_p): Likewise.
4913 (process_alt_operands, curr_insn_transform): Use them.
4914 (process_address): Pass the constraint to valid_address_p when
4915 checking address operands.
4917 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4919 * config/mips/mips.c (mips_isa_rev): New variable.
4920 (mips_set_architecture): Set it.
4921 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
4923 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
4924 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
4925 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
4926 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
4927 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
4928 conditions in terms of mips_isa_rev.
4929 (mips_isa_rev): Declare.
4931 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4933 * config/sh/sh-mem.cc: Use tabs instead of spaces.
4934 (prob_unlikely, prob_likely): Make variables const.
4936 2014-05-03 Denis Chertykov <chertykov@gmail.com>
4938 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
4940 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4942 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
4944 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4946 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
4947 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
4948 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
4950 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
4951 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
4953 Replace usage of ROUND_REG with sh_round_reg.
4954 Use CEIL instead of ROUND_ADVANCE.
4956 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4959 * config/sh/sh.c: Include stdlib headers before everything else.
4961 2014-05-02 Jakub Jelinek <jakub@redhat.com>
4963 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
4964 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
4965 (gimplify_adjust_omp_clauses): Simd region is never
4966 directly nested in combined parallel. Instead, for linear
4967 with copyin/copyout, if in combined for simd loop, make decl
4968 firstprivate/lastprivate on OMP_FOR.
4969 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
4970 expand_omp_for_static_chunk): When setting endvar, also set
4971 fd->loop.v to the same value.
4973 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4975 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
4977 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
4979 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
4982 2014-05-02 Marek Polacek <polacek@redhat.com>
4984 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
4986 2014-05-02 Kito Cheng <kito@0xlab.org>
4988 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
4989 to a C expression marco.
4990 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
4991 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
4992 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
4993 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
4994 HONOR_REG_ALLOC_ORDER.
4995 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
4997 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4999 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
5001 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5003 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
5005 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
5007 * tree-if-conv.c (is_cond_scalar_reduction): New function.
5008 (convert_scalar_cond_reduction): Likewise.
5009 (predicate_scalar_phi): Add recognition and transformation
5010 of simple conditioanl reduction to be vectorizable.
5012 2014-05-01 Marek Polacek <polacek@redhat.com>
5015 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
5017 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
5019 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
5020 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
5021 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
5022 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
5023 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
5024 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
5025 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
5026 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
5028 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
5030 * config/arc/arc.opt (mlra): Move comment above option name
5031 to avoid mis-parsing as language options.
5033 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5035 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
5036 * config/sol2.h: ... here.
5037 * config/sol2-10.h: Remove.
5039 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
5040 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
5041 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
5042 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
5043 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
5044 * config/sol2.h: ... here.
5045 (SECTION_NAME_FORMAT): Don't redefine.
5046 (STARTFILE_ARCH32_SPEC): Rename to ...
5047 (STARTFILE_ARCH_SPEC): ... this.
5048 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
5049 * config/sparc/sol2.h: ... here.
5050 (SECTION_NAME_FORMAT): Don't undef.
5051 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
5052 (SUBTARGET_EXTRA_SPECS): Remove.
5053 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
5055 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
5056 (MD_STARTFILE_PREFIX): Remove.
5057 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
5058 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
5059 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
5060 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
5061 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
5062 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
5063 * config/i386/sol2.h: ... here.
5064 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
5065 * config/i386/sol2-bi.h: Remove.
5066 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
5067 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
5069 * config/i386/t-sol2-64: Rename to ...
5070 * config/i386/t-sol2: ... this.
5071 * config/sparc/t-sol2-64: Rename to ...
5072 * config/sparc/t-sol2: ... this.
5074 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
5075 sol2_tm_file_head, sol2_tm_file_tail.
5076 Include ${cpu_type}/sol2.h before sol2.h.
5078 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
5079 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
5080 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
5081 Reflect i386/t-sol2-64 renaming.
5082 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
5083 Reflect sparc/t-sol2-64 renaming.
5085 2014-04-30 Richard Biener <rguenther@suse.de>
5087 * passes.c (execute_function_todo): Move TODO_verify_stmts
5088 and TODO_verify_ssa under the TODO_verify_il umbrella.
5089 * tree-ssa.h (verify_ssa): Adjust prototype.
5090 * tree-ssa.c (verify_ssa): Add parameter to tell whether
5091 we should verify SSA operands.
5092 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
5093 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
5094 whether we should verify whether not throwing stmts have EH info.
5095 * graphite-scop-detection.c (create_sese_edges): Adjust.
5096 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
5097 * tree-eh.c (lower_try_finally_switch): Do not add the
5098 default case label twice.
5100 2014-04-30 Marek Polacek <polacek@redhat.com>
5102 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
5103 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
5104 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
5105 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
5107 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
5109 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
5110 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
5111 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
5112 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
5113 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
5114 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
5115 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
5116 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
5118 2014-04-29 David Malcolm <dmalcolm@redhat.com>
5120 * tree-cfg.c (dump_function_to_file): Dump the return type of
5121 functions, in a line to itself before the function body, mimicking
5122 the layout of a C function.
5124 2014-04-29 Jakub Jelinek <jakub@redhat.com>
5126 PR tree-optimization/60971
5127 * tree-tailcall.c (process_assignment): Reject conversions which
5130 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
5132 * calls.c (initialize_argument_information): Always treat
5133 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
5134 (expand_call): Likewise.
5135 (emit_library_call_calue_1): Likewise.
5136 * expr.c (PUSH_ARGS_REVERSED): Do not define.
5137 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
5140 2014-04-29 Nick Clifton <nickc@redhat.com>
5142 * config/msp430/msp430.md (umulsidi): Fix typo.
5143 (mulhisi3): Enable even inside interrupt handlers.
5144 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
5145 bigger return address pushed in large mode.
5147 2014-04-29 Nick Clifton <nickc@redhat.com>
5149 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
5150 (arc_init_reg_tables): Use a machine_mode enum to iterate over
5152 * config/m32r/m32r.c (init_reg_tables): Likewise.
5153 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
5154 enum to hold the modes.
5156 2014-04-29 Richard Biener <rguenther@suse.de>
5158 * dominance.c (free_dominance_info): Add overload with
5160 (dom_info_state): Likewise.
5161 (dom_info_available_p): Likewise.
5162 * basic-block.h (free_dominance_info, dom_info_state,
5163 dom_info_available_p): Declare overloads.
5164 * passes.c (execute_function_todo): Verify that verifiers
5165 don't change dominator info state. Drop dominator info
5166 for IPA pass invocations.
5167 * cgraph.c (release_function_body): Restore asserts that
5168 dominator information is released.
5170 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
5172 * doc/invoke.texi: Fix typo.
5173 * tree-vrp.c: Fix typos.
5174 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
5176 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5178 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
5180 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
5182 * config/aarch64/aarch64-builtins.c
5183 (aarch64_types_storestruct_lane_qualifiers): New.
5184 (TYPES_STORESTRUCT_LANE): Likewise.
5185 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
5186 (st3_lane): Likewise.
5187 (st4_lane): Likewise.
5188 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
5189 (vec_store_lanesci_lane<mode>): Likewise.
5190 (vec_store_lanesxi_lane<mode>): Likewise.
5191 (aarch64_st2_lane<VQ:mode>): Likewise.
5192 (aarch64_st3_lane<VQ:mode>): Likewise.
5193 (aarch64_st4_lane<VQ:mode>): Likewise.
5194 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
5195 * config/aarch64/arm_neon.h
5196 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
5197 use new macro arguments.
5198 (__ST3_LANE_FUNC): Likewise.
5199 (__ST4_LANE_FUNC): Likewise.
5200 * config/aarch64/iterators.md (V_TWO_ELEM): New.
5201 (V_THREE_ELEM): Likewise.
5202 (V_FOUR_ELEM): Likewise.
5204 2014-04-28 David Malcolm <dmalcolm@redhat.com>
5206 * doc/gimple.texi: Replace the description of the now-defunct
5207 union gimple_statement_d with a diagram showing the
5208 gimple_statement_base class hierarchy and its relationships to
5209 the GSS_ and GIMPLE_ enums.
5211 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
5213 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
5214 * config/aarch64/aarch64.c
5215 (aarch64_cannot_change_mode_class): Weaken conditions.
5216 (aarch64_modes_tieable_p): New.
5217 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
5219 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
5221 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5222 (loadsync_<mode>): Change mode.
5223 (load_quadpti, store_quadpti): New.
5224 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5225 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5227 2014-04-28 Martin Jambor <mjambor@suse.cz>
5229 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
5230 same alias type as the original statement.
5231 (subreplacement_assignment_data): New type.
5232 (handle_unscalarized_data_in_subtree): New type of parameter,
5233 generate new memory accesses with same alias type as the original
5235 (load_assign_lhs_subreplacements): Likewise.
5236 (sra_modify_constructor_assign): Generate new memory accesses with
5237 same alias type as the original statement.
5239 2014-04-28 Richard Biener <rguenther@suse.de>
5241 * tree-pass.h (TODO_verify_il): Define.
5242 (TODO_verify_all): Complete properly.
5243 * passes.c (execute_function_todo): Move existing loop-closed
5244 SSA verification under TODO_verify_il.
5245 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
5246 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
5247 Fix tree sharing issue.
5249 2014-04-28 Richard Biener <rguenther@suse.de>
5252 * builtins.def (DEF_C11_BUILTIN): Add.
5253 (BUILT_IN_ALIGNED_ALLOC): Likewise.
5254 * coretypes.h (enum function_class): Add function_c11_misc.
5255 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
5256 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
5257 (call_may_clobber_ref_p_1): Likewise.
5258 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
5259 (mark_all_reaching_defs_necessary_1): Likewise.
5260 (propagate_necessity): Likewise.
5261 (eliminate_unnecessary_stmts): Likewise.
5262 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
5264 2014-04-28 Richard Biener <rguenther@suse.de>
5266 * tree-vrp.c (vrp_var_may_overflow): Remove.
5267 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
5268 with overflow immediately bump to one before that value and
5269 let iteration figure out overflow status.
5271 2014-04-28 Richard Biener <rguenther@suse.de>
5273 * configure.ac: Do valgrind header checks unconditionally.
5274 Add --enable-valgrind-annotations.
5275 * system.h: Guard valgrind header inclusion with
5276 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
5277 * alloc-pool.c (pool_alloc, pool_free): Use
5278 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
5279 to guard possibly dead code.
5280 * config.in: Regenerated.
5281 * configure: Likewise.
5283 2014-04-28 Jeff Law <law@redhat.com>
5285 PR tree-optimization/60902
5286 * tree-ssa-threadedge.c
5287 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
5288 over real defs when invalidating outputs from statements that do not
5289 produce useful outputs for threading.
5291 2014-04-28 Richard Biener <rguenther@suse.de>
5293 PR tree-optimization/60979
5294 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
5295 SCOPs that end in a block with a successor with abnormal
5298 2014-04-28 Richard Biener <rguenther@suse.de>
5300 * tree-pass.h (execute_pass_list): Adjust prototype.
5301 * passes.c (pass_manager::execute_early_local_passes): Adjust.
5302 (do_per_function): Change callback signature, push all actual
5303 work to the callbals.
5304 (do_per_function_toporder): Likewise.
5305 (execute_function_dump): Adjust.
5306 (execute_function_todo): Likewise.
5307 (clear_last_verified): Likewise.
5308 (verify_curr_properties): Likewise.
5309 (update_properties_after_pass): Likewise.
5310 (execute_pass_list_1): Split out from ...
5311 (execute_pass_list): ... here. Adjust.
5312 (execute_ipa_pass_list): Likewise.
5313 * cgraphunit.c (cgraph_add_new_function): Adjust.
5314 (analyze_function): Likewise.
5315 (expand_function): Likewise.
5316 * cgraph.c (release_function_body): Free dominance info
5317 here instead of asserting it was magically freed elsewhere.
5319 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
5321 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
5322 * configure: Regenerate.
5323 * config/sparc/sparc.opt (muser-mode): New option.
5324 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
5326 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
5327 * doc/invoke.texi (SPARC options): Document -muser-mode.
5329 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
5331 * cselib.c (find_slot_memmode): Delete.
5332 (cselib_hasher): Change compare_type to a struct.
5333 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
5335 (preserve_constants_and_equivs): Adjust for new compare_type.
5336 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
5337 (wrap_constant): Delete.
5338 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
5340 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
5342 * doc/install.texi (Building with profile feedback): Remove
5345 2014-04-26 Tom de Vries <tom@codesourcery.com>
5347 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
5350 2014-04-25 Cary Coutant <ccoutant@google.com>
5353 * dwarf2out.c (should_move_die_to_comdat): A type definition
5354 can contain a subprogram definition, but don't move it to a
5356 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
5357 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
5359 (clone_tree_hash): Rename to...
5360 (clone_tree_partial): ...this; change callers. Copy
5361 DW_TAG_subprogram DIEs as declarations.
5362 (copy_decls_walk): Don't copy children of a declaration into a
5365 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
5368 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
5371 2014-04-25 Jiong Wang <jiong.wang@arm.com>
5373 * config/arm/predicates.md (call_insn_operand): Add long_call check.
5374 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
5376 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
5379 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5381 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
5383 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5385 PR tree-optimization/60930
5386 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
5387 creating a multiply candidate by folding two constant
5388 multiplicands when the result overflows.
5390 2014-04-25 Jakub Jelinek <jakub@redhat.com>
5392 PR tree-optimization/60960
5393 * tree-vect-generic.c (expand_vector_operation): Only call
5394 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
5396 2014-04-25 Tom de Vries <tom@codesourcery.com>
5398 * expr.c (clobber_reg_mode): New function.
5399 * expr.h (clobber_reg): New function.
5401 2014-04-25 Tom de Vries <tom@codesourcery.com>
5403 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
5406 2014-04-25 Radovan Obradovic <robradovic@mips.com>
5407 Tom de Vries <tom@codesourcery.com>
5409 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
5411 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
5412 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
5413 new argument to find_all_hard_reg_sets call.
5415 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5417 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
5418 Use HOST_WIDE_INT_C for mask literal.
5419 (aarch_rev16_shleft_mask_imm_p): Likewise.
5421 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
5424 * config/sparc/sparc.md (ashlsi3_extend): Delete.
5426 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
5428 PR preprocessor/56540
5429 * config/i386/i386-c.c (ix86_target_macros): Define
5430 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
5432 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5434 * configure.ac (tga_func): Remove.
5435 (LIB_TLS_SPEC): Remove.
5436 * configure: Regenerate.
5437 * config.in: Regenerate.
5438 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
5440 2014-04-25 Richard Biener <rguenther@suse.de>
5443 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
5444 call stmt use/clobber sets during stmt walk instead of
5445 walking the possibly incomplete set of caller edges.
5447 2014-04-25 Richard Biener <rguenther@suse.de>
5450 * passes.c (apply_ipa_transforms): Inline into only caller ...
5451 (execute_one_pass): ... here. Properly bring in function
5452 bodies for nodes we want to apply IPA transforms to.
5454 2014-04-24 Cong Hou <congh@google.com>
5456 PR tree-optimization/60896
5457 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
5458 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
5459 (vect_mark_pattern_stmts): Set the def type of all statements in
5460 PATTERN_DEF_SEQ as vect_internal_def.
5462 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
5464 * doc/extend.texi (PowerPC Built-in Functions): Document new
5465 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
5466 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
5468 * config/rs6000/predicates.md (const_0_to_3_operand): New
5469 predicate to match 0..3 integer constants.
5471 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
5472 to support adding miscellaneous builtin functions.
5473 (BU_DFP_MISC_2): Likewise.
5474 (BU_P7_MISC_1): Likewise.
5475 (BU_P7_MISC_2): Likewise.
5476 (BU_P8V_MISC_3): Likewise.
5477 (BU_MISC_1): Likewise.
5478 (BU_MISC_2): Likewise.
5479 (DIVWE): Add extended divide builtin functions.
5482 (DIVWEUO): Likewise.
5486 (DIVDEUO): Likewise.
5487 (DXEX): Add decimal floating-point builtin functions.
5490 (DDEDPDQ): Likewise.
5492 (DENBCDQ): Likewise.
5499 (CDTBCD): Add new BCD builtin functions.
5503 (BCDADD_LT): Likewise.
5504 (BCDADD_EQ): Likewise.
5505 (BCDADD_GT): Likewise.
5506 (BCDADD_OV): Likewise.
5508 (BCDSUB_LT): Likewise.
5509 (BCDSUB_EQ): Likewise.
5510 (BCDSUB_GT): Likewise.
5511 (BCDSUB_OV): Likewise.
5512 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
5513 (UNPACK_TD): Likewise.
5514 (PACK_TF): Likewise.
5515 (UNPACK_TF): Likewise.
5516 (UNPACK_TF_0): Likewise.
5517 (UNPACK_TF_1): Likewise.
5518 (PACK_V1TI): Likewise.
5519 (UNPACK_V1TI): Likewise.
5521 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
5522 support for decimal floating point builtin functions.
5523 (rs6000_expand_ternop_builtin): Add checks for the new builtin
5524 functions that take constant arguments.
5525 (rs6000_invalid_builtin): Add decimal floating point builtin support.
5526 (rs6000_init_builtins): Setup long double, _Decimal64, and
5527 _Decimal128 types for new builtin functions.
5528 (builtin_function_type): Set the unsigned flags appropriately for
5529 the new builtin functions.
5530 (rs6000_opt_masks): Add support for decimal floating point builtin
5533 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
5534 floating point builtin functions.
5535 (RS6000_BTM_COMMON): Likewise.
5536 (RS6000_BTI_long_double): Likewise.
5537 (RS6000_BTI_dfloat64): Likewise.
5538 (RS6000_BTI_dfloat128): Likewise.
5539 (long_double_type_internal_node): Likewise.
5540 (dfloat64_type_internal_node): Likewise.
5541 (dfloat128_type_internal_node): Likewise.
5543 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
5544 2.07 bcd arithmetic instructions.
5545 (UNSPEC_BCDSUB): Likewise.
5546 (UNSPEC_BCD_OVERFLOW): Likewise.
5547 (UNSPEC_BCD_ADD_SUB): Likewise.
5548 (bcd_add_sub): Likewise.
5549 (BCD_TEST): Likewise.
5550 (bcd<bcd_add_sub>): Likewise.
5551 (bcd<bcd_add_sub>_test): Likewise.
5552 (bcd<bcd_add_sub>_test2): Likewise.
5553 (bcd<bcd_add_sub>_<code>): Likewise.
5554 (peephole2 for combined bcd ops): Likewise.
5556 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
5557 decimal floating point builtin functions.
5558 (UNSPEC_DENBCD): Likewise.
5559 (UNSPEC_DXEX): Likewise.
5560 (UNSPEC_DIEX): Likewise.
5561 (UNSPEC_DSCLI): Likewise.
5562 (UNSPEC_DSCRI): Likewise.
5563 (D64_D128): Likewise.
5564 (dfp_suffix): Likewise.
5565 (dfp_ddedpd_<mode>): Likewise.
5566 (dfp_denbcd_<mode>): Likewise.
5567 (dfp_dxex_<mode>): Likewise.
5568 (dfp_diex_<mode>): Likewise.
5569 (dfp_dscli_<mode>): Likewise.
5570 (dfp_dscri_<mode>): Likewise.
5572 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
5574 (UNSPEC_CDTBCD): Likewise.
5575 (UNSPEC_CBCDTD): Likewise.
5576 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
5577 (UNSPEC_DIVEO): Likewise.
5578 (UNSPEC_DIVEU): Likewise.
5579 (UNSPEC_DIVEUO): Likewise.
5580 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
5581 pack/unpack 128-bit types.
5582 (UNSPEC_PACK_128BIT): Likewise.
5583 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
5584 (udiv<mode>3): Use idiv_ldiv mode attribute.
5585 (div<mode>3): Likewise.
5586 (addg6s): Add new BCD builtin functions.
5589 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
5590 (div_extend): Likewise.
5591 (div<div_extend>_<mode>"): Likewise.
5592 (FP128_64): Add support for new builtin functions to pack/unpack
5594 (unpack<mode>): Likewise.
5595 (unpacktf_0): Likewise.
5596 (unpacktf_1): Likewise.
5597 (unpack<mode>_dm): Likewise.
5598 (unpack<mode>_nodm): Likewise.
5599 (pack<mode>): Likewise.
5600 (unpackv1ti): Likewise.
5601 (packv1ti): Likewise.
5603 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
5605 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
5608 2014-04-24 Jakub Jelinek <jakub@redhat.com>
5610 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
5611 * gimplify.c (omp_is_private): Change last argument's type to int.
5612 Only diagnose lastprivate if the simd argument is 1, only diagnose
5613 linear if the simd argument is 2.
5614 (gimplify_omp_for): Adjust omp_is_private callers. When adding
5615 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
5616 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
5617 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
5618 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
5619 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5620 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
5621 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5622 * tree-nested.c (convert_nonlocal_omp_clauses,
5623 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
5625 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
5628 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
5631 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
5633 * flag-types.h (enum ivar_visibility): Add.
5635 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
5637 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
5638 function * argument.
5640 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
5642 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
5644 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5645 Tom de Vries <tom@codesourcery.com>
5647 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
5648 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
5650 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
5651 * emit-rtl.c (try_split): Same.
5653 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5654 Tom de Vries <tom@codesourcery.com>
5656 * common.opt (fuse-caller-save): New option.
5658 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
5660 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
5661 elements for big-endian.
5663 2014-04-24 Richard Biener <rguenther@suse.de>
5665 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
5666 during TER and instead use the sepops interface for expanding
5667 non-GIMPLE_SINGLE_RHS.
5669 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5671 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
5672 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
5674 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5676 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
5677 assembler 64-bit option.
5678 * configure: Regenerate.
5680 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5682 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
5683 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
5684 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
5685 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
5686 (TARGET_CRYPTO): Take TARGET_SIMD into account.
5688 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5690 * config/aarch64/aarch64-builtins.c
5691 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
5692 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
5693 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
5694 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
5696 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
5697 (Vrevsuff): New mode attribute.
5699 2014-04-24 Terry Guo <terry.guo@arm.com>
5701 * config/arm/arm.h (machine_function): Define variable
5702 after_arm_reorg here.
5703 * config/arm/arm.c (after_arm_reorg): Remove the definition.
5704 (arm_split_constant): Update the way to access variable
5707 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
5709 2014-04-23 Tom de Vries <tom@codesourcery.com>
5711 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
5713 2014-04-23 David Malcolm <dmalcolm@redhat.com>
5715 * is-a.h: Update comments to reflect the following changes to the
5716 "pointerness" of the API, making the template parameter match the
5717 return type, allowing use of is-a.h with typedefs of pointers.
5718 (is_a_helper::cast): Return a T rather then a pointer to a T, so
5719 that the return type matches the parameter to the is_a_helper.
5721 (dyn_cast): Likewise.
5723 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
5724 pointer from the is-a.h API.
5726 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
5727 (is_a_helper <cgraph_node *>::test): ...this, matching change to
5729 (is_a_helper <varpool_node>::test): Likewise, convert to...
5730 (is_a_helper <varpool_node *>::test): ...this.
5732 (varpool_first_variable): Update for removal of implicit pointer
5733 from the is-a.h API.
5734 (varpool_next_variable): Likewise.
5735 (varpool_first_static_initializer): Likewise.
5736 (varpool_next_static_initializer): Likewise.
5737 (varpool_first_defined_variable): Likewise.
5738 (varpool_next_defined_variable): Likewise.
5739 (cgraph_first_defined_function): Likewise.
5740 (cgraph_next_defined_function): Likewise.
5741 (cgraph_first_function): Likewise.
5742 (cgraph_next_function): Likewise.
5743 (cgraph_first_function_with_gimple_body): Likewise.
5744 (cgraph_next_function_with_gimple_body): Likewise.
5745 (cgraph_alias_target): Likewise.
5746 (varpool_alias_target): Likewise.
5747 (cgraph_function_or_thunk_node): Likewise.
5748 (varpool_variable_node): Likewise.
5749 (symtab_real_symbol_p): Likewise.
5750 * cgraphunit.c (referred_to_p): Likewise.
5751 (analyze_functions): Likewise.
5752 (handle_alias_pairs): Likewise.
5753 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
5754 * gimple-ssa.h (gimple_vuse_op): Likewise.
5755 (gimple_vdef_op): Likewise.
5756 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
5757 * gimple.c (gimple_build_asm_1): Likewise.
5758 (gimple_build_try): Likewise.
5759 (gimple_build_resx): Likewise.
5760 (gimple_build_eh_dispatch): Likewise.
5761 (gimple_build_omp_for): Likewise.
5762 (gimple_omp_for_set_clauses): Likewise.
5764 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
5765 (is_a_helper <gimple_statement_asm *>::test): ...this.
5766 (is_a_helper <gimple_statement_bind>::test): Convert to...
5767 (is_a_helper <gimple_statement_bind *>::test): ...this.
5768 (is_a_helper <gimple_statement_call>::test): Convert to...
5769 (is_a_helper <gimple_statement_call *>::test): ...this.
5770 (is_a_helper <gimple_statement_catch>::test): Convert to...
5771 (is_a_helper <gimple_statement_catch *>::test): ...this.
5772 (is_a_helper <gimple_statement_resx>::test): Convert to...
5773 (is_a_helper <gimple_statement_resx *>::test): ...this.
5774 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
5775 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
5776 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
5777 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
5778 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
5779 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
5780 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
5781 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
5782 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
5783 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
5784 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
5785 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
5786 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
5787 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
5788 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
5789 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
5790 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
5791 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
5792 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
5793 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
5794 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
5795 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
5796 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
5797 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
5798 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
5799 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
5800 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
5801 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
5802 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
5803 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
5804 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
5805 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
5806 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
5807 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
5808 (is_a_helper <gimple_statement_phi>::test): Convert to...
5809 (is_a_helper <gimple_statement_phi *>::test): ...this.
5810 (is_a_helper <gimple_statement_transaction>::test): Convert to...
5811 (is_a_helper <gimple_statement_transaction *>::test): ...this.
5812 (is_a_helper <gimple_statement_try>::test): Convert to...
5813 (is_a_helper <gimple_statement_try *>::test): ...this.
5814 (is_a_helper <gimple_statement_wce>::test): Convert to...
5815 (is_a_helper <gimple_statement_wce *>::test): ...this.
5816 (is_a_helper <const gimple_statement_asm>::test): Convert to...
5817 (is_a_helper <const gimple_statement_asm *>::test): ...this.
5818 (is_a_helper <const gimple_statement_bind>::test): Convert to...
5819 (is_a_helper <const gimple_statement_bind *>::test): ...this.
5820 (is_a_helper <const gimple_statement_call>::test): Convert to...
5821 (is_a_helper <const gimple_statement_call *>::test): ...this.
5822 (is_a_helper <const gimple_statement_catch>::test): Convert to...
5823 (is_a_helper <const gimple_statement_catch *>::test): ...this.
5824 (is_a_helper <const gimple_statement_resx>::test): Convert to...
5825 (is_a_helper <const gimple_statement_resx *>::test): ...this.
5826 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
5827 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
5828 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
5829 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
5830 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
5832 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
5834 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
5836 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
5838 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
5839 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
5840 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
5842 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
5843 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
5845 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
5846 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
5847 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
5848 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
5849 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
5850 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
5852 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
5853 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
5854 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
5855 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
5857 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
5858 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
5859 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
5860 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
5861 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
5862 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
5863 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
5864 (is_a_helper <const gimple_statement_phi>::test): Convert to...
5865 (is_a_helper <const gimple_statement_phi *>::test): ...this.
5866 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
5867 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
5868 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
5869 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
5870 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
5871 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
5872 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
5874 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
5876 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
5877 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
5879 (gimple_use_ops): Update for removal of implicit pointer from the
5881 (gimple_set_use_ops): Likewise.
5882 (gimple_vuse): Likewise.
5883 (gimple_vdef): Likewise.
5884 (gimple_vuse_ptr): Likewise.
5885 (gimple_vdef_ptr): Likewise.
5886 (gimple_set_vuse): Likewise.
5887 (gimple_set_vdef): Likewise.
5888 (gimple_omp_return_set_lhs): Likewise.
5889 (gimple_omp_return_lhs): Likewise.
5890 (gimple_omp_return_lhs_ptr): Likewise.
5891 (gimple_call_fntype): Likewise.
5892 (gimple_call_set_fntype): Likewise.
5893 (gimple_call_set_internal_fn): Likewise.
5894 (gimple_call_use_set): Likewise.
5895 (gimple_call_clobber_set): Likewise.
5896 (gimple_bind_vars): Likewise.
5897 (gimple_bind_set_vars): Likewise.
5898 (gimple_bind_body_ptr): Likewise.
5899 (gimple_bind_set_body): Likewise.
5900 (gimple_bind_add_stmt): Likewise.
5901 (gimple_bind_block): Likewise.
5902 (gimple_bind_set_block): Likewise.
5903 (gimple_asm_ninputs): Likewise.
5904 (gimple_asm_noutputs): Likewise.
5905 (gimple_asm_nclobbers): Likewise.
5906 (gimple_asm_nlabels): Likewise.
5907 (gimple_asm_input_op): Likewise.
5908 (gimple_asm_input_op_ptr): Likewise.
5909 (gimple_asm_output_op): Likewise.
5910 (gimple_asm_output_op_ptr): Likewise.
5911 (gimple_asm_set_output_op): Likewise.
5912 (gimple_asm_clobber_op): Likewise.
5913 (gimple_asm_set_clobber_op): Likewise.
5914 (gimple_asm_label_op): Likewise.
5915 (gimple_asm_set_label_op): Likewise.
5916 (gimple_asm_string): Likewise.
5917 (gimple_catch_types): Likewise.
5918 (gimple_catch_types_ptr): Likewise.
5919 (gimple_catch_handler_ptr): Likewise.
5920 (gimple_catch_set_types): Likewise.
5921 (gimple_catch_set_handler): Likewise.
5922 (gimple_eh_filter_types): Likewise.
5923 (gimple_eh_filter_types_ptr): Likewise.
5924 (gimple_eh_filter_failure_ptr): Likewise.
5925 (gimple_eh_filter_set_types): Likewise.
5926 (gimple_eh_filter_set_failure): Likewise.
5927 (gimple_eh_must_not_throw_fndecl): Likewise.
5928 (gimple_eh_must_not_throw_set_fndecl): Likewise.
5929 (gimple_eh_else_n_body_ptr): Likewise.
5930 (gimple_eh_else_e_body_ptr): Likewise.
5931 (gimple_eh_else_set_n_body): Likewise.
5932 (gimple_eh_else_set_e_body): Likewise.
5933 (gimple_try_eval_ptr): Likewise.
5934 (gimple_try_cleanup_ptr): Likewise.
5935 (gimple_try_set_eval): Likewise.
5936 (gimple_try_set_cleanup): Likewise.
5937 (gimple_wce_cleanup_ptr): Likewise.
5938 (gimple_wce_set_cleanup): Likewise.
5939 (gimple_phi_capacity): Likewise.
5940 (gimple_phi_num_args): Likewise.
5941 (gimple_phi_result): Likewise.
5942 (gimple_phi_result_ptr): Likewise.
5943 (gimple_phi_set_result): Likewise.
5944 (gimple_phi_arg): Likewise.
5945 (gimple_phi_set_arg): Likewise.
5946 (gimple_resx_region): Likewise.
5947 (gimple_resx_set_region): Likewise.
5948 (gimple_eh_dispatch_region): Likewise.
5949 (gimple_eh_dispatch_set_region): Likewise.
5950 (gimple_omp_critical_name): Likewise.
5951 (gimple_omp_critical_name_ptr): Likewise.
5952 (gimple_omp_critical_set_name): Likewise.
5953 (gimple_omp_for_clauses): Likewise.
5954 (gimple_omp_for_clauses_ptr): Likewise.
5955 (gimple_omp_for_set_clauses): Likewise.
5956 (gimple_omp_for_collapse): Likewise.
5957 (gimple_omp_for_index): Likewise.
5958 (gimple_omp_for_index_ptr): Likewise.
5959 (gimple_omp_for_set_index): Likewise.
5960 (gimple_omp_for_initial): Likewise.
5961 (gimple_omp_for_initial_ptr): Likewise.
5962 (gimple_omp_for_set_initial): Likewise.
5963 (gimple_omp_for_final): Likewise.
5964 (gimple_omp_for_final_ptr): Likewise.
5965 (gimple_omp_for_set_final): Likewise.
5966 (gimple_omp_for_incr): Likewise.
5967 (gimple_omp_for_incr_ptr): Likewise.
5968 (gimple_omp_for_set_incr): Likewise.
5969 (gimple_omp_for_pre_body_ptr): Likewise.
5970 (gimple_omp_for_set_pre_body): Likewise.
5971 (gimple_omp_parallel_clauses): Likewise.
5972 (gimple_omp_parallel_clauses_ptr): Likewise.
5973 (gimple_omp_parallel_set_clauses): Likewise.
5974 (gimple_omp_parallel_child_fn): Likewise.
5975 (gimple_omp_parallel_child_fn_ptr): Likewise.
5976 (gimple_omp_parallel_set_child_fn): Likewise.
5977 (gimple_omp_parallel_data_arg): Likewise.
5978 (gimple_omp_parallel_data_arg_ptr): Likewise.
5979 (gimple_omp_parallel_set_data_arg): Likewise.
5980 (gimple_omp_task_clauses): Likewise.
5981 (gimple_omp_task_clauses_ptr): Likewise.
5982 (gimple_omp_task_set_clauses): Likewise.
5983 (gimple_omp_task_child_fn): Likewise.
5984 (gimple_omp_task_child_fn_ptr): Likewise.
5985 (gimple_omp_task_set_child_fn): Likewise.
5986 (gimple_omp_task_data_arg): Likewise.
5987 (gimple_omp_task_data_arg_ptr): Likewise.
5988 (gimple_omp_task_set_data_arg): Likewise.
5989 (gimple_omp_taskreg_clauses): Likewise.
5990 (gimple_omp_taskreg_clauses_ptr): Likewise.
5991 (gimple_omp_taskreg_set_clauses): Likewise.
5992 (gimple_omp_taskreg_child_fn): Likewise.
5993 (gimple_omp_taskreg_child_fn_ptr): Likewise.
5994 (gimple_omp_taskreg_set_child_fn): Likewise.
5995 (gimple_omp_taskreg_data_arg): Likewise.
5996 (gimple_omp_taskreg_data_arg_ptr): Likewise.
5997 (gimple_omp_taskreg_set_data_arg): Likewise.
5998 (gimple_omp_task_copy_fn): Likewise.
5999 (gimple_omp_task_copy_fn_ptr): Likewise.
6000 (gimple_omp_task_set_copy_fn): Likewise.
6001 (gimple_omp_task_arg_size): Likewise.
6002 (gimple_omp_task_arg_size_ptr): Likewise.
6003 (gimple_omp_task_set_arg_size): Likewise.
6004 (gimple_omp_task_arg_align): Likewise.
6005 (gimple_omp_task_arg_align_ptr): Likewise.
6006 (gimple_omp_task_set_arg_align): Likewise.
6007 (gimple_omp_single_clauses): Likewise.
6008 (gimple_omp_single_clauses_ptr): Likewise.
6009 (gimple_omp_single_set_clauses): Likewise.
6010 (gimple_omp_target_clauses): Likewise.
6011 (gimple_omp_target_clauses_ptr): Likewise.
6012 (gimple_omp_target_set_clauses): Likewise.
6013 (gimple_omp_target_child_fn): Likewise.
6014 (gimple_omp_target_child_fn_ptr): Likewise.
6015 (gimple_omp_target_set_child_fn): Likewise.
6016 (gimple_omp_target_data_arg): Likewise.
6017 (gimple_omp_target_data_arg_ptr): Likewise.
6018 (gimple_omp_target_set_data_arg): Likewise.
6019 (gimple_omp_teams_clauses): Likewise.
6020 (gimple_omp_teams_clauses_ptr): Likewise.
6021 (gimple_omp_teams_set_clauses): Likewise.
6022 (gimple_omp_sections_clauses): Likewise.
6023 (gimple_omp_sections_clauses_ptr): Likewise.
6024 (gimple_omp_sections_set_clauses): Likewise.
6025 (gimple_omp_sections_control): Likewise.
6026 (gimple_omp_sections_control_ptr): Likewise.
6027 (gimple_omp_sections_set_control): Likewise.
6028 (gimple_omp_for_set_cond): Likewise.
6029 (gimple_omp_for_cond): Likewise.
6030 (gimple_omp_atomic_store_set_val): Likewise.
6031 (gimple_omp_atomic_store_val): Likewise.
6032 (gimple_omp_atomic_store_val_ptr): Likewise.
6033 (gimple_omp_atomic_load_set_lhs): Likewise.
6034 (gimple_omp_atomic_load_lhs): Likewise.
6035 (gimple_omp_atomic_load_lhs_ptr): Likewise.
6036 (gimple_omp_atomic_load_set_rhs): Likewise.
6037 (gimple_omp_atomic_load_rhs): Likewise.
6038 (gimple_omp_atomic_load_rhs_ptr): Likewise.
6039 (gimple_omp_continue_control_def): Likewise.
6040 (gimple_omp_continue_control_def_ptr): Likewise.
6041 (gimple_omp_continue_set_control_def): Likewise.
6042 (gimple_omp_continue_control_use): Likewise.
6043 (gimple_omp_continue_control_use_ptr): Likewise.
6044 (gimple_omp_continue_set_control_use): Likewise.
6045 (gimple_transaction_body_ptr): Likewise.
6046 (gimple_transaction_label): Likewise.
6047 (gimple_transaction_label_ptr): Likewise.
6048 (gimple_transaction_set_body): Likewise.
6049 (gimple_transaction_set_label): Likewise.
6051 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
6052 * ipa-inline-analysis.c (inline_write_summary): Likewise.
6053 * ipa-ref.c (ipa_record_reference): Likewise.
6054 * ipa-reference.c (analyze_function): Likewise.
6055 (ipa_reference_write_optimization_summary): Likewise.
6056 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
6057 (address_taken_from_non_vtable_p): Likewise.
6058 (comdat_can_be_unshared_p_1): Likewise.
6059 * lto-cgraph.c (lto_output_ref): Likewise.
6060 (add_references): Likewise.
6061 (compute_ltrans_boundary): Likewise.
6062 (output_symtab): Likewise.
6063 (input_ref): Likewise.
6064 (input_cgraph_1): Likewise.
6065 (output_cgraph_opt_summary): Likewise.
6066 * lto-streamer-out.c (lto_output): Likewise.
6067 (output_symbol_p): Likewise.
6068 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
6069 (lsei_start_function_in_partition): Likewise.
6070 (lsei_next_variable_in_partition): Likewise.
6071 (lsei_start_variable_in_partition): Likewise.
6072 * symtab.c (insert_to_assembler_name_hash): Likewise.
6073 (unlink_from_assembler_name_hash): Likewise.
6074 (symtab_unregister_node): Likewise.
6075 (symtab_remove_node): Likewise.
6076 (dump_symtab_node): Likewise.
6077 (verify_symtab_base): Likewise.
6078 (verify_symtab_node): Likewise.
6079 (symtab_make_decl_local): Likewise.
6080 (symtab_alias_ultimate_target): Likewise.
6081 (symtab_resolve_alias): Likewise.
6082 (symtab_get_symbol_partitioning_class): Likewise.
6083 * tree-phinodes.c (allocate_phi_node): Likewise.
6084 (reserve_phi_args_for_new_edge): Likewise.
6085 (remove_phi_args): Likewise.
6086 * varpool.c (varpool_node_for_asm): Likewise.
6087 (varpool_remove_unreferenced_decls): Likewise.
6089 2014-04-23 Jeff Law <law@redhat.com>
6091 PR tree-optimization/60902
6092 * tree-ssa-threadedge.c
6093 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
6094 invalidate outputs from statements that do not produce useful
6095 outputs for threading.
6097 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6099 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
6100 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
6101 machine descriptions for Stack Smashing Protector.
6103 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
6105 * aarch64.md (<optab>_rol<mode>3): New pattern.
6106 (<optab>_rolsi3_uxtw): Likewise.
6107 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
6109 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
6111 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
6112 (arm_cortex_a12_tune): Likewise.
6114 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6116 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
6118 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6120 * config/arm/arm.md (arm_rev16si2): New pattern.
6121 (arm_rev16si2_alt): Likewise.
6122 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
6124 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6126 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
6127 (rev16<mode>2_alt): Likewise.
6128 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
6129 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
6130 (aarch_rev16_shleft_mask_imm_p): Likewise.
6131 (aarch_rev16_p_1): Likewise.
6132 (aarch_rev16_p): Likewise.
6133 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
6134 (aarch_rev16_shright_mask_imm_p): Likewise.
6135 (aarch_rev16_shleft_mask_imm_p): Likewise.
6137 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6139 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
6140 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
6142 (cortex_a53_extra_costs): Likewise.
6143 (cortex_a57_extra_costs): Likewise.
6144 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
6145 (cortexa7_extra_costs): Likewise.
6146 (cortexa8_extra_costs): Likewise.
6147 (cortexa12_extra_costs): Likewise.
6148 (cortexa15_extra_costs): Likewise.
6149 (v7m_extra_costs): Likewise.
6150 (arm_new_rtx_costs): Handle BSWAP.
6152 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6154 * config/arm/arm.c (cortexa8_extra_costs): New table.
6155 (arm_cortex_a8_tune): New tuning struct.
6156 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
6158 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6160 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
6162 2014-04-23 Richard Biener <rguenther@suse.de>
6164 * Makefile.in (OBJS): Remove loop-unswitch.o.
6165 * tree-pass.h (make_pass_rtl_unswitch): Remove.
6166 * passes.def (pass_rtl_unswitch): Likewise.
6167 * loop-init.c (gate_rtl_unswitch): Likewise.
6168 (rtl_unswitch): Likewise.
6169 (pass_data_rtl_unswitch): Likewise.
6170 (pass_rtl_unswitch): Likewise.
6171 (make_pass_rtl_unswitch): Likewise.
6172 * rtl.h (reversed_condition): Likewise.
6173 (compare_and_jump_seq): Likewise.
6174 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
6176 * loop-unroll.c (compare_and_jump_seq): Likewise.
6178 2014-04-23 Richard Biener <rguenther@suse.de>
6180 PR tree-optimization/60903
6181 * tree-ssa-loop-im.c (analyze_memory_references): Remove
6182 commented code block.
6183 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
6184 loop flags to newly created BBs and edges.
6186 2014-04-23 Nick Clifton <nickc@redhat.com>
6188 * config/msp430/msp430.c (msp430_handle_option): Move function
6190 (msp430_option_override): Simplify mcu and mcpu option handling.
6191 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
6192 support for -mhwmult command line option.
6193 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
6194 -mhwmult command line option.
6195 (msp430_hwmult_enabled): Delete.
6196 (msp43o_output_labelref): Add support for -mhwmult command line option.
6197 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
6198 (umulsidi3): Likewise.
6199 * config/msp430/msp430.opt (mmcu): Add Report attribute.
6200 (mcpu, mlarge, msmall): Likewise.
6201 (mhwmult): New option.
6202 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
6204 (msp430_is_f5_mcu): Remove prototype.
6205 (msp430_use_f5_series_hwmult): Add prototype.
6206 * config/msp430/msp430-opts.h: New file.
6207 * common/config/msp430: New directory.
6208 * common/config/msp430/msp430-common.c: New file.
6209 * config.gcc (msp430): Remove target_has_targetm_common.
6210 * doc/invoke.texi: Document -mhwmult command line option.
6212 2014-04-23 Nick Clifton <nickc@redhat.com>
6214 * config/i386/cygwin.h (ENDFILE_SPEC): Include
6215 default-manifest.o if it can be found in the search path.
6216 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
6218 2014-04-23 Terry Guo <terry.guo@arm.com>
6220 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
6222 2014-04-23 Richard Biener <rguenther@suse.de>
6225 * tree-inline.c (declare_return_variable): Use mark_addressable.
6227 2014-04-23 Richard Biener <rguenther@suse.de>
6230 * loop-init.c (loop_optimizer_init): Make sure to apply
6231 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
6233 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6236 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
6238 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
6239 if flag_sanitize_undefined_trap_on_error.
6240 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
6241 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
6242 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
6243 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
6244 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
6245 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
6246 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
6247 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
6248 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
6249 * ubsan.c (ubsan_instrument_unreachable): Return
6250 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
6251 (ubsan_expand_null_ifn): Emit __builtin_trap ()
6252 if flag_sanitize_undefined_trap_on_error and
6253 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
6254 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
6255 instrument_bool_enum_load): Emit __builtin_trap () if
6256 flag_sanitize_undefined_trap_on_error and
6257 __builtin_handle_*_abort () if !flag_sanitize_recover.
6258 * doc/invoke.texi (-fsanitize-recover,
6259 -fsanitize-undefined-trap-on-error): Document.
6261 2014-04-22 Christian Bruel <christian.bruel@st.com>
6263 * config/sh/sh.md (mov<mode>): Replace movQIHI.
6264 Force immediates to SImode.
6266 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
6268 * config/nios2/nios2.md (UNSPEC_ROUND): New.
6270 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
6271 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
6272 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
6273 (nios2_fpu_insn): Add entry for round.
6274 (N2FPU_NO_ERRNO_P): Define.
6275 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
6277 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
6279 2014-04-22 Richard Henderson <rth@redhat.com>
6281 * config/aarch64/aarch64 (addti3, subti3): New expanders.
6282 (add<GPI>3_compare0): Remove leading * from name.
6283 (add<GPI>3_carryin): Likewise.
6284 (sub<GPI>3_compare0): Likewise.
6285 (sub<GPI>3_carryin): Likewise.
6286 (<su_optab>mulditi3): New expander.
6287 (multi3): New expander.
6288 (madd<GPI>): Remove leading * from name.
6290 2014-04-22 Martin Jambor <mjambor@suse.cz>
6292 * cgraphclones.c (cgraph_function_versioning): Copy
6293 ipa_transforms_to_apply instead of asserting it is empty.
6295 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
6298 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
6299 on count_exp to get mode.
6301 2014-04-22 Andrew Pinski <apinski@cavium.com>
6303 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6304 Handle TLS for ILP32.
6305 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
6306 (tlsie_small_<mode>): this and handle PTR.
6307 (tlsie_small_sidi): New pattern.
6308 (tlsle_small): Change to an expand to handle ILP32.
6309 (tlsle_small_<mode>): New pattern.
6310 (tlsdesc_small): Rename to ...
6311 (tlsdesc_small_<mode>): this and handle PTR.
6313 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6315 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
6317 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6319 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6320 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
6321 (aarch64_types_signed_poly_qualifiers): Likewise.
6322 (aarch64_types_unsigned_signed_qualifiers): Likewise.
6323 (aarch64_types_poly_signed_qualifiers): Likewise.
6324 (TYPES_REINTERP_SS): Type macro added.
6325 (TYPES_REINTERP_SU): Likewise.
6326 (TYPES_REINTERP_SP): Likewise.
6327 (TYPES_REINTERP_US): Likewise.
6328 (TYPES_REINTERP_PS): Likewise.
6329 (aarch64_fold_builtin): New expression folding added.
6330 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
6331 Declarations removed.
6332 (REINTERP_SS): Declarations added.
6333 (REINTERP_US): Likewise.
6334 (REINTERP_PS): Likewise.
6335 (REINTERP_SU): Likewise.
6336 (REINTERP_SP): Likewise.
6337 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
6338 (vreinterpretq_p8_f64): Likewise.
6339 (vreinterpret_p16_f64): Likewise.
6340 (vreinterpretq_p16_f64): Likewise.
6341 (vreinterpret_f32_f64): Likewise.
6342 (vreinterpretq_f32_f64): Likewise.
6343 (vreinterpret_f64_f32): Likewise.
6344 (vreinterpret_f64_p8): Likewise.
6345 (vreinterpret_f64_p16): Likewise.
6346 (vreinterpret_f64_s8): Likewise.
6347 (vreinterpret_f64_s16): Likewise.
6348 (vreinterpret_f64_s32): Likewise.
6349 (vreinterpret_f64_s64): Likewise.
6350 (vreinterpret_f64_u8): Likewise.
6351 (vreinterpret_f64_u16): Likewise.
6352 (vreinterpret_f64_u32): Likewise.
6353 (vreinterpret_f64_u64): Likewise.
6354 (vreinterpretq_f64_f32): Likewise.
6355 (vreinterpretq_f64_p8): Likewise.
6356 (vreinterpretq_f64_p16): Likewise.
6357 (vreinterpretq_f64_s8): Likewise.
6358 (vreinterpretq_f64_s16): Likewise.
6359 (vreinterpretq_f64_s32): Likewise.
6360 (vreinterpretq_f64_s64): Likewise.
6361 (vreinterpretq_f64_u8): Likewise.
6362 (vreinterpretq_f64_u16): Likewise.
6363 (vreinterpretq_f64_u32): Likewise.
6364 (vreinterpretq_f64_u64): Likewise.
6365 (vreinterpret_s64_f64): Likewise.
6366 (vreinterpretq_s64_f64): Likewise.
6367 (vreinterpret_u64_f64): Likewise.
6368 (vreinterpretq_u64_f64): Likewise.
6369 (vreinterpret_s8_f64): Likewise.
6370 (vreinterpretq_s8_f64): Likewise.
6371 (vreinterpret_s16_f64): Likewise.
6372 (vreinterpretq_s16_f64): Likewise.
6373 (vreinterpret_s32_f64): Likewise.
6374 (vreinterpretq_s32_f64): Likewise.
6375 (vreinterpret_u8_f64): Likewise.
6376 (vreinterpretq_u8_f64): Likewise.
6377 (vreinterpret_u16_f64): Likewise.
6378 (vreinterpretq_u16_f64): Likewise.
6379 (vreinterpret_u32_f64): Likewise.
6380 (vreinterpretq_u32_f64): Likewise.
6382 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6384 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6385 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
6386 (vreinterpret_p8_s8): Likewise.
6387 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
6388 (vreinterpret_p8_s16): Likewise.
6389 (vreinterpret_p8_s32): Likewise.
6390 (vreinterpret_p8_s64): Likewise.
6391 (vreinterpret_p8_f32): Likewise.
6392 (vreinterpret_p8_u8): Likewise.
6393 (vreinterpret_p8_u16): Likewise.
6394 (vreinterpret_p8_u32): Likewise.
6395 (vreinterpret_p8_u64): Likewise.
6396 (vreinterpret_p8_p16): Likewise.
6397 (vreinterpretq_p8_s8): Likewise.
6398 (vreinterpretq_p8_s16): Likewise.
6399 (vreinterpretq_p8_s32): Likewise.
6400 (vreinterpretq_p8_s64): Likewise.
6401 (vreinterpretq_p8_f32): Likewise.
6402 (vreinterpretq_p8_u8): Likewise.
6403 (vreinterpretq_p8_u16): Likewise.
6404 (vreinterpretq_p8_u32): Likewise.
6405 (vreinterpretq_p8_u64): Likewise.
6406 (vreinterpretq_p8_p16): Likewise.
6407 (vreinterpret_p16_s8): Likewise.
6408 (vreinterpret_p16_s16): Likewise.
6409 (vreinterpret_p16_s32): Likewise.
6410 (vreinterpret_p16_s64): Likewise.
6411 (vreinterpret_p16_f32): Likewise.
6412 (vreinterpret_p16_u8): Likewise.
6413 (vreinterpret_p16_u16): Likewise.
6414 (vreinterpret_p16_u32): Likewise.
6415 (vreinterpret_p16_u64): Likewise.
6416 (vreinterpret_p16_p8): Likewise.
6417 (vreinterpretq_p16_s8): Likewise.
6418 (vreinterpretq_p16_s16): Likewise.
6419 (vreinterpretq_p16_s32): Likewise.
6420 (vreinterpretq_p16_s64): Likewise.
6421 (vreinterpretq_p16_f32): Likewise.
6422 (vreinterpretq_p16_u8): Likewise.
6423 (vreinterpretq_p16_u16): Likewise.
6424 (vreinterpretq_p16_u32): Likewise.
6425 (vreinterpretq_p16_u64): Likewise.
6426 (vreinterpretq_p16_p8): Likewise.
6427 (vreinterpret_f32_s8): Likewise.
6428 (vreinterpret_f32_s16): Likewise.
6429 (vreinterpret_f32_s32): Likewise.
6430 (vreinterpret_f32_s64): Likewise.
6431 (vreinterpret_f32_u8): Likewise.
6432 (vreinterpret_f32_u16): Likewise.
6433 (vreinterpret_f32_u32): Likewise.
6434 (vreinterpret_f32_u64): Likewise.
6435 (vreinterpret_f32_p8): Likewise.
6436 (vreinterpret_f32_p16): Likewise.
6437 (vreinterpretq_f32_s8): Likewise.
6438 (vreinterpretq_f32_s16): Likewise.
6439 (vreinterpretq_f32_s32): Likewise.
6440 (vreinterpretq_f32_s64): Likewise.
6441 (vreinterpretq_f32_u8): Likewise.
6442 (vreinterpretq_f32_u16): Likewise.
6443 (vreinterpretq_f32_u32): Likewise.
6444 (vreinterpretq_f32_u64): Likewise.
6445 (vreinterpretq_f32_p8): Likewise.
6446 (vreinterpretq_f32_p16): Likewise.
6447 (vreinterpret_s64_s8): Likewise.
6448 (vreinterpret_s64_s16): Likewise.
6449 (vreinterpret_s64_s32): Likewise.
6450 (vreinterpret_s64_f32): Likewise.
6451 (vreinterpret_s64_u8): Likewise.
6452 (vreinterpret_s64_u16): Likewise.
6453 (vreinterpret_s64_u32): Likewise.
6454 (vreinterpret_s64_u64): Likewise.
6455 (vreinterpret_s64_p8): Likewise.
6456 (vreinterpret_s64_p16): Likewise.
6457 (vreinterpretq_s64_s8): Likewise.
6458 (vreinterpretq_s64_s16): Likewise.
6459 (vreinterpretq_s64_s32): Likewise.
6460 (vreinterpretq_s64_f32): Likewise.
6461 (vreinterpretq_s64_u8): Likewise.
6462 (vreinterpretq_s64_u16): Likewise.
6463 (vreinterpretq_s64_u32): Likewise.
6464 (vreinterpretq_s64_u64): Likewise.
6465 (vreinterpretq_s64_p8): Likewise.
6466 (vreinterpretq_s64_p16): Likewise.
6467 (vreinterpret_u64_s8): Likewise.
6468 (vreinterpret_u64_s16): Likewise.
6469 (vreinterpret_u64_s32): Likewise.
6470 (vreinterpret_u64_s64): Likewise.
6471 (vreinterpret_u64_f32): Likewise.
6472 (vreinterpret_u64_u8): Likewise.
6473 (vreinterpret_u64_u16): Likewise.
6474 (vreinterpret_u64_u32): Likewise.
6475 (vreinterpret_u64_p8): Likewise.
6476 (vreinterpret_u64_p16): Likewise.
6477 (vreinterpretq_u64_s8): Likewise.
6478 (vreinterpretq_u64_s16): Likewise.
6479 (vreinterpretq_u64_s32): Likewise.
6480 (vreinterpretq_u64_s64): Likewise.
6481 (vreinterpretq_u64_f32): Likewise.
6482 (vreinterpretq_u64_u8): Likewise.
6483 (vreinterpretq_u64_u16): Likewise.
6484 (vreinterpretq_u64_u32): Likewise.
6485 (vreinterpretq_u64_p8): Likewise.
6486 (vreinterpretq_u64_p16): Likewise.
6487 (vreinterpret_s8_s16): Likewise.
6488 (vreinterpret_s8_s32): Likewise.
6489 (vreinterpret_s8_s64): Likewise.
6490 (vreinterpret_s8_f32): Likewise.
6491 (vreinterpret_s8_u8): Likewise.
6492 (vreinterpret_s8_u16): Likewise.
6493 (vreinterpret_s8_u32): Likewise.
6494 (vreinterpret_s8_u64): Likewise.
6495 (vreinterpret_s8_p8): Likewise.
6496 (vreinterpret_s8_p16): Likewise.
6497 (vreinterpretq_s8_s16): Likewise.
6498 (vreinterpretq_s8_s32): Likewise.
6499 (vreinterpretq_s8_s64): Likewise.
6500 (vreinterpretq_s8_f32): Likewise.
6501 (vreinterpretq_s8_u8): Likewise.
6502 (vreinterpretq_s8_u16): Likewise.
6503 (vreinterpretq_s8_u32): Likewise.
6504 (vreinterpretq_s8_u64): Likewise.
6505 (vreinterpretq_s8_p8): Likewise.
6506 (vreinterpretq_s8_p16): Likewise.
6507 (vreinterpret_s16_s8): Likewise.
6508 (vreinterpret_s16_s32): Likewise.
6509 (vreinterpret_s16_s64): Likewise.
6510 (vreinterpret_s16_f32): Likewise.
6511 (vreinterpret_s16_u8): Likewise.
6512 (vreinterpret_s16_u16): Likewise.
6513 (vreinterpret_s16_u32): Likewise.
6514 (vreinterpret_s16_u64): Likewise.
6515 (vreinterpret_s16_p8): Likewise.
6516 (vreinterpret_s16_p16): Likewise.
6517 (vreinterpretq_s16_s8): Likewise.
6518 (vreinterpretq_s16_s32): Likewise.
6519 (vreinterpretq_s16_s64): Likewise.
6520 (vreinterpretq_s16_f32): Likewise.
6521 (vreinterpretq_s16_u8): Likewise.
6522 (vreinterpretq_s16_u16): Likewise.
6523 (vreinterpretq_s16_u32): Likewise.
6524 (vreinterpretq_s16_u64): Likewise.
6525 (vreinterpretq_s16_p8): Likewise.
6526 (vreinterpretq_s16_p16): Likewise.
6527 (vreinterpret_s32_s8): Likewise.
6528 (vreinterpret_s32_s16): Likewise.
6529 (vreinterpret_s32_s64): Likewise.
6530 (vreinterpret_s32_f32): Likewise.
6531 (vreinterpret_s32_u8): Likewise.
6532 (vreinterpret_s32_u16): Likewise.
6533 (vreinterpret_s32_u32): Likewise.
6534 (vreinterpret_s32_u64): Likewise.
6535 (vreinterpret_s32_p8): Likewise.
6536 (vreinterpret_s32_p16): Likewise.
6537 (vreinterpretq_s32_s8): Likewise.
6538 (vreinterpretq_s32_s16): Likewise.
6539 (vreinterpretq_s32_s64): Likewise.
6540 (vreinterpretq_s32_f32): Likewise.
6541 (vreinterpretq_s32_u8): Likewise.
6542 (vreinterpretq_s32_u16): Likewise.
6543 (vreinterpretq_s32_u32): Likewise.
6544 (vreinterpretq_s32_u64): Likewise.
6545 (vreinterpretq_s32_p8): Likewise.
6546 (vreinterpretq_s32_p16): Likewise.
6547 (vreinterpret_u8_s8): Likewise.
6548 (vreinterpret_u8_s16): Likewise.
6549 (vreinterpret_u8_s32): Likewise.
6550 (vreinterpret_u8_s64): Likewise.
6551 (vreinterpret_u8_f32): Likewise.
6552 (vreinterpret_u8_u16): Likewise.
6553 (vreinterpret_u8_u32): Likewise.
6554 (vreinterpret_u8_u64): Likewise.
6555 (vreinterpret_u8_p8): Likewise.
6556 (vreinterpret_u8_p16): Likewise.
6557 (vreinterpretq_u8_s8): Likewise.
6558 (vreinterpretq_u8_s16): Likewise.
6559 (vreinterpretq_u8_s32): Likewise.
6560 (vreinterpretq_u8_s64): Likewise.
6561 (vreinterpretq_u8_f32): Likewise.
6562 (vreinterpretq_u8_u16): Likewise.
6563 (vreinterpretq_u8_u32): Likewise.
6564 (vreinterpretq_u8_u64): Likewise.
6565 (vreinterpretq_u8_p8): Likewise.
6566 (vreinterpretq_u8_p16): Likewise.
6567 (vreinterpret_u16_s8): Likewise.
6568 (vreinterpret_u16_s16): Likewise.
6569 (vreinterpret_u16_s32): Likewise.
6570 (vreinterpret_u16_s64): Likewise.
6571 (vreinterpret_u16_f32): Likewise.
6572 (vreinterpret_u16_u8): Likewise.
6573 (vreinterpret_u16_u32): Likewise.
6574 (vreinterpret_u16_u64): Likewise.
6575 (vreinterpret_u16_p8): Likewise.
6576 (vreinterpret_u16_p16): Likewise.
6577 (vreinterpretq_u16_s8): Likewise.
6578 (vreinterpretq_u16_s16): Likewise.
6579 (vreinterpretq_u16_s32): Likewise.
6580 (vreinterpretq_u16_s64): Likewise.
6581 (vreinterpretq_u16_f32): Likewise.
6582 (vreinterpretq_u16_u8): Likewise.
6583 (vreinterpretq_u16_u32): Likewise.
6584 (vreinterpretq_u16_u64): Likewise.
6585 (vreinterpretq_u16_p8): Likewise.
6586 (vreinterpretq_u16_p16): Likewise.
6587 (vreinterpret_u32_s8): Likewise.
6588 (vreinterpret_u32_s16): Likewise.
6589 (vreinterpret_u32_s32): Likewise.
6590 (vreinterpret_u32_s64): Likewise.
6591 (vreinterpret_u32_f32): Likewise.
6592 (vreinterpret_u32_u8): Likewise.
6593 (vreinterpret_u32_u16): Likewise.
6594 (vreinterpret_u32_u64): Likewise.
6595 (vreinterpret_u32_p8): Likewise.
6596 (vreinterpret_u32_p16): Likewise.
6597 (vreinterpretq_u32_s8): Likewise.
6598 (vreinterpretq_u32_s16): Likewise.
6599 (vreinterpretq_u32_s32): Likewise.
6600 (vreinterpretq_u32_s64): Likewise.
6601 (vreinterpretq_u32_f32): Likewise.
6602 (vreinterpretq_u32_u8): Likewise.
6603 (vreinterpretq_u32_u16): Likewise.
6604 (vreinterpretq_u32_u64): Likewise.
6605 (vreinterpretq_u32_p8): Likewise.
6606 (vreinterpretq_u32_p16): Likewise.
6608 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6610 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
6612 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
6614 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
6615 (vqnegd_s64): Likewise.
6616 (vqabs_s64): Likewise.
6617 (vqabsd_s64): Likewise.
6619 2014-04-22 Richard Henderson <rth@redhat.com>
6621 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
6622 computation to the top of the loop.
6624 2014-04-22 Renlin <renlin.li@arm.com>
6625 Jiong Wang <jiong.wang@arm.com>
6627 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
6628 * config/aarch64/aarch64.c (aarch64_layout_frame)
6629 (aarch64_initial_elimination_offset): Likewise.
6631 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
6633 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
6636 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
6638 * machmode.h (bitwise_mode_for_mode): Declare.
6639 * stor-layout.h (bitwise_type_for_mode): Likewise.
6640 * stor-layout.c (bitwise_mode_for_mode): New function.
6641 (bitwise_type_for_mode): Likewise.
6642 * builtins.c (fold_builtin_memory_op): Use it instead of
6643 int_mode_for_mode and build_nonstandard_integer_type.
6645 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6647 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
6648 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
6649 (*-*-solaris2*): Simplify.
6650 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
6651 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
6652 *-*-solaris2.9* handling.
6654 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
6656 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
6657 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
6659 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
6660 * configure: Regenerate.
6662 * config/i386/sol2-9.h: Remove.
6664 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
6665 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
6666 Remove Solaris 9 references.
6668 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
6670 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
6671 (floatuns<GPI:mode><GPF:mode>2): Remove.
6672 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
6673 and floatuns conversions.
6674 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
6675 and floatuns conversions.
6676 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
6677 (w1,w2): New mode attributes for inequal width conversions.
6679 2014-04-22 Renlin Li <Renlin.Li@arm.com>
6681 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
6682 the output asm format.
6684 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
6686 * config/aarch64/aarch64-simd.md
6687 (aarch64_cm<optab>di): Always split.
6688 (*aarch64_cm<optab>di): New.
6689 (aarch64_cmtstdi): Always split.
6690 (*aarch64_cmtstdi): New.
6692 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6694 PR tree-optimization/60823
6695 * omp-low.c (ipa_simd_modify_function_body): Go through
6696 all SSA_NAMEs and for those refering to vector arguments
6697 which are going to be replaced adjust SSA_NAME_VAR and,
6698 if it is a default definition, change it into a non-default
6699 definition assigned at the beginning of function from new_decl.
6700 (ipa_simd_modify_stmt_ops): Rewritten.
6701 * tree-dfa.c (set_ssa_default_def): When removing default def,
6702 check for NULL loc instead of NULL *loc.
6704 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6706 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
6707 restrictions on core registers for DImode values in Thumb2.
6709 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6711 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
6712 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
6714 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6716 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
6717 (*iordi_notzesidi_di): Likewise.
6718 (*iordi_notsesidi_di): Likewise.
6720 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6722 * config/arm/arm-protos.h (tune_params): New struct members.
6723 * config/arm/arm.c: Initialise tune_params per processor.
6724 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
6725 for speed, based on new tune_params.
6727 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6729 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
6730 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
6731 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
6732 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
6733 * config/aarch64/arm_neon.h (vrnd_f64): Added.
6734 (vrnda_f64): Likewise.
6735 (vrndi_f64): Likewise.
6736 (vrndm_f64): Likewise.
6737 (vrndn_f64): Likewise.
6738 (vrndp_f64): Likewise.
6739 (vrndx_f64): Likewise.
6741 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6743 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
6744 GET_MODE_SIZE argument is enum machine_mode.
6746 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6749 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
6750 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
6752 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
6755 * asan.c (asan_emit_stack_protection): Force the base to align to
6756 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
6757 appropriate bits if STRICT_ALIGNMENT.
6758 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
6760 (expand_used_vars): Leave a space in the stack frame for alignment
6761 if STRICT_ALIGNMENT.
6763 2014-04-21 David Malcolm <dmalcolm@redhat.com>
6765 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
6767 (gimple_store_p): Likewise.
6768 (gimple_assign_load_p): Likewise.
6769 (gimple_assign_cast_p): Likewise.
6770 (gimple_clobber_p): Likewise.
6772 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
6773 rather than a gimple.
6774 (gimple_assign_cast_p): Likewise.
6776 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
6779 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
6780 If mode is DDmode and TARGET_E500_DOUBLE allow move.
6782 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
6783 more debug information for E500 if -mdebug=reg.
6785 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
6788 * config/i386/i386.c (ix86_expand_builtin)
6789 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
6790 register for target RTX.
6791 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
6793 2014-04-18 Cong Hou <congh@google.com>
6795 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
6796 the widen-mult pattern by handling two operands with different sizes,
6797 and operands whose size is smaller than half of the result type.
6799 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6801 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
6802 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
6803 (do_estimate_edge_time): Compute it.
6804 * ipa-inline.c (want_inline_small_function_p): Bypass
6805 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
6807 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6809 * ipa-inline.c (spec_rem): New static variable.
6810 (dump_overall_stats): New function.
6811 (dump_inline_stats): New function.
6813 2014-04-18 Richard Henderson <rth@redhat.com>
6815 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
6816 to GET_MODE_SIZE, not a reg_class_t.
6818 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6820 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
6821 (vsx_xxmrglw_<mode>): Likewise.
6823 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
6826 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
6827 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
6828 (rs6000_init_hard_regno_mode_ok): Likewise.
6830 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
6832 * ipa-inline.c (inline_small_functions): Account only non-cold
6834 * doc/invoke.texi (inline-unit-growth): Update documentation.
6836 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
6838 * config/rs6000/rs6000.md (addti3, subti3): New.
6840 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
6843 * config/i386/i386.c (ix86_expand_clear): Remove outdated
6844 comment. Check optimize_insn_for_size_p instead of
6845 optimize_insn_for_speed_p.
6847 2014-04-17 Martin Jambor <mjambor@suse.cz>
6849 * gimple-iterator.c (gsi_start_edge): New function.
6850 * gimple-iterator.h (gsi_start_edge): Declare.
6851 * tree-sra.c (single_non_eh_succ): New function.
6852 (disqualify_ops_if_throwing_stmt): Renamed to
6853 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
6854 having one non-EH successor BB.
6855 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
6856 generate loads into replacements.
6857 (sra_modify_assign): Likewise and and also use the simple path for
6859 (sra_modify_function_body): Commit statements on edges.
6861 2014-04-17 Richard Biener <rguenther@suse.de>
6864 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
6865 comparison results and add clarifying comment.
6867 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6869 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
6870 (blank_mode): Initialize it.
6871 (emit_mode_size_inline, emit_mode_nunits_inline,
6872 emit_mode_inner_inline): New functions.
6873 (emit_insn_modes_h): Call them and surround their output with
6874 #if GCC_VERSION >= 4001 ... #endif.
6875 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
6876 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
6877 mode_* arrays if the argument is __builtin_constant_p.
6878 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
6879 is enum machine_mode.
6881 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6883 * passes.c (opt_pass::execute): Adjust.
6884 (pass_manager::execute_pass_mode_switching): Likewise.
6885 (early_local_passes::execute): Likewise.
6886 (execute_one_pass): Pass cfun to the pass's execute method.
6887 * tree-pass.h (opt_pass::execute): Add function * argument.
6888 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6889 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6890 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6891 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6892 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6893 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
6894 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
6895 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6896 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6897 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
6898 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
6899 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
6900 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
6901 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
6902 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6903 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6904 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6905 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
6906 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
6907 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6908 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6909 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6910 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6911 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6912 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6913 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6914 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6915 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6916 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6919 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6921 * passes.c (opt_pass::gate): Take function * argument.
6922 (gate_all_early_local_passes): Merge into
6923 (early_local_passes::gate): this.
6924 (gate_all_early_optimizations): Merge into
6925 (all_early_optimizations::gate): this.
6926 (gate_all_optimizations): Mege into
6927 (all_optimizations::gate): this.
6928 (gate_all_optimizations_g): Merge into
6929 (all_optimizations_g::gate): this.
6930 (gate_rest_of_compilation): Mege into
6931 (rest_of_compilation::gate): this.
6932 (gate_postreload): Merge into
6933 (postreload::gate): this.
6934 (dump_one_pass): Pass cfun to the pass's gate method.
6935 (execute_ipa_summary_passes): Likewise.
6936 (execute_one_pass): Likewise.
6937 (ipa_write_summaries_2): Likewise.
6938 (ipa_write_optimization_summaries_1): Likewise.
6939 (ipa_read_summaries_1): Likewise.
6940 (ipa_read_optimization_summaries_1): Likewise.
6941 (execute_ipa_stmt_fixups): Likewise.
6942 * tree-pass.h (opt_pass::gate): Add function * argument.
6943 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
6944 combine-stack-adj.c, combine.c, compare-elim.c,
6945 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6946 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
6947 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
6948 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
6949 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6950 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6951 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6952 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
6953 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
6954 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
6955 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6956 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6957 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
6958 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6959 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6960 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6961 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6962 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6963 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6964 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6965 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6966 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6967 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6968 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
6969 var-tracking.c, vtable-verify.c, web.c: Adjust.
6971 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6973 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
6974 * configure: Regenerate.
6976 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6978 * passes.c (dump_one_pass): don't check pass->has_gate.
6979 (execute_ipa_summary_passes): Likewise.
6980 (execute_one_pass): Likewise.
6981 (ipa_write_summaries_2): Likewise.
6982 (ipa_write_optimization_summaries_1): Likewise.
6983 (ipa_read_optimization_summaries_1): Likewise.
6984 (execute_ipa_stmt_fixups): Likewise.
6985 * tree-pass.h (pass_data::has_gate): Remove.
6986 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6987 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6988 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6989 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6990 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6991 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
6992 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
6993 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
6994 gimple-low.c, gimple-ssa-isolate-paths.c,
6995 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
6996 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6997 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6998 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
6999 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
7000 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
7001 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
7002 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
7003 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
7004 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
7005 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
7006 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
7007 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7008 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
7009 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
7010 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
7011 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7012 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7013 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
7014 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
7015 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
7016 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
7019 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7021 * pass_manager.h (pass_manager::register_dump_files_1): Remove
7023 * passes.c (pass_manager::register_dump_files_1): Merge into
7024 (pass_manager::register_dump_files): this, and remove its handling of
7025 properties since the pass always has the properties anyway.
7026 (pass_manager::pass_manager): Adjust.
7028 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
7030 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
7031 * passes.c (pass_manager::register_dump_files_1): Remove dead code
7032 dealing with properties.
7033 (pass_manager::register_dump_files): Adjust.
7035 2014-03-20 Mark Wielaard <mjw@redhat.com>
7037 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
7038 then represent the bound as normal constant value.
7040 2014-04-17 Jakub Jelinek <jakub@redhat.com>
7043 Forward port from 4.8 branch
7044 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
7046 * config/i386/bmiintrin.h (_blsi_u32): New.
7050 (_blsmsk_u32): Ditto.
7051 (_blsmsk_u64): Ditto.
7052 (_tzcnt_u32): Ditto.
7053 (_tzcnt_u64): Ditto.
7055 2014-04-17 Kito Cheng <kito@0xlab.org>
7057 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
7059 2014-04-17 Richard Biener <rguenther@suse.de>
7062 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
7063 boolean results for comparisons.
7065 2014-04-17 Richard Biener <rguenther@suse.de>
7067 PR tree-optimization/60836
7068 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
7069 initial PHI args to be gimple values.
7071 2014-04-17 Richard Biener <rguenther@suse.de>
7073 PR tree-optimization/60841
7074 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
7075 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
7076 of stmts to SLP build.
7077 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
7078 (vect_analyze_slp): Likewise.
7079 (vect_analyze_slp_instance): Likewise.
7080 (vect_build_slp_tree): Limit overall SLP tree growth.
7081 * tree-vectorizer.h (vect_analyze_data_refs,
7082 vect_analyze_slp): Adjust prototypes.
7084 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
7086 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
7089 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
7091 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
7092 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
7093 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
7094 for TARGET_SLOW_PSHUFB
7096 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
7098 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
7099 * config/i386/i386.c (intel_cost): Ditto.
7101 2014-04-17 Joey Ye <joey.ye@arm.com>
7103 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
7105 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
7107 * opts.c (common_handle_option): Disable -fipa-reference coorectly
7110 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
7112 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
7113 (type_all_derivations_known_p): New predicate.
7114 (type_all_ctors_visible_p): New predicate.
7115 (type_possibly_instantiated_p): New predicate.
7116 (get_odr_type): Compute all_derivations_known.
7117 (dump_odr_type): Dump the flag.
7118 (maybe_record_type): Cleanup.
7119 (record_target_from_binfo): Add bases_to_consider array;
7120 record bases for types w/o instances and skip CXX destructor.
7121 (possible_polymorphic_call_targets_1): Add bases_to_consider
7122 and consider_construction parameters; check if type may have instance.
7123 (get_polymorphic_call_info): Set maybe_in_construction to true
7124 when we know nothing.
7125 (record_targets_from_bases): Skip CXX destructors; they are
7126 never called for types in construction.
7127 (possible_polymorphic_call_targets): Do not record target when
7128 type may not have instance.
7130 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
7133 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
7134 external aliases alive, too.
7136 2014-04-16 Andrew Pinski <apinski@cavium.com>
7138 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
7141 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
7143 * final.c (compute_alignments): Do not apply loop alignment to a block
7144 falling through to the exit.
7146 2014-04-16 Catherine Moore <clm@codesourcery.com>
7148 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
7149 Adjust constraints for microMIPS store patterns.
7151 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
7153 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
7155 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
7157 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
7158 (append_use): Run at -O0.
7159 (append_vdef): Likewise.
7160 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
7161 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
7163 2014-04-16 Jakub Jelinek <jakub@redhat.com>
7165 PR tree-optimization/60844
7166 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
7167 (propagate_op_to_single_use, remove_visited_stmt_chain,
7168 linearize_expr, repropagate_negates, reassociate_bb): Use it
7169 instead of gsi_remove.
7171 2014-04-16 Martin Jambor <mjambor@suse.cz>
7173 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
7174 ipa_transforms_to_apply.
7175 (cgraph_function_versioning): Assert that old_node has empty
7176 ipa_transforms_to_apply.
7177 * trans-mem.c (ipa_tm_create_version): Likewise.
7178 * tree-inline.c (tree_function_versioning): Do not duplicate
7179 ipa_transforms_to_apply.
7181 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7184 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
7186 Pass necessary as flags on 64-bit Solaris/x86.
7187 Use lowercase relocs for x86_64-*-*.
7188 * configure: Regenerate.
7190 2014-04-15 Jan Hubicka <jh@suse.cz>
7192 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
7193 (maybe_record_node, likely_target_p): Use it.
7195 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7198 Revert following patch
7200 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7203 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7204 software floating point or no floating point registers, do not
7205 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7206 in GPRs that occurs after we tested for GPRs that would never be
7209 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7210 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7211 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7212 specifically allow DDmode, since that does not use the SPE SIMD
7215 2014-03-21 Mark Wielaard <mjw@redhat.com>
7217 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
7218 as unsigned or int depending on type and value used.
7220 2014-04-15 Richard Biener <rguenther@suse.de>
7222 PR rtl-optimization/56965
7223 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
7224 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
7226 * alias.c (true_dependence_1): Do not call
7227 nonoverlapping_component_refs_p.
7228 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
7229 nonoverlapping_component_refs_p.
7230 (indirect_refs_may_alias_p): Likewise.
7232 2014-04-15 Teresa Johnson <tejohnson@google.com>
7234 * cfg.c (dump_bb_info): Fix flags check.
7235 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
7237 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7239 PR rtl-optimization/60663
7240 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
7243 2014-04-15 Richard Biener <rguenther@suse.de>
7245 * lto-streamer.h (LTO_major_version): Bump to 4.
7247 2014-04-15 Richard Biener <rguenther@suse.de>
7249 * common.opt (lto_partition_model): New enum.
7250 (flto-partition=): Merge separate options with a single with argument,
7251 add -flto-partition=one support.
7252 * flag-types.h (enum lto_partition_model): Declare.
7253 * opts.c (finish_options): Remove duplicate -flto-partition=
7255 * lto-wrapper.c (run_gcc): Adjust.
7257 2014-04-15 Richard Biener <rguenther@suse.de>
7259 * alias.c (ncr_compar): New function.
7260 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
7262 2014-04-15 Richard Biener <rguenther@suse.de>
7264 * alias.c (record_component_aliases): Do not walk BINFOs.
7266 2014-04-15 Richard Biener <rguenther@suse.de>
7268 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7269 Add struct function argument and adjust.
7270 (find_func_aliases_for_call): Likewise.
7271 (find_func_aliases): Likewise.
7272 (find_func_clobbers): Likewise.
7273 (intra_create_variable_infos): Likewise.
7274 (compute_points_to_sets): Likewise.
7275 (ipa_pta_execute): Adjust. Do not push/pop cfun.
7277 2014-04-15 Richard Biener <rguenther@suse.de>
7279 * tree.c (iterative_hash_expr): Use enum tree_code_class
7280 to store TREE_CODE_CLASS.
7281 (tree_block): Likewise.
7282 (tree_set_block): Likewise.
7283 * tree.h (fold_build_pointer_plus_loc): Use
7284 convert_to_ptrofftype_loc.
7286 2014-04-15 Jakub Jelinek <jakub@redhat.com>
7289 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
7292 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
7294 * cfgloop.h (struct loop): Move force_vectorize down.
7295 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
7296 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
7297 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
7298 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
7299 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
7300 * tree-core.h (enum annot_expr_kind): Add new kind values.
7301 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
7302 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
7304 * tree.def (ANNOTATE_EXPR): Tweak comment.
7306 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7308 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
7311 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
7313 * tree.h (TYPE_IDENTIFIER): Declare.
7314 * tree.c (subrange_type_for_debug_p): Use it.
7315 * godump.c (go_format_type): Likewise.
7316 * dwarf2out.c (is_cxx_auto, modified_type_die,
7317 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
7318 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
7320 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7323 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
7325 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
7327 * config/i386/i386.c (examine_argument): Return bool. Return true if
7328 parameter should be passed in memory.
7329 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
7330 (construct_container): Update calls to examine_argument.
7331 (function_arg_advance_64): Ditto.
7332 (return_in_memory_32): Merge with ix86_return_in_memory.
7333 (return_in_memory_64): Ditto.
7334 (return_in_memory_ms_64): Ditto.
7336 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7338 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
7339 * coverage.c (coverage_compute_profile_id): Handle externally visible
7342 2014-04-14 Martin Jambor <mjambor@suse.cz>
7344 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
7345 DECL_DISREGARD_INLINE_LIMITS functions.
7347 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7350 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
7352 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7355 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
7356 optimize_insn_for_speed_p instead of
7357 optimize_function_for_speed_p.
7359 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
7361 * doc/invoke.texi (free): Document AArch64.
7363 2014-04-14 Richard Biener <rguenther@suse.de>
7365 PR tree-optimization/60042
7366 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
7367 (insert_into_preds_of_block): Do not prevent PHI insertion
7368 for REFERENCE exprs here ...
7369 (eliminate_dom_walker::before_dom_children): ... but prevent
7370 their use here under similar conditions when applied to the
7371 IL after PRE optimizations.
7373 2014-04-14 Richard Biener <rguenther@suse.de>
7375 * passes.def: Move early points-to after early SRA.
7377 2014-04-14 Richard Biener <rguenther@suse.de>
7379 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
7380 check for which sign-changes we allow when forwarding
7381 a converted value into a switch.
7383 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7385 * stor-layout.c (place_field): Finalize non-constant offset for the
7388 2014-04-14 Richard Biener <rguenther@suse.de>
7390 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
7392 (expand_switch_using_bit_tests_p): Likewise.
7393 (process_switch): Compute and pass on speed_p based on the
7395 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
7396 optimize_bb_for_speed_p.
7398 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7400 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
7401 * function.h (struct function): Rename has_force_vect_loops into
7402 has_force_vectorize_loops.
7403 * lto-streamer-in.c (input_cfg): Adjust for renaming.
7404 (input_struct_function_base): Likewise.
7405 * lto-streamer-out.c (output_cfg): Likewise.
7406 (output_struct_function_base): Likewise.
7407 * omp-low.c (expand_omp_simd): Likewise.
7408 * tree-cfg.c (move_sese_region_to_fn): Likewise.
7409 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
7410 (version_loop_for_if_conversion): Likewise.
7411 (tree_if_conversion): Likewise.
7412 (main_tree_if_conversion): Likewise.
7413 (gate_tree_if_conversion): Likewise.
7414 * tree-inline.c (copy_loops): Likewise.
7415 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
7416 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
7417 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
7418 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
7419 * tree-vectorizer.c (vectorize_loops): Likewise.
7420 * tree-vectorizer.h (unlimited_cost_model): Likewise.
7422 2014-04-14 Richard Biener <rguenther@suse.de>
7425 * lto-streamer-out.c (wrap_refs): New function.
7426 (lto_output): Wrap symbol references in global initializes in
7427 type-preserving MEM_REFs.
7429 2014-04-14 Christian Bruel <christian.bruel@st.com>
7431 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
7433 2014-04-14 Christian Bruel <christian.bruel@st.com>
7435 * config/sh/sh.md (setmemqi): New expand pattern.
7436 * config/sh/sh.h (CLEAR_RATIO): Define.
7437 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
7438 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
7440 2014-04-14 Richard Biener <rguenther@suse.de>
7443 * fold-const.c (negate_expr_p): Don't negate directional rounding
7445 (fold_negate_expr): Likewise.
7447 2014-04-14 Richard Biener <rguenther@suse.de>
7449 PR tree-optimization/59817
7450 PR tree-optimization/60453
7451 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
7452 recursion to catch all CHRECs in the scalar evolution and restrict
7453 the predicate for the remains appropriately.
7455 2014-04-12 Catherine Moore <clm@codesourcery.com>
7457 * config/mips/constraints.md: Add new register constraint "kb".
7458 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
7459 (*movhi_internal): Likewise.
7460 (*movqi_internal): Likewise.
7461 * config/mips/mips.h (M16_STORE_REGS): New register class.
7462 (REG_CLASS_NAMES): Add M16_STORE_REGS.
7463 (REG_CLASS_CONTENTS): Likewise.
7464 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
7466 2014-04-11 Tobias Burnus <burnus@net-b.de>
7469 * doc/invoke.texi (-Wformat-signedness): Document it.
7470 (Wformat=2): Mention that this enables -Wformat-signedness.
7472 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7474 * common/config/epiphany/epiphany-common.c
7475 (epiphany_option_optimization_table): Enable section anchors by
7476 default at -O1 or higher.
7477 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
7478 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
7479 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
7480 carries no extra cost.
7481 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
7482 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
7483 * config/epiphany/predicates.md (memclob_operand): New predicate.
7484 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
7485 Use memclob_operand predicate and X constraint for operand 3.
7487 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7489 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
7490 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
7493 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7495 PR rtl-optimization/60651
7496 * mode-switching.c (optimize_mode_switching): Make sure to emit
7497 sets of a lower numbered entity before sets of a higher numbered
7498 entity to a mode of the same or lower priority.
7499 When creating a seginfo for a basic block that starts with a code
7500 label, move the insertion point past the code label.
7501 (new_seginfo): Document and enforce requirement that
7502 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
7503 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
7504 * doc/tm.texi: Regenerate.
7506 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
7509 * config/arc/arc.c (arc_save_restore): Fix assert typo.
7511 2013-04-11 Jakub Jelinek <jakub@redhat.com>
7513 * BASE-VER: Set to 4.10.0.
7515 2014-04-11 Tobias Burnus <burnus@net-b.de>
7518 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
7519 * doc/gcc.texi (Service): Update description in the @menu
7520 * doc/invoke.texi (Option Summary): Remove misplaced and
7523 2014-04-11 Steve Ellcey <sellcey@mips.com>
7524 Jakub Jelinek <jakub@redhat.com>
7527 * expr.c (convert_move): Use emit_store_flag_force instead of
7528 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
7531 2014-04-11 Richard Biener <rguenther@suse.de>
7534 * varasm.c (assemble_alias): Avoid endless error reporting
7535 recursion by setting TREE_ASM_WRITTEN.
7537 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7539 * config/s390/s390.md: Add a splitter for NOT rtx.
7541 2014-04-11 Jakub Jelinek <jakub@redhat.com>
7543 PR rtl-optimization/60663
7544 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
7546 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
7547 Jakub Jelinek <jakub@redhat.com>
7550 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
7551 flag from decl_node to node.
7553 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7556 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
7557 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
7558 ameliorating the cases where it can be.
7560 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
7563 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7565 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7566 (loadsync_<mode>): Change mode.
7567 (load_quadpti, store_quadpti): New.
7568 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7569 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7570 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
7572 2014-04-09 Cong Hou <congh@google.com>
7575 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
7578 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7580 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
7581 instead of vnor to exploit possible fusion opportunity in the
7583 (altivec_expand_vec_perm_const_le): Likewise.
7585 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7587 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7588 (loadsync_<mode>): Change mode.
7589 (load_quadpti, store_quadpti): New.
7590 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7591 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7593 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
7596 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
7597 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
7598 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
7600 2014-04-08 Richard Biener <rguenther@suse.de>
7603 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
7604 a 64bit widest int print double-int similar to on HWI64 hosts.
7606 2014-04-08 Richard Biener <rguenther@suse.de>
7608 PR tree-optimization/60785
7609 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
7610 default defs properly.
7612 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
7614 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
7615 (Weffc++): Likewise.
7617 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
7619 * ipa-devirt.c (maybe_record_node): When node is not recorded,
7620 set completep to false rather than true.
7622 2014-04-07 Douglas B Rupp <rupp@adacore.com>
7625 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
7626 ARM_TARGET2_DWARF_FORMAT.
7628 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
7631 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
7632 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
7635 2014-04-07 Richard Biener <rguenther@suse.de>
7637 PR tree-optimization/60766
7638 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
7639 (may_eliminate_iv): Convert cand_value_at result to desired type.
7641 2014-04-07 Jason Merrill <jason@redhat.com>
7644 * common.opt (-fno-gnu-unique): Add.
7645 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
7647 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7649 * haifa-sched.c: Fix outdated function reference and minor
7650 grammar errors in introductory comment.
7652 2014-04-07 Richard Biener <rguenther@suse.de>
7655 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
7657 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
7659 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
7662 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
7663 size accounting for thunks.
7664 (pa_asm_output_mi_thunk): Use final_start_function() and
7665 final_end_function() to output function start and end directives.
7667 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
7669 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
7670 device specific ISA/ feature information. Remove short_sp and
7671 errata_skip ds. Add avr_device_specific_features enum to have device
7673 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
7674 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
7675 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
7676 updated device specific info.
7677 * config/avr/avr-mcus.def: Merge device specific details to
7678 dev_attribute field.
7679 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
7681 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
7682 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
7683 assembler if RMW isa supported by current device.
7684 * config/avr/genmultilib.awk: Update as device info structure changed.
7685 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
7687 2014-04-04 Cong Hou <congh@google.com>
7689 PR tree-optimization/60656
7690 * tree-vect-stmts.c (supportable_widening_operation):
7691 Fix a bug that elements in a vector with vect_used_by_reduction
7692 property are incorrectly reordered when the operation on it is not
7693 consistant with the one in reduction operation.
7695 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
7697 PR rtl-optimization/60155
7698 * gcse.c (record_set_data): New function.
7699 (single_set_gcse): New function.
7700 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
7701 (hoist_code): Likewise.
7702 (get_pressure_class_and_nregs): Likewise.
7704 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
7706 * explow.c (probe_stack_range): Emit a final optimization blockage.
7708 2014-04-04 Anthony Green <green@moxielogic.com>
7710 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
7713 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
7716 * lto-cgraph.c (input_overwrite_node): Check that partitioning
7717 flags are set only during streaming.
7718 * ipa.c (process_references, walk_polymorphic_call_targets,
7719 symtab_remove_unreachable_nodes): Drop bodies of always inline
7720 after early inlining.
7721 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
7723 2014-04-04 Jakub Jelinek <jakub@redhat.com>
7724 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7727 * dwarf2out.c (const_ok_for_output_1): Reject expressions
7730 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7733 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
7735 (cortex_a53_fdivd): Likewise.
7737 2014-04-04 Martin Jambor <mjambor@suse.cz>
7740 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
7742 * cgraph.c (clone_of_p): Also return true if thunks match.
7743 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
7744 cgraph_function_or_thunk_node and an obsolete comment.
7745 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
7747 (build_function_decl_skip_args): Likewise.
7748 (set_new_clone_decl_and_node_flags): New function.
7749 (duplicate_thunk_for_node): Likewise.
7750 (redirect_edge_duplicating_thunks): Likewise.
7751 (cgraph_clone_node): New parameter args_to_skip, pass it to
7752 redirect_edge_duplicating_thunks which is called instead of
7753 cgraph_redirect_edge_callee.
7754 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
7755 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
7757 2014-04-04 Jeff Law <law@redhat.com>
7760 * config/arm/predicates.md (const_int_I_operand): New predicate.
7761 (const_int_M_operand): Similarly.
7762 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
7764 (insv_t2, extv_reg, extzv_t2): Likewise.
7765 (load_multiple_with_writeback): Similarly for const_int_I_operand.
7766 (pop_multiple_with_writeback_and_return): Likewise.
7767 (vfp_pop_multiple_with_writeback): Likewise
7769 2014-04-04 Richard Biener <rguenther@suse.de>
7772 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
7773 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
7775 * gimplify.h (gimple_add_tmp_var_fn): Declare.
7776 * gimplify.c (gimple_add_tmp_var_fn): New function.
7777 * gimple-expr.h (create_tmp_reg_fn): Declare.
7778 * gimple-expr.c (create_tmp_reg_fn): New function.
7779 * gimple-low.c (record_vars_into): Don't change cfun.
7780 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
7781 code generation without cfun.
7783 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
7786 * Makefile.in (install-driver): Fix shell scripting.
7788 2014-04-03 Cong Hou <congh@google.com>
7790 PR tree-optimization/60505
7791 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
7792 threshold of number of iterations below which no vectorization
7794 * tree-vect-loop.c (new_loop_vec_info):
7795 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
7796 * tree-vect-loop.c (vect_analyze_loop_operations):
7797 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
7798 * tree-vect-loop.c (vect_transform_loop):
7799 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
7800 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
7801 of iterations of the loop and see if we should build the epilogue.
7803 2014-04-03 Richard Biener <rguenther@suse.de>
7805 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
7806 (streamer_tree_cache_create): Adjust.
7807 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
7808 to allow optional nodes array.
7809 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
7810 (streamer_tree_cache_append): Likewise.
7811 (streamer_tree_cache_create): Create nodes array optionally
7812 as specified by parameter.
7813 * lto-streamer-out.c (create_output_block): Avoid maintaining
7814 the node array in the writer cache.
7815 (DFS_write_tree): Remove assertion.
7816 (produce_asm_for_decls): Free the out decl state hash table early.
7817 * lto-streamer-in.c (lto_data_in_create): Adjust for
7818 streamer_tree_cache_create prototype change.
7820 2014-04-03 Richard Biener <rguenther@suse.de>
7822 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
7823 set TREE_CHAIN to NULL_TREE.
7825 2014-04-03 Richard Biener <rguenther@suse.de>
7827 PR tree-optimization/60740
7828 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
7829 over all GIMPLE_COND operands.
7831 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
7833 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
7834 (Weffc++): Remove Scott's numbering, merge lists and reference
7837 2014-04-03 Nick Clifton <nickc@redhat.com>
7839 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
7842 2014-04-03 Martin Jambor <mjambor@suse.cz>
7844 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
7845 mention gcc_unreachable before failing.
7846 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
7849 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
7852 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
7853 inconsistent code and instead mark the context inconsistent.
7854 (possible_polymorphic_call_targets): For inconsistent contexts
7855 return empty complete list.
7857 2014-04-02 Anthony Green <green@moxielogic.com>
7859 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
7860 (extendqisi2, extendhisi2): Define.
7861 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
7862 (WCHAR_TYPE): Change to unsigned int.
7864 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7866 PR tree-optimization/60733
7867 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
7868 insertion point for PHI candidates to be the end of the feeding
7869 block for the PHI argument.
7871 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
7873 PR rtl-optimization/60650
7874 * lra-constraints.c (process_alt_operands): Decrease reject for
7875 earlyclobber matching.
7877 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7879 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
7881 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7883 * config/spu/spu.c (pad_bb): Do not crash when the last
7884 insn is CODE_FOR_blockage.
7886 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7888 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
7889 lies outside the target mode.
7891 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7894 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7895 software floating point or no floating point registers, do not
7896 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7897 in GPRs that occurs after we tested for GPRs that would never be
7900 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7901 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7902 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7903 specifically allow DDmode, since that does not use the SPE SIMD
7906 2014-04-02 Richard Biener <rguenther@suse.de>
7909 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
7910 MODE_INTs. Properly use negv_optab.
7911 (expand_abs): Likewise.
7913 2014-04-02 Richard Biener <rguenther@suse.de>
7916 * Makefile.in (install-driver): Guard extra installs with special
7919 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
7921 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7924 2014-04-01 Richard Henderson <rth@redhat.com>
7927 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
7928 alternative enabled before register allocation.
7930 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
7932 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
7933 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
7935 (nios2_large_got_address): Remove unneeded 'sym' parameter.
7936 (nios2_got_address): Update nios2_large_got_address call site.
7937 (nios2_delegitimize_address): New function.
7938 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
7939 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
7940 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
7942 2014-04-01 Martin Husemann <martin@duskware.de>
7944 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
7947 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
7949 PR rtl-optimization/60604
7950 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
7951 check from register_operand.
7952 (register_operand): Redefine in terms of general_operand.
7953 (nonmemory_operand): Use register_operand for the non-constant cases.
7955 2014-04-01 Richard Biener <rguenther@suse.de>
7957 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
7959 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
7961 * doc/invoke.texi (mapp-regs): Clarify.
7963 2014-03-31 Ulrich Drepper <drepper@gmail.com>
7965 * config/i386/avx512fintrin.h (__v32hi): Define type.
7966 (__v64qi): Likewise.
7967 (_mm512_set1_epi8): Define.
7968 (_mm512_set1_epi16): Define.
7969 (_mm512_set4_epi32): Define.
7970 (_mm512_set4_epi64): Define.
7971 (_mm512_set4_pd): Define.
7972 (_mm512_set4_ps): Define.
7973 (_mm512_setr4_epi64): Define.
7974 (_mm512_setr4_epi32): Define.
7975 (_mm512_setr4_pd): Define.
7976 (_mm512_setr4_ps): Define.
7977 (_mm512_setzero_epi32): Define.
7979 2014-03-31 Martin Jambor <mjambor@suse.cz>
7982 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
7983 callsite_arguments_match_p. Updated all callers. Also check types of
7984 corresponding formal parameters and actual arguments.
7985 (not_all_callers_have_enough_arguments_p) Renamed to
7986 some_callers_have_mismatched_arguments_p.
7988 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
7990 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
7992 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
7995 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
7998 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
8000 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
8001 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
8003 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
8004 Use FMAMODE_NOVF512 mode iterator.
8005 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
8006 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
8007 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
8008 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
8010 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
8011 Use VF_128_256 mode iterator.
8012 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
8015 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
8017 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
8018 static chain if needed.
8020 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
8023 * lra-constraints.c (index_part_to_reg): New.
8024 (process_address): Use it.
8026 2014-03-27 Jeff Law <law@redhat.com>
8027 Jakub Jelinek <jakub@redhat.com>
8030 * expr.c (do_tablejump): Use simplify_gen_binary rather than
8031 gen_rtx_{PLUS,MULT} to build up the address expression.
8033 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
8034 creating non-canonical RTL.
8036 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
8039 * ipa-inline.c (want_inline_small_function_p): Short circuit large
8040 functions; reorganize to make cheap checks first.
8041 (inline_small_functions): Do not estimate growth when dumping;
8043 * ipa-inline.h (inline_summary): Add min_size.
8044 (growth_likely_positive): New function.
8045 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
8046 (set_cond_stmt_execution_predicate): Cleanup.
8047 (estimate_edge_size_and_time): Compute min_size.
8048 (estimate_calls_size_and_time): Likewise.
8049 (estimate_node_size_and_time): Likewise.
8050 (inline_update_overall_summary): Update min_size.
8051 (do_estimate_edge_time): Likewise.
8052 (do_estimate_edge_size): Update.
8053 (do_estimate_edge_hints): Update.
8054 (growth_likely_positive): New function.
8056 2014-03-28 Jakub Jelinek <jakub@redhat.com>
8059 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
8060 also if addr has VOIDmode.
8062 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8064 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
8065 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
8067 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
8068 instructions as well as AdvancedSIMD loads.
8070 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8072 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
8073 Use crypto_aese type.
8074 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
8075 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
8076 crypto_aese, crypto_aesmc. Move to types.md.
8077 * config/arm/types.md (crypto_aes): Split into crypto_aese,
8079 * config/arm/iterators.md (crypto_type): Likewise.
8081 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
8083 * cgraph.c: Include expr.h and tree-dfa.h.
8084 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
8087 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
8090 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
8091 regs from checking multi-reg pseudos.
8093 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8095 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
8097 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8099 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
8100 if it would clobber the stack pointer, even temporarily.
8102 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
8104 * mode-switching.c: Make small adjustments to the top comment.
8106 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
8108 * config/rs6000/constraints.md (wD constraint): New constraint to
8109 match the constant integer to get the top DImode/DFmode out of a
8110 vector in a VSX register.
8112 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
8113 match the constant integer to get the top DImode/DFmode out of a
8114 vector in a VSX register.
8116 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
8119 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8122 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
8123 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
8125 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
8126 Optimize vec_extract of 64-bit values, where the value being
8127 extracted is in the top word, where we can use scalar
8128 instructions. Add direct move and store support. Combine the big
8129 endian/little endian vector select load support into a single insn.
8130 (vsx_extract_<mode>_internal1): Likewise.
8131 (vsx_extract_<mode>_internal2): Likewise.
8132 (vsx_extract_<mode>_load): Likewise.
8133 (vsx_extract_<mode>_store): Likewise.
8134 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
8135 combined into vsx_extract_<mode>_load.
8136 (vsx_extract_<mode>_one_le): Likewise.
8138 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
8139 define the top 64-bit vector element.
8141 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
8144 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8145 Document vec_vbpermq builtin.
8148 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
8149 enable use of xxsldwi and xxpermdi builtin functions.
8150 (vec_xxpermdi): Likewise.
8152 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8153 Document use of vec_xxsldwi and vec_xxpermdi builtins.
8155 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
8157 PR rtl-optimization/60650
8158 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
8160 (find_spills_for): New.
8161 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
8162 Spill all pseudos on the second iteration.
8164 2014-03-27 Marek Polacek <polacek@redhat.com>
8167 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
8170 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8172 * config/s390/s390.c (s390_can_use_return_insn): Check for
8173 call-saved FPRs on 31 bit.
8175 2014-03-27 Jakub Jelinek <jakub@redhat.com>
8178 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
8179 if they need regimplification, just drop them instead of
8180 calling gimple_regimplify_operands on them.
8182 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
8185 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
8186 (aarch64_frame_pointer_required): Adjust logic.
8187 (aarch64_can_eliminate): Adjust logic.
8188 (aarch64_override_options_after_change): Adjust logic.
8190 2014-03-27 Dehao Chen <dehao@google.com>
8192 * ipa-inline.c (early_inliner): Update node's inline info.
8194 2014-03-26 Dehao Chen <dehao@google.com>
8196 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
8197 compiler inserted conditional jumps for NAN float check.
8199 2014-03-26 Jakub Jelinek <jakub@redhat.com>
8201 * ubsan.h (ubsan_create_data): Change second argument's type
8202 to const location_t *.
8203 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
8205 (ubsan_create_data): Change second argument to const location_t *PLOC.
8206 Create Loc field whenever PLOC is non-NULL.
8207 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
8208 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
8212 * real.c (real_to_integer2): Change type of low to UHWI.
8214 2014-03-26 Tobias Burnus <burnus@net-b.de>
8216 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
8217 (CILK_SELF_SPECS): New define.
8218 (driver_self_specs): Use it.
8220 2014-03-26 Richard Biener <rguenther@suse.de>
8222 * tree-pretty-print.c (percent_K_format): Implement special
8223 case for LTO and its stripped down BLOCK tree.
8225 2014-03-26 Jakub Jelinek <jakub@redhat.com>
8228 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
8230 * tree-vrp.c (simplify_internal_call_using_ranges): If only
8231 one range is range_int_cst_p, but not both, at least optimize
8232 addition/subtraction of 0 and multiplication by 0 or 1.
8233 * gimple-fold.c (gimple_fold_call): Fold
8234 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
8235 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
8236 INTEGER_CSTs, try to fold at least x * 0 and y - y.
8238 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
8240 PR rtl-optimization/60452
8241 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
8242 <case REG>: Return 1 for invalid offsets from the frame pointer.
8244 2014-03-26 Marek Polacek <polacek@redhat.com>
8247 * doc/extend.texi (C Extensions): Mention variable-length arrays in
8250 2014-03-26 Marek Polacek <polacek@redhat.com>
8253 * doc/extend.texi (Designated Inits): Describe what happens to omitted
8256 2014-03-26 Marek Polacek <polacek@redhat.com>
8259 * ira-color.c (update_conflict_hard_regno_costs): Perform the
8260 multiplication in unsigned type.
8262 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
8264 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
8266 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
8268 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
8270 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
8273 * cif-code.def (UNREACHABLE) New code.
8274 * ipa-inline.c (inline_small_functions): Skip edges to
8275 __builtlin_unreachable.
8276 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
8277 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
8278 predicate to __bulitin_unreachable.
8279 (set_cond_stmt_execution_predicate): Fix issue when
8280 invert_tree_comparison returns ERROR_MARK.
8281 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
8282 propagate to inline clones.
8283 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
8285 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
8286 * cgraphclones.c (cgraph_clone_node): If call destination is already
8287 ureachable, do not redirect it back.
8288 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
8291 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
8293 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
8294 Do not modify inline clones.
8296 2014-03-25 Jakub Jelinek <jakub@redhat.com>
8298 * config/i386/i386.md (general_sext_operand): New mode attr.
8299 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
8300 don't generate (sign_extend (const_int)).
8301 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
8302 operands[2]. Use We constraint instead of <i> and
8303 <general_sext_operand> predicate instead of <general_operand>.
8304 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
8305 * config/i386/constraints.md (We): New constraint.
8306 * config/i386/predicates.md (x86_64_sext_operand,
8307 sext_operand): New predicates.
8309 2014-03-25 Martin Jambor <mjambor@suse.cz>
8312 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
8313 inconsistent devirtualizations to __builtin_unreachable.
8315 2014-03-25 Marek Polacek <polacek@redhat.com>
8318 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
8320 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
8322 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
8323 order of elements for big-endian.
8325 2014-03-25 Richard Biener <rguenther@suse.de>
8328 * gimplify-me.c (gimple_regimplify_operands): Update the
8331 2014-03-25 Martin Jambor <mjambor@suse.cz>
8334 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
8335 (lto_output_varpool_node): Likewise.
8336 (input_overwrite_node): Likewise.
8337 (input_varpool_node): Likewise.
8339 2014-03-25 Richard Biener <rguenther@suse.de>
8341 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
8342 (run_gcc): Likewise.
8344 2014-03-25 Jakub Jelinek <jakub@redhat.com>
8346 * combine.c (simplify_compare_const): Add MODE argument.
8347 Handle mode_width 0 as very large mode_width.
8348 (try_combine, simplify_comparison): Adjust callers.
8350 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
8351 type to avoid signed integer overflow.
8352 * explow.c (plus_constant): Likewise.
8354 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
8356 * doc/generic.texi: Correct typos.
8358 2014-03-24 Tobias Burnus <burnus@net-b.de>
8360 * doc/invoke.texi (-flto): Expand section about
8361 using static libraries with LTO.
8363 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8365 PR rtl-optimization/60501
8366 * optabs.def (addptr3_optab): New optab.
8367 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
8368 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
8369 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
8371 * lra.c (emit_add3_insn): Use the addptr pattern if available.
8373 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
8375 2014-03-24 Ulrich Drepper <drepper@gmail.com>
8377 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
8380 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
8381 (_mm256_undefined_ps): Define.
8382 (_mm256_undefined_pd): Define.
8383 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
8384 (_mm_undefined_pd): Define.
8385 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
8386 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
8387 (_mm512_undefined_ps): Define.
8388 (_mm512_undefined_pd): Define.
8389 Use _mm*_undefined_*.
8390 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
8392 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
8394 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
8395 (lshr_simd): DI mode added.
8396 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
8397 (aarch64_ushr_simddi): Likewise.
8398 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
8399 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
8400 (vshrd_n_u64): Likewise.
8402 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8404 * Makefile.in (s-macro_list): Depend on cc1.
8406 2014-03-23 Teresa Johnson <tejohnson@google.com>
8408 * ipa-utils.c (ipa_print_order): Use specified dump file.
8410 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
8412 PR rtl-optimization/60601
8413 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
8415 * gcc.c (eval_spec_function): Initialize save_growing_value.
8417 2014-03-22 Jakub Jelinek <jakub@redhat.com>
8420 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
8421 code == MINUS_EXPR, never swap op0 with op1.
8423 * toplev.c (init_local_tick): Avoid signed integer multiplication
8425 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
8426 shift by first operand's bitsize.
8428 2014-03-21 Jakub Jelinek <jakub@redhat.com>
8431 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
8433 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
8434 TARGET_ISA_64BIT_P(x).
8436 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8438 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
8439 pattern for vector nor instead of subtract from splat(-1).
8440 (altivec_expand_vec_perm_const_le): Likewise.
8442 2014-03-21 Richard Henderson <rth@twiddle.net>
8445 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
8446 related insns after epilogue_completed.
8448 2014-03-21 Martin Jambor <mjambor@suse.cz>
8451 * cgraph.h (symtab_node): New flag body_removed.
8452 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
8453 when removing bodies.
8454 * symtab.c (dump_symtab_base): Dump body_removed flag.
8455 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
8456 had their bodies removed.
8458 2014-03-21 Martin Jambor <mjambor@suse.cz>
8461 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
8464 2014-03-21 Richard Biener <rguenther@suse.de>
8466 PR tree-optimization/60577
8467 * tree-core.h (struct tree_base): Document nothrow_flag use
8468 in VAR_DECL_NONALIASED.
8469 * tree.h (VAR_DECL_NONALIASED): New.
8470 (may_be_aliased): Adjust.
8471 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
8473 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8475 * expr.c (expand_expr_real_1): Remove outdated comment.
8477 2014-03-20 Jakub Jelinek <jakub@redhat.com>
8480 * ira.c (adjust_cleared_regs): Call copy_rtx on
8481 *reg_equiv[REGNO (loc)].src_p before passing it to
8482 simplify_replace_fn_rtx.
8485 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
8486 into CONST, put pic register as first operand of PLUS. Use
8487 gen_const_mem for both 32-bit and 64-bit PIC got loads.
8489 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8491 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
8493 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8495 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
8496 around for store forwarding issue in the FPU on the UT699.
8497 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
8498 loads and operations if -mfix-ut699 is specified.
8499 (divtf3_hq): Tweak attribute.
8500 (sqrttf2_hq): Likewise.
8502 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8504 * calls.c (store_one_arg): Remove incorrect const qualification on the
8505 type of the temporary.
8506 * cfgexpand.c (expand_return): Likewise.
8507 * expr.c (expand_constructor): Likewise.
8508 (expand_expr_real_1): Likewise.
8510 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8512 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
8515 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
8518 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
8520 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
8522 * config/arm/aarch-common-protos.h
8523 (alu_cost_table): Fix spelling of "extend".
8524 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
8526 2014-03-19 Richard Biener <rguenther@suse.de>
8529 * tree-core.h (tree_type_common): Re-order pointer members
8530 to reduce recursion depth during GC walks.
8532 2014-03-19 Marek Polacek <polacek@redhat.com>
8535 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
8536 before accessing it.
8538 2014-03-19 Richard Biener <rguenther@suse.de>
8541 * lto-streamer-in.c (input_function): In WPA stage do not drop
8544 2014-03-19 Jakub Jelinek <jakub@redhat.com>
8546 PR tree-optimization/60559
8547 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
8548 with build_zero_cst assignment.
8550 2014-03-18 Kai Tietz <ktietz@redhat.com>
8552 PR rtl-optimization/56356
8553 * sdbout.c (sdbout_parms): Verify that parms'
8554 incoming argument is valid.
8555 (sdbout_reg_parms): Likewise.
8557 2014-03-18 Richard Henderson <rth@redhat.com>
8560 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
8561 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
8562 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
8564 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
8566 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
8567 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
8568 Italicize plugin event names in description. Explain that
8569 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
8570 Remind that no GCC functions should be called after PLUGIN_FINISH.
8571 Explain what pragmas with expansion are.
8573 2014-03-18 Martin Liska <mliska@suse.cz>
8575 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
8576 gimple call statement is update.
8577 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
8578 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
8580 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8583 * ubsan.c (ubsan_instrument_unreachable): Call
8584 initialize_sanitizer_builtins.
8585 (ubsan_pass): Likewise.
8588 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
8589 varpool_finalize_decl instead of rest_of_decl_compilation.
8591 2014-03-18 Richard Biener <rguenther@suse.de>
8593 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
8594 by using bitmap_and_compl instead of bitmap_and_compl_into.
8595 (df_rd_transfer_function): Likewise.
8597 2014-03-18 Richard Biener <rguenther@suse.de>
8599 * doc/lto.texi (fresolution): Fix typo.
8601 2014-03-18 Richard Biener <rguenther@suse.de>
8603 * doc/invoke.texi (flto): Update for changes in 4.9.
8605 2014-03-18 Richard Biener <rguenther@suse.de>
8607 * doc/loop.texi: Remove section on the removed lambda framework.
8608 Update loop docs with recent changes in preserving loop structure.
8610 2014-03-18 Richard Biener <rguenther@suse.de>
8612 * doc/lto.texi (-fresolution): Document.
8614 2014-03-18 Richard Biener <rguenther@suse.de>
8616 * doc/contrib.texi: Adjust my name.
8618 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8621 * internal-fn.c: Include diagnostic-core.h.
8622 (expand_BUILTIN_EXPECT): New function.
8623 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
8624 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
8625 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
8626 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
8628 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
8629 Revert 3 argument __builtin_expect code.
8630 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
8631 * gimple-fold.c (gimple_fold_call): Likewise.
8632 * tree.h (fold_builtin_expect): New prototype.
8633 * builtins.c (build_builtin_expect_predicate): Add predictor
8634 argument, if non-NULL, create 3 argument __builtin_expect.
8635 (fold_builtin_expect): No longer static. Add ARG2 argument,
8636 pass it through to build_builtin_expect_predicate.
8637 (fold_builtin_2): Adjust caller.
8638 (fold_builtin_3): Handle BUILT_IN_EXPECT.
8639 * internal-fn.def (BUILTIN_EXPECT): New.
8641 2014-03-18 Tobias Burnus <burnus@net-b.de>
8644 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
8645 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
8646 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
8648 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
8651 * predict.c (combine_predictions_for_bb): Fix up formatting.
8652 (expr_expected_value_1, expr_expected_value): Add predictor argument,
8653 fill what it points to if non-NULL.
8654 (tree_predict_by_opcode): Adjust caller, use the predictor.
8655 * predict.def (PRED_COMPARE_AND_SWAP): Add.
8657 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
8659 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
8660 proper constant for the store mode.
8662 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
8664 * symtab.c (change_decl_assembler_name): Fix transparent alias
8667 2014-03-16 Renlin Li <Renlin.Li@arm.com>
8669 * config/aarch64/aarch64.c: Correct the comments about the
8670 aarch64 stack layout.
8672 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
8674 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
8675 check for GF_OMP_FOR_KIND_FOR.
8677 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
8679 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
8680 ymm and zmm register names.
8682 2014-03-17 Jakub Jelinek <jakub@redhat.com>
8685 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
8686 note creation for the 2010-08-31 changes.
8688 2014-03-17 Marek Polacek <polacek@redhat.com>
8691 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
8692 as -fno-tree-loop-vectorize.
8693 (expand_omp_simd): Likewise.
8695 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
8697 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
8698 (eligible_for_call_delay): New prototype.
8699 * config/sparc/sparc.c (tls_call_delay): Rename into...
8700 (eligible_for_call_delay): ...this. Return false if the instruction
8701 cannot be put in the delay slot of a branch.
8702 (eligible_for_restore_insn): Simplify.
8703 (eligible_for_return_delay): Return false if the instruction cannot be
8704 put in the delay slot of a branch and simplify.
8705 (eligible_for_sibcall_delay): Return false if the instruction cannot be
8706 put in the delay slot of a branch.
8707 * config/sparc/sparc.md (fix_ut699): New attribute.
8708 (tls_call_delay): Delete.
8709 (in_call_delay): Reimplement.
8710 (eligible_for_sibcall_delay): Rename into...
8711 (in_sibcall_delay): ...this.
8712 (eligible_for_return_delay): Rename into...
8713 (in_return_delay): ...this.
8714 (in_branch_delay): Reimplement.
8715 (in_uncond_branch_delay): Delete.
8716 (in_annul_branch_delay): Delete.
8718 2014-03-14 Richard Henderson <rth@redhat.com>
8721 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
8722 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
8723 (*floathi<X87MODEF>2_i387_with_temp): Remove.
8724 (floathi splitters): Remove.
8725 (float<SWI48x>xf2): New pattern.
8726 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
8727 code that tried to handle DImode for 32-bit, but which was excluded
8728 by the pattern's condition. Drop allocation of stack temporary.
8729 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
8730 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
8731 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
8732 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
8733 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
8734 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
8735 (*float<SWI48><MODEF>2_sse_interunit): Remove.
8736 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
8737 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
8738 (*float<SWI48x><X87MODEF>2_i387): Remove.
8739 (all float _with_temp splitters): Remove.
8740 (*float<SWI48x><MODEF>2_i387): New pattern.
8741 (*float<SWI48><MODEF>2_sse): New pattern.
8742 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
8743 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
8745 2014-03-14 Jakub Jelinek <jakub@redhat.com>
8746 Marek Polacek <polacek@redhat.com>
8749 * common.opt (dump_base_name_prefixed): New Variable.
8750 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
8751 if x_dump_base_name_prefixed is already set, set it at the end.
8753 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
8755 PR rtl-optimization/60508
8756 * lra-constraints.c (get_reload_reg): Add new parameter
8758 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
8759 Pass the new parameter values.
8761 2014-03-14 Richard Biener <rguenther@suse.de>
8763 * common.opt: Revert unintented changes from r205065.
8766 2014-03-14 Richard Biener <rguenther@suse.de>
8769 * cfghooks.c (split_block): Properly adjust all loops the
8770 block was a latch of.
8772 2014-03-14 Martin Jambor <mjambor@suse.cz>
8775 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
8778 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
8781 * config/avr/avr.c (avr_set_current_function): Pass function name
8782 through default_strip_name_encoding before sanity checking instead
8783 of skipping the first char of the assembler name.
8785 2014-03-13 Richard Henderson <rth@redhat.com>
8788 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
8789 (ix86_force_to_memory, ix86_free_from_memory): Remove.
8790 * config/i386/i386-protos.h: Likewise.
8791 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
8792 in the expander instead of a splitter.
8793 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
8794 any possibility of requiring a memory.
8795 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
8796 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
8797 (fp branch splitters): Update for ix86_split_fp_branch.
8798 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
8799 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
8800 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
8801 (*fop_<MODEF>_2_i387): Remove f/r alternative.
8802 (*fop_<MODEF>_3_i387): Likewise.
8803 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
8804 (splitters for the fop_* register patterns): Remove.
8805 (fscalexf4_i387): Rename from *fscalexf4_i387.
8806 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
8808 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8810 PR tree-optimization/59779
8811 * tree-dfa.c (get_ref_base_and_extent): Use double_int
8812 type for bitsize and maxsize instead of HOST_WIDE_INT.
8814 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
8816 PR rtl-optimization/57320
8817 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
8818 the CFG after thread_prologue_and_epilogue_insns.
8820 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
8822 PR rtl-optimization/57189
8823 * lra-constraints.c (process_alt_operands): Disfavor spilling
8826 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
8828 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
8830 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8832 PR tree-optimization/59025
8834 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
8835 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
8837 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
8840 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
8841 calls of avr_out_plus_1.
8843 2014-03-13 Bin Cheng <bin.cheng@arm.com>
8845 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
8846 BB's single pred and update the father loop's latch info later.
8848 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8850 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
8854 (VEC_base): Likewise.
8855 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
8856 registers, we need to swap double words in little endian mode.
8858 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
8859 to be a container mode for 128-bit integer operations added in ISA
8860 2.07. Unlike TImode and PTImode, the preferred register set is
8861 the Altivec/VMX registers for the 128-bit operations.
8863 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
8865 (rs6000_split_128bit_ok_p): Likewise.
8867 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
8868 macros for creating ISA 2.07 normal and overloaded builtin
8869 functions with 3 arguments.
8870 (BU_P8V_OVERLOAD_3): Likewise.
8871 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
8872 for use as overloaded functions.
8873 (VPERM_1TI_UNS): Likewise.
8874 (VSEL_1TI): Likewise.
8875 (VSEL_1TI_UNS): Likewise.
8876 (ST_INTERNAL_1ti): Likewise.
8877 (LD_INTERNAL_1ti): Likewise.
8878 (XXSEL_1TI): Likewise.
8879 (XXSEL_1TI_UNS): Likewise.
8880 (VPERM_1TI): Likewise.
8881 (VPERM_1TI_UNS): Likewise.
8882 (XXPERMDI_1TI): Likewise.
8883 (SET_1TI): Likewise.
8884 (LXVD2X_V1TI): Likewise.
8885 (STXVD2X_V1TI): Likewise.
8886 (VEC_INIT_V1TI): Likewise.
8887 (VEC_SET_V1TI): Likewise.
8888 (VEC_EXT_V1TI): Likewise.
8889 (EQV_V1TI): Likewise.
8890 (NAND_V1TI): Likewise.
8891 (ORC_V1TI): Likewise.
8892 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
8893 added in ISA 2.07. Add both normal 'altivec' builtins, and the
8895 (VADDUQM): Likewise.
8896 (VSUBCUQ): Likewise.
8897 (VADDEUQM): Likewise.
8898 (VADDECUQ): Likewise.
8899 (VSUBEUQM): Likewise.
8900 (VSUBECUQ): Likewise.
8902 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
8903 __int128_t and __uint128_t types.
8904 (__uint128_type): Likewise.
8905 (altivec_categorize_keyword): Add support for vector __int128_t,
8906 vector __uint128_t, vector __int128, and vector unsigned __int128
8907 as a container type for TImode operations that need to be done in
8908 VSX/Altivec registers.
8909 (rs6000_macro_to_expand): Likewise.
8910 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
8911 to support 128-bit integer instructions vaddcuq, vadduqm,
8912 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
8913 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
8915 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
8916 for V1TImode, and set up preferences to use VSX/Altivec registers.
8917 Setup VSX reload handlers.
8918 (rs6000_debug_reg_global): Likewise.
8919 (rs6000_init_hard_regno_mode_ok): Likewise.
8920 (rs6000_preferred_simd_mode): Likewise.
8921 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
8922 (easy_altivec_constant): Likewise.
8923 (output_vec_const_move): Likewise.
8924 (rs6000_expand_vector_set): Convert V1TImode set and extract to
8926 (rs6000_expand_vector_extract): Likewise.
8927 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
8929 (rs6000_const_vec): Add support for V1TImode.
8930 (rs6000_emit_le_vsx_load): Swap double words when loading or
8931 storing TImode/V1TImode.
8932 (rs6000_emit_le_vsx_store): Likewise.
8933 (rs6000_emit_le_vsx_move): Likewise.
8934 (rs6000_emit_move): Add support for V1TImode.
8935 (altivec_expand_ld_builtin): Likewise.
8936 (altivec_expand_st_builtin): Likewise.
8937 (altivec_expand_vec_init_builtin): Likewise.
8938 (altivec_expand_builtin): Likewise.
8939 (rs6000_init_builtins): Add support for V1TImode type. Add
8940 support for ISA 2.07 128-bit integer builtins. Define type names
8941 for the VSX/Altivec vector types.
8942 (altivec_init_builtins): Add support for overloaded vector
8943 functions with V1TImode type.
8944 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
8945 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
8947 (rs6000_split_128bit_ok_p): Likewise.
8948 (rs6000_handle_altivec_attribute): Create V1TImode from vector
8949 __int128_t and vector __uint128_t.
8951 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
8952 and mode attributes.
8959 (VS_scalar): Likewise.
8960 (VS_double): Likewise.
8961 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
8963 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
8964 we support the ISA 2.07 128-bit integer arithmetic instructions.
8965 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
8966 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
8967 and TImode types for use with the builtin functions.
8968 (V1TI_type_node): Likewise.
8969 (unsigned_V1TI_type_node): Likewise.
8970 (intTI_type_internal_node): Likewise.
8971 (uintTI_type_internal_node): Likewise.
8973 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
8974 128-bit builtin functions.
8975 (UNSPEC_VADDEUQM): Likewise.
8976 (UNSPEC_VADDECUQ): Likewise.
8977 (UNSPEC_VSUBCUQ): Likewise.
8978 (UNSPEC_VSUBEUQM): Likewise.
8979 (UNSPEC_VSUBECUQ): Likewise.
8980 (VM): Add V1TImode to vector mode iterators.
8982 (VI_unit): Likewise.
8983 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
8984 (altivec_vaddcuq): Likewise.
8985 (altivec_vsubuqm): Likewise.
8986 (altivec_vsubcuq): Likewise.
8987 (altivec_vaddeuqm): Likewise.
8988 (altivec_vaddecuq): Likewise.
8989 (altivec_vsubeuqm): Likewise.
8990 (altivec_vsubecuq): Likewise.
8992 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
8994 (BOOL_128): Likewise.
8995 (BOOL_REGS_OUTPUT): Likewise.
8996 (BOOL_REGS_OP1): Likewise.
8997 (BOOL_REGS_OP2): Likewise.
8998 (BOOL_REGS_UNARY): Likewise.
8999 (BOOL_REGS_AND_CR0): Likewise.
9001 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
9002 128-bit integer builtin support.
9003 (vec_vadduqm): Likewise.
9004 (vec_vaddecuq): Likewise.
9005 (vec_vaddeuqm): Likewise.
9006 (vec_vsubecuq): Likewise.
9007 (vec_vsubeuqm): Likewise.
9008 (vec_vsubcuq): Likewise.
9009 (vec_vsubuqm): Likewise.
9011 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
9012 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
9013 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
9014 128-bit integer add/subtract to ISA 2.07.
9016 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
9018 * config/arc/arc.c (arc_predicate_delay_insns):
9019 Fix third argument passed to conditionalize_nonjump.
9021 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
9023 * config/aarch64/aarch64-builtins.c
9024 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
9025 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
9026 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
9027 instead of __builtin_lfloor.
9028 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
9030 2014-03-12 Jakub Jelinek <jakub@redhat.com>
9032 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
9033 (tree_ssa_ifcombine_bb_1): New function.
9034 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
9035 is an empty forwarder block to then_bb or vice versa and then_bb
9036 and else_bb are effectively swapped.
9038 2014-03-12 Christian Bruel <christian.bruel@st.com>
9041 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
9042 REG_CFA_DEF_CFA note.
9043 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
9044 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
9046 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
9048 PR tree-optimization/60454
9049 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
9051 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9053 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
9054 Do not define target_cpu_default2 to generic.
9055 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
9056 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
9057 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
9059 2014-03-12 Jakub Jelinek <jakub@redhat.com>
9060 Marc Glisse <marc.glisse@inria.fr>
9062 PR tree-optimization/60502
9063 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
9064 instead of build_low_bits_mask.
9066 2014-03-12 Jakub Jelinek <jakub@redhat.com>
9069 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
9070 if there are multiple uses, but op doesn't live on E edge.
9071 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
9072 clobber stmts before __builtin_unreachable.
9074 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
9076 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
9077 hard_frame_pointer_rtx.
9078 * cse.c (cse_insn): Remove volatile check.
9079 * cselib.c (cselib_process_insn): Likewise.
9080 * dse.c (scan_insn): Likewise.
9082 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
9084 * config/arc/arc.c (conditionalize_nonjump): New function,
9086 (arc_ifcvt): ... this.
9087 (arc_predicate_delay_insns): Use it.
9089 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
9091 * config/arc/predicates.md (extend_operand): During/after reload,
9092 allow const_int_operand.
9093 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
9094 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
9095 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
9097 (umulsi3_highpart_i): Likewise.
9099 2014-03-11 Richard Biener <rguenther@suse.de>
9101 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
9102 Add asserts to guard possible wrong-code bugs.
9104 2014-03-11 Richard Biener <rguenther@suse.de>
9106 PR tree-optimization/60429
9107 PR tree-optimization/60485
9108 * tree-ssa-structalias.c (set_union_with_increment): Properly
9109 take into account all fields that overlap the shifted vars.
9110 (do_sd_constraint): Likewise.
9111 (do_ds_constraint): Likewise.
9112 (get_constraint_for_ptr_offset): Likewise.
9114 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
9116 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
9117 (nios2_compute_frame_layout):
9118 Add calculation of cfun->machine->fp_save_offset.
9119 (nios2_expand_prologue): Correct setting of frame pointer register
9121 (nios2_expand_epilogue): Update recovery of stack pointer from
9122 frame pointer accordingly.
9123 (nios2_initial_elimination_offset): Update calculation of offset
9124 for eliminating to HARD_FRAME_POINTER_REGNUM.
9126 2014-03-10 Jakub Jelinek <jakub@redhat.com>
9129 * ipa.c (symtab_remove_unreachable_nodes): Don't call
9130 cgraph_get_create_node on VAR_DECLs.
9132 2014-03-10 Richard Biener <rguenther@suse.de>
9135 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
9137 2014-03-08 Douglas B Rupp <rupp@gnat.com>
9139 * config/vms/vms.opt (vms_float_format): New variable.
9141 2014-03-08 Tobias Burnus <burnus@net-b.de>
9143 * doc/invoke.texi (-fcilkplus): Update implementation status.
9145 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
9146 Richard Biener <rguenther@suse.de>
9148 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
9149 consistently accross all TUs.
9150 (run_gcc): Enable -fshort-double automatically at link at link-time
9151 and disallow override.
9153 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
9156 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
9157 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
9158 if they can't be used.
9160 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9162 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
9163 for Solaris 11/x86 ld.
9164 * configure: Regenerate.
9166 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9168 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
9169 (LIB_TLS_SPEC): Save as ld_tls_libs.
9170 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
9171 (HAVE_AS_IX86_TLSLDM): New test.
9172 * configure, config.in: Regenerate.
9173 * config/i386/i386.c (legitimize_tls_address): Fall back to
9174 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
9175 cannot support TLS_MODEL_LOCAL_DYNAMIC.
9176 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
9177 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
9179 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
9181 * common.opt (fira-loop-pressure): Mark as optimization.
9183 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
9185 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
9186 an OpenMP mappable type.
9188 2014-03-06 Matthias Klose <doko@ubuntu.com>
9190 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
9191 MULTILIB_OSDIRNAMES is not defined.
9193 2014-03-06 Jakub Jelinek <jakub@redhat.com>
9194 Meador Inge <meadori@codesourcery.com>
9197 * config/arm/arm.c (arm_tls_symbol_p): Remove.
9198 (arm_legitimize_address): Call legitimize_tls_address for any
9199 arm_tls_referenced_p expression, handle constant addend. Call it
9200 before testing for !TARGET_ARM.
9201 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
9203 2014-03-06 Richard Biener <rguenther@suse.de>
9209 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
9211 * tree-streamer.c (record_common_node): Assert we don't record
9212 nodes with type double.
9213 (preload_common_node): Skip type double, complex double and double
9214 pointer since it is now frontend dependent due to fshort-double option.
9216 2014-03-06 Richard Biener <rguenther@suse.de>
9218 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
9219 or -fno-lto is specified and the linker has full plugin support.
9220 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
9221 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
9222 * lto-wrapper.c (merge_and_complain): Merge compile-time
9223 optimization levels.
9224 (run_gcc): And pass it through to the link options.
9226 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
9230 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9232 * cselib.c (remove_useless_values): Skip to avoid quadratic
9233 behavior if the condition moved from...
9234 (cselib_process_insn): ... here holds.
9236 2014-03-05 Jakub Jelinek <jakub@redhat.com>
9239 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
9240 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
9243 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
9244 (TM_H): Add x86-tune.def.
9246 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9248 * config/aarch64/aarch64.c (generic_tunings):
9249 Use cortexa57_extra_costs.
9251 2014-03-05 Jakub Jelinek <jakub@redhat.com>
9254 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
9255 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
9256 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
9259 2014-03-04 Heiher <r@hev.cc>
9261 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
9262 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
9264 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
9266 * config/i386/predicates.md (const2356_operand): Change to ...
9267 (const2367_operand): ... this.
9268 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
9270 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9271 (*avx512pf_scatterpf<mode>sf): Ditto.
9272 (avx512pf_scatterpf<mode>df): Ditto.
9273 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9274 (*avx512pf_scatterpf<mode>df): Ditto.
9275 * config/i386/i386.c (ix86_expand_builtin): Update
9276 incorrect hint operand error message.
9278 2014-03-04 Richard Biener <rguenther@suse.de>
9280 * lto-section-in.c (lto_get_section_data): Fix const cast.
9282 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
9284 * tree-streamer.c (record_common_node): Assert we don't record
9285 nodes with type double.
9286 (preload_common_node): Skip type double, complex double and double
9287 pointer since it is now frontend dependent due to fshort-double option.
9289 2014-03-04 Richard Biener <rguenther@suse.de>
9292 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
9293 (lto_input_toplevel_asms): Likewise.
9294 * lto-section-in.c (lto_get_section_data): Instead do it here
9297 2014-03-04 Richard Biener <rguenther@suse.de>
9299 PR tree-optimization/60382
9300 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
9301 dead PHIs a reduction.
9303 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
9305 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
9307 (_mm_prefetch): Move out of GCC target("sse") pragma.
9308 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
9309 GCC target("prfchw") pragma.
9310 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
9312 * config/i386/i386.c (ix86_option_override_internal): Enable
9313 -mprfchw with -mprefetchwt1.
9315 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
9317 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
9320 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
9322 * opts.h (CL_PCH_IGNORE): Define.
9323 * targhooks.c (option_affects_pch_p):
9324 Return false for options that have CL_PCH_IGNORE set.
9325 * opt-functions.awk: Process PchIgnore.
9326 * doc/options.texi: Document PchIgnore.
9328 * config/arc/arc.opt (misize): Add PchIgnore property.
9330 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9332 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
9333 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
9334 constraint on constants to permit them being loaded into
9335 GENERAL_REGS or BASE_REGS.
9337 2014-03-03 Nick Clifton <nickc@redhat.com>
9339 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
9340 anti-cacnonical alternatives.
9341 (negandhi3_real): New pattern.
9342 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
9344 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9346 * config/avr/avr-mcus.def: Remove atxmega16x1.
9347 * config/avr/avr-tables.opt: Regenerate.
9348 * config/avr/t-multilib: Regenerate.
9349 * doc/avr-mmcu.texi: Regenerate.
9351 2014-03-03 Tobias Grosser <tobias@grosser.es>
9352 Mircea Namolaru <mircea.namolaru@inria.fr>
9354 PR tree-optimization/58028
9355 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
9358 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9360 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
9361 not handled by recognizers.
9363 2014-03-03 Jakub Jelinek <jakub@redhat.com>
9366 * function.c (expand_function_end): Don't emit
9367 clobber_return_register sequence if clobber_after is a BARRIER.
9368 * cfgexpand.c (construct_exit_block): Append instructions before
9369 return_label to prev_bb.
9371 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9373 * config/rs6000/constraints.md: Document reserved use of "wc".
9375 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9378 * ipa.c (function_and_variable_visibility): When dissolving comdat
9379 group, also set all symbols to local.
9381 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9386 2013-12-14 Jan Hubicka <jh@suse.cz>
9388 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
9390 2014-03-02 Jon Beniston <jon@beniston.com>
9396 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
9397 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
9398 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
9399 (simple_return, *simple_return): New patterns
9400 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
9401 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
9403 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
9405 * dwarf2out.c (gen_subprogram_die): Tidy.
9407 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
9410 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
9411 (*mov_t_msb_neg_negc): ... this new insn.
9413 2014-02-28 Jason Merrill <jason@redhat.com>
9416 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
9419 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
9422 * dwarf2out.c (decltype_auto_die): New static.
9423 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
9424 (gen_type_die_with_usage): Handle 'decltype(auto)'.
9425 (is_cxx_auto): Likewise.
9427 2014-02-28 Ian Bolton <ian.bolton@arm.com>
9429 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
9430 we are not using general regs only.
9432 2014-02-28 Richard Biener <rguenther@suse.de>
9435 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
9436 previous fix and only allow to remove trivial pre-headers
9437 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
9438 (remove_forwarder_block): Properly update the latch of a loop.
9440 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9443 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
9444 (cselib_preserved_hash_table): New.
9445 (preserve_constants_and_equivs): Move preserved vals to it.
9446 (cselib_find_slot): Look it up first.
9447 (cselib_init): Initialize it.
9448 (cselib_finish): Release it.
9449 (dump_cselib_table): Dump it.
9451 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9454 * cselib.c (remove_useless_values): Skip to avoid quadratic
9455 behavior if the condition moved from...
9456 (cselib_process_insn): ... here holds.
9458 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9461 * var-tracking.c (vt_initialize): Apply the same condition to
9462 preserve the CFA base value.
9464 2014-02-28 Joey Ye <joey.ye@arm.com>
9467 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
9468 if reload in progress or completed.
9470 2014-02-28 Tobias Burnus <burnus@net-b.de>
9473 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
9476 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
9478 * doc/tm.texi.in (Condition Code Status): Update documention for
9479 relative locations of cc0-setter and cc0-user.
9481 2014-02-27 Jeff Law <law@redhat.com>
9483 PR rtl-optimization/52714
9484 * combine.c (try_combine): When splitting an unrecognized PARALLEL
9485 into two independent simple sets, if I3 is a jump, ensure the
9486 pattern we place into I3 is a (set (pc) ...).
9488 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
9489 Jeff Law <law@redhat.com>
9491 PR rtl-optimization/49847
9492 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
9493 are in different blocks.
9494 * doc/tm.texi (Condition Code Status): Update documention for
9495 relative locations of cc0-setter and cc0-user.
9497 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
9500 * lra.c (lra_emit_add): Check SUBREG too.
9502 2014-02-27 Andreas Schwab <schwab@suse.de>
9504 * config/m68k/m68k.c (m68k_option_override): Disable
9505 -flive-range-shrinkage for classic m68k.
9506 (m68k_override_options_after_change): Likewise.
9508 2014-02-27 Marek Polacek <polacek@redhat.com>
9511 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
9512 -Wmaybe-uninitialized.
9514 2014-02-27 Alan Modra <amodra@gmail.com>
9517 * reload1.c (emit_input_reload_insns): When reload_override_in,
9518 set old to rl->in_reg when rl->in_reg is a subreg.
9520 2014-02-26 Richard Biener <rguenther@suse.de>
9523 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
9525 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9527 * common/config/i386/predicates.md (const1256_operand): Remove.
9528 (const2356_operand): New.
9529 (const_1_to_2_operand): Remove.
9530 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9531 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9532 (*avx512pf_gatherpf<mode>sf): Ditto.
9533 (avx512pf_gatherpf<mode>df): Ditto.
9534 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9535 (*avx512pf_gatherpf<mode>df): Ditto.
9536 (avx512pf_scatterpf<mode>sf): Ditto.
9537 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9538 (*avx512pf_scatterpf<mode>sf): Ditto.
9539 (avx512pf_scatterpf<mode>df): Ditto.
9540 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9541 (*avx512pf_scatterpf<mode>df): Ditto.
9542 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
9544 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
9546 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
9547 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
9548 (_mm512_mask_testn_epi64_mask): Move to ...
9549 * config/i386/avx512cdintrin.h: Here.
9550 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
9551 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
9552 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
9553 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
9554 TARGET_AVX512F from TARGET_AVX512CD.
9556 2014-02-26 Richard Biener <rguenther@suse.de>
9559 * ipa.c (walk_polymorphic_call_targets): Properly guard
9560 call to inline_update_overall_summary.
9562 2014-02-26 Bin Cheng <bin.cheng@arm.com>
9565 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
9566 and latches only if requested. Fix latch if it is removed.
9567 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
9568 LOOPS_HAVE_PREHEADERS.
9570 2014-02-25 Andrew Pinski <apinski@cavium.com>
9572 * builtins.c (expand_builtin_thread_pointer): Create a new target
9573 when the target is NULL.
9575 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
9577 PR rtl-optimization/60317
9578 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9579 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9580 * lra-assigns.c: Include params.h.
9581 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
9582 other reload pseudos considerations.
9584 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9586 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
9587 to use canonical form for nor<mode>3.
9589 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9592 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
9595 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9597 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
9598 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
9599 (ix86_handle_option): Handle OPT_mprefetchwt1.
9600 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
9601 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9603 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9604 OPTION_MASK_ISA_PREFETCHWT1.
9605 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
9606 (PTA_PREFETCHWT1): New.
9607 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
9608 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
9609 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
9610 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
9611 (*prefetch_avx512pf_<mode>_: Change into ...
9612 (*prefetch_prefetchwt1_<mode>: This.
9613 * config/i386/i386.opt (mprefetchwt1): New.
9614 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
9615 (_mm_prefetch): Handle intent to write.
9616 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
9618 2014-02-25 Richard Biener <rguenther@suse.de>
9621 * emit-rtl.c (mem_attrs_htab): Remove.
9622 (mem_attrs_htab_hash): Likewise.
9623 (mem_attrs_htab_eq): Likewise.
9624 (set_mem_attrs): Always allocate new mem-attrs when something changed.
9625 (init_emit_once): Do not allocate mem_attrs_htab.
9627 2014-02-25 Richard Biener <rguenther@suse.de>
9630 * lto-opts.c (lto_write_options): Output non-explicit conservative
9631 -fwrapv, -fno-trapv and -fno-strict-overflow.
9632 * lto-wrapper.c (merge_and_complain): Handle merging those options.
9633 (run_gcc): And pass them through.
9635 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9637 * sel-sched.c (calculate_new_fences): New parameter ptime.
9638 Calculate it as a maximum over all fence cycles.
9639 (sel_sched_region_2): Adjust the call to calculate_new_fences.
9640 Print the final schedule timing when sched_verbose.
9642 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9644 PR rtl-optimization/60292
9645 * sel-sched.c (fill_vec_av_set): Do not reset target availability
9646 bit fot the fence instruction.
9648 2014-02-24 Alangi Derick <alangiderick@gmail.com>
9650 * calls.h: Fix typo in comment.
9652 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
9654 * config/pa/pa.c (pa_output_move_double): Don't valididate when
9655 adjusting offsetable addresses.
9657 2014-02-24 Guozhi Wei <carrot@google.com>
9659 * sparseset.h (sparseset_pop): Fix the wrong index.
9661 2014-02-24 Walter Lee <walt@tilera.com>
9663 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
9664 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
9666 * common/config/tilegx/tilegx-common.c
9667 (TARGET_DEFAULT_TARGET_FLAGS): Define.
9668 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
9670 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
9671 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
9672 (tilegx_gimplify_va_arg_expr): Handle big endian.
9673 (tilegx_expand_unaligned_load): Ditto.
9674 (tilegx_expand_unaligned_store): Ditto.
9675 (TARGET_RETURN_IN_MSB): New.
9676 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
9677 (TARGET_ENDIAN_DEFAULT): New.
9678 (TARGET_BIG_ENDIAN): Handle big endian.
9679 (BYTES_BIG_ENDIAN): Ditto.
9680 (WORDS_BIG_ENDIAN): Ditto.
9681 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
9684 * config/tilegx/tilegx.md (extv): Handle big endian.
9686 (insn_st<n>): Ditto.
9687 (insn_st<n>_add<bitsuffix>): Ditto.
9688 (insn_stnt<n>): Ditto.
9689 (insn_stnt<n>_add<bitsuffix>):Ditto.
9690 (vec_interleave_highv8qi): Handle big endian.
9691 (vec_interleave_highv8qi_be): New.
9692 (vec_interleave_highv8qi_le): New.
9693 (insn_v1int_h): Handle big endian.
9694 (vec_interleave_lowv8qi): Handle big endian.
9695 (vec_interleave_lowv8qi_be): New.
9696 (vec_interleave_lowv8qi_le): New.
9697 (insn_v1int_l): Handle big endian.
9698 (vec_interleave_highv4hi): Handle big endian.
9699 (vec_interleave_highv4hi_be): New.
9700 (vec_interleave_highv4hi_le): New.
9701 (insn_v2int_h): Handle big endian.
9702 (vec_interleave_lowv4hi): Handle big endian.
9703 (vec_interleave_lowv4hi_be): New.
9704 (vec_interleave_lowv4hi_le): New.
9705 (insn_v2int_l): Handle big endian.
9706 (vec_interleave_highv2si): Handle big endian.
9707 (vec_interleave_highv2si_be): New.
9708 (vec_interleave_highv2si_le): New.
9709 (insn_v4int_h): Handle big endian.
9710 (vec_interleave_lowv2si): Handle big endian.
9711 (vec_interleave_lowv2si_be): New.
9712 (vec_interleave_lowv2si_le): New.
9713 (insn_v4int_l): Handle big endian.
9714 * config/tilegx/tilegx.opt (mbig-endian): New option.
9715 (mlittle-endian): New option.
9716 * doc/install.texi: Document tilegxbe-linux.
9717 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
9719 2014-02-24 Martin Jambor <mjambor@suse.cz>
9722 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
9723 there are no parameter descriptors.
9725 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
9727 PR rtl-optimization/60268
9728 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
9729 initialization to ...
9730 (sched_rgn_init): ... here.
9731 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
9733 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9735 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
9738 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9740 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
9743 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9745 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
9746 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
9748 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9750 * config/microblaze/predicates.md: Add cmp_op predicate.
9751 * config/microblaze/microblaze.md: Add branch_compare instruction
9752 which uses cmp_op predicate and emits cmp insn before branch.
9753 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
9754 to microblaze_expand_conditional_branch and consolidate logic.
9755 (microblaze_expand_conditional_branch): emit branch_compare
9756 insn instead of handling cmp op separate from branch insn.
9758 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9760 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
9763 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9765 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
9766 define_insn with define_expand and new define_insn
9767 *altivec_lve<VI_char>x_internal.
9768 (altivec_stve<VI_char>x): Replace define_insn with define_expand
9769 and new define_insn *altivec_stve<VI_char>x_internal.
9770 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
9772 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
9774 (altivec_expand_stvex_be): New function.
9776 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
9778 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
9779 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
9780 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
9781 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
9783 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
9786 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
9787 instead of emit_move_insn.
9789 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9791 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
9793 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
9794 gen_altivec_vsumsws.
9796 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9798 * config/rs6000/altivec.md (altivec_lvxl): Rename as
9799 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
9800 (altivec_lvxl_<mode>): New define_expand incorporating
9801 -maltivec=be semantics where needed.
9802 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
9803 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
9804 semantics where needed.
9805 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
9806 (altivec_stvx_<mode>): New define_expand incorporating
9807 -maltivec=be semantics where needed.
9808 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
9809 VM2 iterator instead of V4SI.
9810 (altivec_stvxl_<mode>): New define_expand incorporating
9811 -maltivec=be semantics where needed.
9812 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
9813 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
9814 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
9815 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
9816 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
9817 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
9818 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
9819 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
9820 ALTIVEC_BUILTIN_STVXL.
9821 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
9822 (altivec_expand_stvx_be): Likewise.
9823 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
9824 (altivec_expand_lvx_be): Likewise.
9825 (altivec_expand_stvx_be): Likewise.
9826 (altivec_expand_builtin): Add cases for
9827 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
9828 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
9829 (altivec_init_builtins): Add definitions for
9830 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
9831 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
9833 2014-02-21 Catherine Moore <clm@codesourcery.com>
9835 * doc/invoke.texi (mvirt, mno-virt): Document.
9836 * config/mips/mips.opt (mvirt): New option.
9837 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
9839 2014-02-21 Richard Biener <rguenther@suse.de>
9841 PR tree-optimization/60276
9842 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
9843 (STMT_VINFO_MIN_NEG_DIST): New macro.
9844 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
9845 STMT_VINFO_MIN_NEG_DIST.
9846 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
9847 made for negative dependence distances still hold.
9849 2014-02-21 Richard Biener <rguenther@suse.de>
9852 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
9853 DECL_INITIAL for globals not in the current function context.
9855 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9857 PR tree-optimization/56490
9858 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
9859 * tree-ssa-uninit.c: Include params.h.
9860 (compute_control_dep_chain): Add num_calls argument, return false
9861 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
9862 num_calls to recursive call.
9863 (find_predicates): Change dep_chain into normal array,
9864 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
9865 variable and adjust compute_control_dep_chain caller.
9866 (find_def_preds): Likewise.
9868 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
9870 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
9871 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
9873 2014-02-21 Nick Clifton <nickc@redhat.com>
9875 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
9876 (pushhi1): Likewise.
9877 (popqi1): Add mode to pre_dec.
9880 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9882 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
9883 mode for mask of V8SFmode permutation.
9885 2014-02-20 Richard Henderson <rth@redhat.com>
9888 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
9889 a new pseudo for OLDVAL.
9891 2014-02-20 Jakub Jelinek <jakub@redhat.com>
9894 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
9895 gen_reg_rtx if d->testing_p.
9896 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
9897 if d->testing_p and we will certainly return true.
9898 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
9901 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
9903 * emit-rtl.c (gen_reg_rtx): Assert that
9904 crtl->emit.regno_pointer_align_length is non-zero.
9906 2014-02-20 Richard Henderson <rth@redhat.com>
9909 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
9910 on failure the store back into EXPECT.
9912 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
9913 Sandra Loosemore <sandra@codesourcery.com>
9915 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
9916 * config/nios2/nios2.c (nios2_function_profiler): Add
9917 -fPIC (flag_pic == 2) support.
9918 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
9919 (nios2_large_offset_p): New function.
9920 (nios2_unspec_reloc_p): Move up position, update to use
9921 nios2_large_offset_p.
9922 (nios2_unspec_address): Remove function.
9923 (nios2_unspec_offset): New function.
9924 (nios2_large_got_address): New function.
9925 (nios2_got_address): Add large offset support.
9926 (nios2_legitimize_tls_address): Update usage of removed and new
9928 (nios2_symbol_binds_local_p): New function.
9929 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
9930 (nios2_legitimize_address): Update to use nios2_large_offset_p.
9931 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
9932 (nios2_print_operand): Merge H/L processing, add hiadj/lo
9933 processing for (const (unspec ...)).
9934 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
9936 2014-02-20 Richard Biener <rguenther@suse.de>
9938 * tree-cfg.c (replace_uses_by): Mark altered BBs before
9939 doing the substitution.
9940 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
9942 2014-02-20 Martin Jambor <mjambor@suse.cz>
9945 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
9946 info when checking whether lattices are bottom.
9948 2014-02-20 Richard Biener <rguenther@suse.de>
9951 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
9954 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
9957 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
9958 parameter specifying the scaling.
9959 (inline_call): Update.
9960 (want_inline_recursively): Guard division by zero.
9961 (recursive_inlining): Update.
9962 * ipa-inline.h (clone_inlined_nodes): Update.
9964 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9967 * config/i386/i386.c (classify_argument): Pass structures of size
9968 64 bytes or less in register.
9970 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9971 Kirill Yukhin <kirill.yukhin@intel.com>
9973 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
9974 (_mm_rcp28_round_ss): Ditto.
9975 (_mm_rsqrt28_round_sd): Ditto.
9976 (_mm_rsqrt28_round_ss): Ditto.
9977 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
9978 (_mm_rcp14_round_ss): Ditto.
9979 (_mm_rsqrt14_round_sd): Ditto.
9980 (_mm_rsqrt14_round_ss): Ditto.
9981 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
9982 the first input operand, get rid of match_dup.
9983 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
9985 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9987 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
9988 operand as the first input operand, set type attribute.
9989 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9991 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
9992 operand as the first input operand, set type attribute.
9994 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9996 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
9999 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
10002 * config/i386/i386.c (construct_container): Remove TFmode check
10003 for X86_64_INTEGER_CLASS.
10005 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
10008 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
10009 only when -Wpsabi is enabled.
10011 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
10014 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
10015 passing arrays in registers are the same as for structs, so remove the
10016 special case for them.
10018 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
10020 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
10021 destination type, extract only the valid bits if the source type is not
10022 integral and has a different mode.
10024 2014-02-19 Richard Biener <rguenther@suse.de>
10027 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
10030 2014-02-19 Richard Biener <rguenther@suse.de>
10033 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
10034 (ipa_modify_call_arguments): Emit an argument load explicitely and
10035 preserve virtual SSA form there and for the replacement call.
10036 Do not update SSA form nor free dominance info.
10038 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
10040 * ipa.c (function_and_variable_visibility): Also clear WEAK
10041 flag when disolving COMDAT_GROUP.
10043 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
10045 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
10046 * ipa-prop.c (ipa_set_jf_known_type): Return early when
10047 not devirtualizing.
10048 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
10049 do more sanity checks.
10050 (detect_type_change): Return true when giving up early.
10051 (compute_complex_assign_jump_func): Fix type parameter of
10052 ipa_set_ancestor_jf.
10053 (compute_complex_ancestor_jump_func): Likewise.
10054 (update_jump_functions_after_inlining): Fix updating of
10056 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
10058 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
10060 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
10061 inline clones when edge disappears.
10063 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
10066 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
10067 Split 64-bit moves into 2 patterns. Do not allow the use of
10068 direct move for TDmode in little endian, since the decimal value
10069 has little endian bytes within a word, but the 64-bit pieces are
10070 ordered in a big endian fashion, and normal subreg's of TDmode are
10072 (mov<mode>_64bit_dm): Likewise.
10073 (movtd_64bit_nodm): Likewise.
10075 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
10077 PR tree-optimization/60174
10078 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
10079 statement of an SSA_NAME that occurs in an abnormal PHI node.
10081 2014-02-18 Jakub Jelinek <jakub@redhat.com>
10084 * final.c (SEEN_BB): Remove.
10085 (SEEN_NOTE, SEEN_EMITTED): Renumber.
10086 (final_scan_insn): Don't force_source_line on second
10087 NOTE_INSN_BASIC_BLOCK.
10089 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
10092 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
10093 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
10094 (type_natural_mode): Warn ABI change when %zmm register is not
10095 available for AVX512F vector value passing.
10097 2014-02-18 Kai Tietz <ktietz@redhat.com>
10100 * config/i386/i386.c (ix86_expand_prologue): Use value in
10101 rax register as displacement when restoring %r10 or %rax.
10102 Fix wrong offset when restoring both registers.
10104 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
10106 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
10107 assertion with conditional return.
10109 2014-02-18 Jakub Jelinek <jakub@redhat.com>
10110 Uros Bizjak <ubizjak@gmail.com>
10113 * config/i386/driver-i386.c (host_detect_local_cpu): If
10114 YMM state is not saved by the OS, also clear has_f16c. Move
10115 CPUID 0x80000001 handling before YMM state saving checking.
10117 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
10119 PR rtl-optimization/58960
10120 * haifa-sched.c (alloc_global_sched_pressure_data): New,
10121 factored out from ...
10122 (sched_init): ... here.
10123 (free_global_sched_pressure_data): New, factored out from ...
10124 (sched_finish): ... here.
10125 * sched-int.h (free_global_sched_pressure_data): Declare.
10126 * sched-rgn.c (nr_regions_initial): New static global.
10127 (haifa_find_rgns): Initialize it.
10128 (schedule_region): Disable sched-pressure for the newly
10131 2014-02-17 Richard Biener <rguenther@suse.de>
10133 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
10134 release SSA defs of pattern stmts.
10136 2014-02-17 Richard Biener <rguenther@suse.de>
10138 * tree-inline.c (expand_call_inline): Release the virtual
10139 operand defined by the call we are about to inline.
10141 2014-02-17 Richard Biener <rguenther@suse.de>
10143 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
10145 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
10146 Ilya Tocar <ilya.tocar@intel.com>
10148 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
10149 arguments order in builtin.
10150 (_mm512_permutexvar_epi64): Ditto.
10151 (_mm512_mask_permutexvar_epi64): Ditto
10152 (_mm512_maskz_permutexvar_epi32): Ditto
10153 (_mm512_permutexvar_epi32): Ditto
10154 (_mm512_mask_permutexvar_epi32): Ditto
10156 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10158 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
10159 (p8_vmrgow): Likewise.
10161 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10163 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
10166 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
10169 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
10170 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
10171 into 64-bit and 32-bit moves. On 64-bit moves, add support for
10172 using direct move instructions on ISA 2.07. Also adjust
10173 instruction length for 64-bit.
10174 (mov<mode>_64bit, TFmode/TDmode): Likewise.
10175 (mov<mode>_32bit, TFmode/TDmode): Likewise.
10177 2014-02-15 Alan Modra <amodra@gmail.com>
10181 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
10182 find_replacement on parts of insn rtl that might be reloaded.
10184 2014-02-15 Richard Biener <rguenther@suse.de>
10186 PR tree-optimization/60183
10187 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
10188 (tree_ssa_phiprop): Calculate and free post-dominators.
10190 2014-02-14 Jeff Law <law@redhat.com>
10192 PR rtl-optimization/60131
10193 * ree.c (get_extended_src_reg): New function.
10194 (combine_reaching_defs): Use it rather than assuming location of REG.
10195 (find_and_remove_re): Verify first operand of extension is
10196 a REG before adding the insns to the copy list.
10198 2014-02-14 Roland McGrath <mcgrathr@google.com>
10200 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
10201 * configure: Regenerated.
10202 * config.in: Regenerated.
10203 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
10204 instead of ASM_SHORT.
10206 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
10207 Richard Earnshaw <rearnsha@arm.com>
10209 PR rtl-optimization/59535
10210 * lra-constraints.c (process_alt_operands): Encourage alternative
10211 when unassigned pseudo class is superset of the alternative class.
10212 (inherit_reload_reg): Don't inherit when optimizing for code size.
10213 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
10214 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
10215 modes not less than 4 for Thumb1.
10217 2014-02-14 Kyle McMartin <kyle@redhat.com>
10220 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
10222 2014-02-14 Richard Biener <rguenther@suse.de>
10224 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
10225 (get_frame_arg): Drop the assert with langhook types_compatible_p.
10226 Do not strip INDIRECT_REFs.
10228 2014-02-14 Richard Biener <rguenther@suse.de>
10231 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
10232 DECL_FUNCTION_SPECIFIC_TARGET.
10233 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
10234 * tree-streamer-out.c (pack_ts_target_option): Remove.
10235 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
10236 (write_ts_function_decl_tree_pointers): Do not stream
10237 DECL_FUNCTION_SPECIFIC_TARGET.
10238 * tree-streamer-in.c (unpack_ts_target_option): Remove.
10239 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
10240 (lto_input_ts_function_decl_tree_pointers): Do not stream
10241 DECL_FUNCTION_SPECIFIC_TARGET.
10243 2014-02-14 Jakub Jelinek <jakub@redhat.com>
10245 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
10246 (get_initial_def_for_induction, vectorizable_induction): Ignore
10247 debug stmts when looking for exit_phi.
10248 (vectorizable_live_operation): Fix up condition.
10250 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10252 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
10253 nreverse() because it changes the content of original tree list.
10255 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10257 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
10258 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
10260 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10262 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
10263 GNU coding standards.
10265 2014-02-13 Jakub Jelinek <jakub@redhat.com>
10268 * dwarf2out.c (gen_subprogram_die): Don't call
10269 add_calling_convention_attribute if subr_die is old_die.
10271 2014-02-13 Sharad Singhai <singhai@google.com>
10273 * doc/optinfo.texi: Fix order of nodes.
10275 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
10277 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
10278 operands[2], not operands[3].
10280 2014-02-13 Richard Biener <rguenther@suse.de>
10283 * doc/install.texi (ISL): Update recommended version to 0.12.2,
10284 mention the possibility of an in-tree build.
10285 (CLooG): Update recommended version to 0.18.1, mention the
10286 possibility of an in-tree build and clarify that the ISL
10287 bundled with CLooG does not work.
10289 2014-02-13 Jakub Jelinek <jakub@redhat.com>
10292 * expr.c (compress_float_constant): If x is a hard register,
10293 extend into a pseudo and then move to x.
10295 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
10297 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
10298 caused by bad second argument to warning_at() with -mhotpatch and
10299 nested functions (e.g. with gfortran).
10301 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
10303 * opts.c (option_name): Remove "enabled by default" rider.
10305 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
10307 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
10309 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
10310 Uros Bizjak <ubizjak@gmail.com>
10313 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
10314 * configure: Regenerated.
10316 2014-02-12 Richard Biener <rguenther@suse.de>
10318 * vec.c (vec_prefix::calculate_allocation): Move as
10319 inline variant to vec.h.
10320 (vec_prefix::calculate_allocation_1): New out-of-line version.
10321 * vec.h (vec_prefix::calculate_allocation_1): Declare.
10322 (vec_prefix::m_has_auto_buf): Rename to ...
10323 (vec_prefix::m_using_auto_storage): ... this.
10324 (vec_prefix::calculate_allocation): Inline the easy cases
10325 and dispatch to calculate_allocation_1 which doesn't need the
10327 (va_heap::reserve): Use gcc_checking_assert.
10328 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
10329 m_using_auto_storage.
10330 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
10332 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
10333 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
10334 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
10336 2014-02-12 Richard Biener <rguenther@suse.de>
10338 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
10339 when we found a dependence.
10341 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
10343 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
10345 (maybe_fold_stmt): ... into this new function.
10346 * omp-low.c (lower_omp): Update comment.
10348 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
10351 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
10354 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
10356 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
10357 identifiers in comments.
10358 (cortexa53_extra_costs): Likewise.
10359 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
10360 (cortexa7_extra_costs): Likewise.
10361 (cortexa12_extra_costs): Likewise.
10362 (cortexa15_extra_costs): Likewise.
10363 (v7m_extra_costs): Likewise.
10365 2014-02-12 Richard Biener <rguenther@suse.de>
10367 PR middle-end/60092
10368 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
10369 of posix_memalign being successful.
10370 (lower_stmt): Restrict lowering of posix_memalign to when
10371 -ftree-bit-ccp is enabled.
10373 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10375 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
10377 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
10379 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
10381 PR rtl-optimization/60116
10382 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
10383 other_insn once the combination has been validated.
10385 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
10388 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
10390 * ipa-devirt.c: Include demangle.h
10391 (odr_violation_reported): New static variable.
10392 (add_type_duplicate): Update odr_violations.
10393 (maybe_record_node): Add completep parameter; update it.
10394 (record_target_from_binfo): Add COMPLETEP parameter;
10395 update it as needed.
10396 (possible_polymorphic_call_targets_1): Likewise.
10397 (struct polymorphic_call_target_d): Add nonconstruction_targets;
10398 rename FINAL to COMPLETE.
10399 (record_targets_from_bases): Sanity check we found the binfo;
10400 fix COMPLETEP updating.
10401 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
10402 parameter, fix computing of COMPLETEP.
10403 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
10404 at LTO time do demangling.
10405 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
10406 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
10408 (gimple_get_virt_method_for_binfo): Likewise.
10409 * gimple-fold.h (gimple_get_virt_method_for_binfo,
10410 gimple_get_virt_method_for_vtable): Update prototypes.
10412 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
10415 * genautomata.c (add_presence_absence): Fix typo with
10416 {final_}presence_list.
10418 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10421 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
10422 for VSX/Altivec vectors that land in GPR registers.
10424 2014-02-11 Richard Henderson <rth@redhat.com>
10425 Jakub Jelinek <jakub@redhat.com>
10428 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
10429 around drhs if type conversion to lacc->type is not useless.
10431 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10433 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
10435 (cortex-a57.cortex-a53): Likewise.
10436 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
10438 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10440 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
10443 2014-02-11 Renlin Li <Renlin.Li@arm.com>
10445 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
10446 add_options_for_arm_vfp3.
10448 2014-02-11 Jeff Law <law@redhat.com>
10450 PR middle-end/54041
10451 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
10452 object with an undesirable mode.
10454 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10457 * config/i386/sol2-9.h: New file.
10458 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
10459 *-*-solaris2.9*): Use it.
10461 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10463 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
10464 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
10466 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10468 * config/microblaze/microblaze.c: Extend mcpu version format
10470 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
10472 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
10474 2014-02-10 Richard Henderson <rth@redhat.com>
10477 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
10478 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
10479 ms-abi vs -mno-accumulate-outgoing-args.
10480 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
10481 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
10484 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
10486 PR middle-end/60080
10487 * cfgexpand.c (expand_asm_operands): Attach source location to
10488 ASM_INPUT rtx objects.
10489 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
10491 2014-02-10 Nick Clifton <nickc@redhat.com>
10493 * config/mn10300/mn10300.c (popcount): New function.
10494 (mn10300_expand_prologue): Include saved registers in stack usage
10497 2014-02-10 Jeff Law <law@redhat.com>
10499 PR middle-end/52306
10500 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
10501 when changing the SET_DEST of a prior insn to avoid an input reload.
10503 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10505 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
10506 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
10507 -mcall-openbsd, or -mcall-linux.
10508 (CC1_ENDIAN_BIG_SPEC): Remove.
10509 (CC1_ENDIAN_LITTLE_SPEC): Remove.
10510 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10511 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
10512 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
10513 and %cc1_endian_default.
10514 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10516 2014-02-10 Richard Biener <rguenther@suse.de>
10518 PR tree-optimization/60115
10519 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
10520 MEM_REF handling. Properly verify that the accesses are not
10521 out of the objects bound.
10523 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10525 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
10528 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
10530 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
10531 proper constants and fix formatting.
10532 (possible_polymorphic_call_targets): Fix formatting.
10534 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
10535 Ilya Tocar <ilya.tocar@intel.com>
10537 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
10538 (_mm512_loadu_epi32): Renamed into...
10539 (_mm512_loadu_si512): This.
10540 (_mm512_storeu_epi32): Renamed into...
10541 (_mm512_storeu_si512): This.
10542 (_mm512_maskz_ceil_ps): Removed.
10543 (_mm512_maskz_ceil_pd): Ditto.
10544 (_mm512_maskz_floor_ps): Ditto.
10545 (_mm512_maskz_floor_pd): Ditto.
10546 (_mm512_floor_round_ps): Ditto.
10547 (_mm512_floor_round_pd): Ditto.
10548 (_mm512_ceil_round_ps): Ditto.
10549 (_mm512_ceil_round_pd): Ditto.
10550 (_mm512_mask_floor_round_ps): Ditto.
10551 (_mm512_mask_floor_round_pd): Ditto.
10552 (_mm512_mask_ceil_round_ps): Ditto.
10553 (_mm512_mask_ceil_round_pd): Ditto.
10554 (_mm512_maskz_floor_round_ps): Ditto.
10555 (_mm512_maskz_floor_round_pd): Ditto.
10556 (_mm512_maskz_ceil_round_ps): Ditto.
10557 (_mm512_maskz_ceil_round_pd): Ditto.
10558 (_mm512_expand_pd): Ditto.
10559 (_mm512_expand_ps): Ditto.
10560 * config/i386/i386.c (ix86_builtins): Remove
10561 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
10562 (bdesc_args): Ditto.
10563 * config/i386/predicates.md (const1256_operand): New.
10564 (const_1_to_2_operand): Ditto.
10565 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
10566 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10567 (*avx512pf_gatherpf<mode>sf): Ditto.
10568 (avx512pf_gatherpf<mode>df): Ditto.
10569 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10570 (*avx512pf_gatherpf<mode>df): Ditto.
10571 (avx512pf_scatterpf<mode>sf): Ditto.
10572 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10573 (*avx512pf_scatterpf<mode>sf): Ditto.
10574 (avx512pf_scatterpf<mode>df): Ditto.
10575 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10576 (*avx512pf_scatterpf<mode>df): Ditto.
10577 (avx512f_expand<mode>): Removed.
10578 (<shift_insn><mode>3<mask_name>): Change predicate type.
10580 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10582 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
10583 not at the end of datarefs vector use ordered_remove to avoid
10584 reordering datarefs vector.
10587 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
10588 mark local addressable non-static vars as GOVD_PRIVATE
10589 instead of GOVD_LOCAL.
10590 * omp-low.c (lower_omp_for): Move gimple_bind_vars
10591 and BLOCK_VARS of gimple_bind_block to new_stmt rather
10594 PR middle-end/60092
10595 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
10596 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
10597 assume_aligned or alloc_align attributes.
10598 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
10599 arguments. Handle also assume_aligned and alloc_align attributes.
10600 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
10601 calls to functions with assume_aligned or alloc_align attributes.
10602 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
10604 2014-02-08 Terry Guo <terry.guo@arm.com>
10606 * doc/invoke.texi: Document ARM -march=armv7e-m.
10608 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10610 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
10611 flag on __cilkrts_rethrow builtin.
10614 * ipa-cp.c (determine_versionability): Fail at -O0
10615 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
10616 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
10619 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10622 * tree-inline.c (copy_forbidden): Fail for
10623 __attribute__((optimize (0))) functions.
10625 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10627 * varpool.c: Include pointer-set.h.
10628 (varpool_remove_unreferenced_decls): Variables in other partitions
10629 will not be output; be however careful to not lose information
10630 about partitioning.
10632 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10634 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
10635 lookup in the vtable constructor.
10637 2014-02-07 Jeff Law <law@redhat.com>
10640 * config/m68k/m68k.md (ashldi_extsi): Turn into a
10641 define_insn_and_split.
10643 * ipa-inline.c (inline_small_functions): Fix typos.
10645 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10647 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
10648 (s390_can_use_return_insn): Declare.
10649 * config/s390/s390.h (EPILOGUE_USES): Define.
10650 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
10652 (s390_chunkify_start): Handle return JUMP_LABELs.
10653 (s390_early_mach): Emit a main_pool instruction on the entry edge.
10654 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
10655 (s390_can_use_return_insn): New functions.
10656 (s390_fix_long_loop_prediction): Handle conditional returns.
10657 (TARGET_SET_UP_BY_PROLOGUE): Define.
10658 * config/s390/s390.md (ANY_RETURN): New code iterator.
10659 (*creturn, *csimple_return, return, simple_return): New patterns.
10661 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10663 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
10664 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
10665 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
10666 REG_CFA_RESTORE list when deciding not to restore a register.
10668 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10670 * config/s390/s390.c: Include tree-pass.h and context.h.
10671 (s390_early_mach): New function, split out from...
10672 (s390_emit_prologue): ...here.
10673 (pass_data_s390_early_mach): New pass structure.
10674 (pass_s390_early_mach): New class.
10675 (s390_option_override): Create and register early_mach pass.
10676 Move to end of file.
10678 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10680 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
10681 to match for the exit block.
10683 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10685 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
10686 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
10687 Reject misaligned operands.
10689 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10691 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
10693 2014-02-07 Richard Biener <rguenther@suse.de>
10695 PR middle-end/60092
10696 * gimple-low.c (lower_builtin_posix_memalign): New function.
10697 (lower_stmt): Call it to lower posix_memalign in a way
10698 to make alignment info accessible.
10700 2014-02-07 Jakub Jelinek <jakub@redhat.com>
10703 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
10704 __builtin_setjmp_receiver.
10706 2014-02-07 Richard Biener <rguenther@suse.de>
10708 PR middle-end/60092
10709 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
10710 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
10711 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
10712 Handle BUILT_IN_POSIX_MEMALIGN.
10713 (find_func_clobbers): Likewise.
10714 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
10715 (call_may_clobber_ref_p_1): Likewise.
10717 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10720 * ipa-devirt.c (record_target_from_binfo): Remove overactive
10723 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10726 * lto-cgraph.c (lto_output_node): Use
10727 symtab_get_symbol_partitioning_class.
10728 (lto_output_varpool_node): likewise.
10729 (symtab_get_symbol_partitioning_class): Move here from
10730 lto/lto-partition.c
10731 * cgraph.h (symbol_partitioning_class): Likewise.
10732 (symtab_get_symbol_partitioning_class): Declare.
10734 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10736 * ggc.h (ggc_internal_cleared_alloc): New macro.
10737 * vec.h (vec_safe_copy): Handle memory stats.
10738 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
10739 * target-globals.c (save_target_globals): Likewise.
10741 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10744 * expr.c (emit_move_resolve_push): Export; be bit more selective
10745 on when to clear alias set.
10746 * expr.h (emit_move_resolve_push): Declare.
10747 * function.h (struct function): Add tail_call_marked.
10748 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
10749 * config/i386/i386-protos.h (ix86_expand_push): Remove.
10750 * config/i386/i386.md (TImode move expander): De not call
10752 (FP push expanders): Preserve memory attributes.
10753 * config/i386/sse.md (push<mode>1): Remove.
10754 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
10755 (ix86_expand_push): Remove.
10756 * config/i386/mmx.md (push<mode>1): Remove.
10758 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10760 PR rtl-optimization/60030
10761 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
10762 lopart with paradoxical subreg before shifting it up by hprec.
10764 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10766 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
10767 Remove extra newline at end of file.
10768 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
10769 (arm_issue_rate): Handle cortexa57.
10770 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
10771 (cortex-a57.cortex-a53): Likewise.
10773 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10776 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
10777 don't record in REG_FRAME_RELATED_EXPR registers not set in that
10779 (arm_expand_prologue): Adjust all callers.
10780 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
10781 info, registers also at the lowest numbered registers side. Use
10782 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
10786 * var-tracking.c (adjust_mems): Before adding a SET to
10787 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
10789 2014-02-06 Alan Modra <amodra@gmail.com>
10792 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
10793 change SDmode to DDmode when lra_in_progress.
10795 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10797 PR middle-end/59150
10798 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
10799 free_data_ref on the dr first, and before goto again also set dr
10800 to the next dr. For simd_lane_access, free old datarefs[i] before
10801 overwriting it. For get_vectype_for_scalar_type failure, don't
10802 free_data_ref if simd_lane_access.
10804 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
10807 * tree.h (opts_for_fn): New inline function.
10808 (opt_for_fn): Define.
10809 * config/i386/i386.c (ix86_function_regparm): Use
10810 opt_for_fn (decl, optimize) instead of optimize.
10812 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
10814 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
10815 for SYMBOL_REF in large memory model.
10817 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10819 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
10820 and crypto support.
10821 (cortex-a57): Likewise.
10822 (cortex-a57.cortex-a53): Likewise.
10824 2014-02-06 Yury Gribov <y.gribov@samsung.com>
10825 Kugan Vivekanandarajah <kuganv@linaro.org>
10827 * config/arm/arm.c (arm_vector_alignment_reachable): Check
10829 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
10831 2014-02-06 Richard Biener <rguenther@suse.de>
10833 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
10834 set_loop_copy and initialize_original_copy_tables.
10836 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
10838 * config/aarch64/aarch64-simd.md
10839 (aarch64_ashr_simddi): Change QI to SI.
10841 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10842 Jakub Jelinek <jakub@redhat.com>
10844 PR middle-end/60013
10845 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
10848 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10850 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
10851 CODE_FOR_altivec_vpku[hw]um to
10852 CODE_FOR_altivec_vpku[hw]um_direct.
10853 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
10854 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
10855 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
10856 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
10858 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10860 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
10861 generation for -maltivec=be.
10862 (altivec_vsumsws): Simplify redundant test.
10864 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10866 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
10867 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
10868 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
10869 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
10870 gen_altivec_vpkuwum.
10871 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
10873 (altivec_vpks<VI_char>ss): Likewise.
10874 (altivec_vpks<VI_char>us): Likewise.
10875 (altivec_vpku<VI_char>us): Likewise.
10876 (altivec_vpku<VI_char>um): Likewise.
10877 (altivec_vpku<VI_char>um_direct): New (copy of
10878 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
10880 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
10881 target is little endian and -maltivec=be is not specified.
10882 (*altivec_vupkhs<VU_char>_direct): New (copy of
10883 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
10884 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
10885 target is little endian and -maltivec=be is not specified.
10886 (*altivec_vupkls<VU_char>_direct): New (copy of
10887 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
10888 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
10889 little endian and -maltivec=be is not specified.
10890 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
10891 little endian and -maltivec=be is not specified.
10893 2014-02-05 Richard Henderson <rth@redhat.com>
10896 * combine-stack-adj.c: Revert r206943.
10897 * sched-int.h (struct deps_desc): Add last_args_size.
10898 * sched-deps.c (init_deps): Initialize it.
10899 (sched_analyze_insn): Add OUTPUT dependencies between insns that
10900 contain REG_ARGS_SIZE notes.
10902 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10904 * lto-cgraph.c (asm_nodes_output): Make global.
10905 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
10906 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
10907 (driver_handle_option): Handle OPT_fwpa.
10909 2014-02-05 Jakub Jelinek <jakub@redhat.com>
10912 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
10913 a comment typo and formatting issue. If odr_hash hasn't been
10914 created, return vNULL and set *completep to false.
10916 PR middle-end/57499
10917 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
10918 bb with no successors.
10920 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
10923 * doc/invoke.texi (-march): Clarify documentation for ARM.
10924 (-mtune): Likewise.
10927 2014-02-05 Richard Biener <rguenther@suse.de>
10929 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
10930 when not vectorizing because of too many alias checks.
10931 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
10932 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
10934 2014-02-05 Nick Clifton <nickc@redhat.com>
10936 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
10937 accept extended registers in any mode when compiling for the MN10300.
10939 2014-02-05 Yury Gribov <y.gribov@samsung.com>
10941 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
10942 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
10943 sanitization attributes.
10944 (can_inline_edge_p): Likewise.
10945 (sanitize_attrs_match_for_inline_p): New function.
10947 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10949 * ipa-prop.c (detect_type_change): Shor circuit testing of
10950 type changes on THIS pointer.
10952 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
10955 * config/pa/pa.c (legitimize_tls_address): Return original address
10956 if not passed a SYMBOL_REF rtx.
10957 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
10959 (pa_emit_move_sequence): Simplify TLS source operands.
10960 (pa_legitimate_constant_p): Reject all TLS constants.
10961 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
10962 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
10964 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10966 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
10967 groups when we know they are controlled by LTO.
10968 * varasm.c (default_binds_local_p_1): If object is in other partition,
10969 it will be resolved locally.
10971 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10973 * config/host-linux.c (linux_gt_pch_use_address): Don't
10974 use SSIZE_MAX because it is not always defined.
10976 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
10979 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
10980 threshold for pseudo splitting.
10981 (update_ebb_live_info): Process call argument hard registers and
10982 hard registers from insn definition too.
10983 (max_small_class_regs_num): New constant.
10984 (inherit_in_ebb): Update live hard regs through EBBs. Update
10985 reloads_num only for small register classes. Don't split for
10988 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
10991 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
10994 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10996 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
10997 visibility is safe.
10999 2014-02-04 Marek Polacek <polacek@redhat.com>
11001 * gdbinit.in (pel): Define.
11003 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
11005 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
11008 2014-02-04 Richard Biener <rguenther@suse.de>
11011 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
11012 in function context local.
11013 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
11014 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
11015 similar to LTO_imported_decl_ref.
11017 2014-02-04 Jakub Jelinek <jakub@redhat.com>
11019 PR tree-optimization/60002
11020 * cgraphclones.c (build_function_decl_skip_args): Clear
11021 DECL_LANG_SPECIFIC.
11023 PR tree-optimization/60023
11024 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
11025 false to gsi_replace.
11026 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
11027 has been in some EH region and vec_stmt could throw, add
11028 vec_stmt into the same EH region.
11029 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
11030 has no lhs, ignore it.
11031 * internal-fn.c (expand_MASK_LOAD): Likewise.
11034 * tree-inline.c (copy_forbidden): Fail for
11035 __attribute__((optimize (0))) functions.
11038 * omp-low.c (simd_clone_struct_copy): If from->inbranch
11039 is set, copy one less argument.
11040 (expand_simd_clones): Don't subtract clone_info->inbranch
11041 from simd_clone_struct_alloc argument.
11043 PR rtl-optimization/57915
11044 * recog.c (simplify_while_replacing): If all unary/binary/relational
11045 operation arguments are constant, attempt to simplify those.
11047 PR middle-end/59261
11048 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
11049 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
11051 2014-02-04 Richard Biener <rguenther@suse.de>
11053 PR tree-optimization/60012
11054 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
11055 TBAA disambiguation to all DDRs.
11057 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11060 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
11061 (LINK_SPEC): Use it for -shared, -shared-libgcc.
11063 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11066 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
11068 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11070 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
11071 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
11073 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11076 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
11077 to figure out targets of polymorphic calls with known decl.
11078 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
11079 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
11080 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
11081 (get_polymorphic_call_info): ... here.
11082 (get_polymorphic_call_info_from_invariant): New function.
11084 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11086 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
11087 lookup via vtable pointer; check for type consistency
11088 and turn inconsitent facts into UNREACHABLE.
11089 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
11090 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
11091 type inconsistent querries; return UNREACHABLE instead.
11093 2014-02-03 Richard Henderson <rth@twiddle.net>
11096 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
11097 already processed this node.
11098 (normalize_one_pred_1): Pass along mark_set.
11099 (normalize_one_pred): Create and destroy a pointer_set_t.
11100 (normalize_one_pred_chain): Likewise.
11102 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
11104 PR gcov-profile/58602
11105 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
11107 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
11110 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
11111 -fno-devirtualize; try to devirtualize by the knowledge of
11112 virtual table pointer given by aggregate propagation.
11113 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
11114 (ipa_print_node_jump_functions): Dump also offset that
11115 is relevant for polymorphic calls.
11116 (determine_known_aggregate_parts): Add arg_type parameter; use it
11117 instead of determining the type from pointer type.
11118 (ipa_compute_jump_functions_for_edge): Update call of
11119 determine_known_aggregate_parts.
11120 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
11121 (gimple_get_virt_method_for_binfo): ... here; simplify using
11122 vtable_pointer_value_to_vtable.
11123 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
11124 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
11125 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
11126 (vtable_pointer_value_to_vtable): Break out from ...; handle also
11128 (vtable_pointer_value_to_binfo): ... here.
11129 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
11131 2014-02-03 Teresa Johnson <tejohnson@google.com>
11133 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
11134 redef of outer loop index variable.
11136 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
11140 * doc/extend.texi (Function Attributes): Typo.
11142 2014-02-03 Cong Hou <congh@google.com>
11144 PR tree-optimization/60000
11145 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
11146 if the vectorized statement is a store. A store statement can only
11147 appear at the end of pattern statements.
11149 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
11151 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
11152 (ix86_option_override_internal): Default long double to 64-bit for
11153 32-bit Bionic and to 128-bit for 64-bit Bionic.
11155 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
11156 TARGET_LONG_DOUBLE_128 is true.
11157 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
11159 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
11160 (mlong-double-64): Negate -mlong-double-128.
11161 (mlong-double-128): New option.
11163 * config/i386/i386-c.c (ix86_target_macros): Define
11164 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
11166 * doc/invoke.texi: Document -mlong-double-128.
11168 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
11170 PR rtl-optimization/60024
11171 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
11173 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
11175 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
11177 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
11179 PR rtl-optimization/57662
11180 * sel-sched.c (code_motion_path_driver): Do not mark already not
11181 existing blocks in the visiting bitmap.
11183 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
11185 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
11186 on the insn being emitted.
11188 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
11189 Will Deacon <will.deacon@arm.com>
11191 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
11193 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11195 * config/arm/arm-tables.opt: Regenerate.
11197 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11199 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
11200 for vector types other than V16QImode.
11201 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
11202 define_expand, and call altivec_expand_vec_perm_le when producing
11203 code with little endian element order.
11204 (*altivec_vperm_<mode>_internal): New insn having previous
11205 behavior of altivec_vperm_<mode>.
11206 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
11207 altivec_expand_vec_perm_le when producing code with little endian
11209 (*altivec_vperm_<mode>_uns_internal): New insn having previous
11210 behavior of altivec_vperm_<mode>_uns.
11212 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11214 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
11215 (altivec_vsumsws): Add handling for -maltivec=be with a little
11217 (altivec_vsumsws_direct): New.
11218 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
11219 gen_altivec_vsumsws.
11221 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
11223 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
11224 vtable_pointer_value_to_binfo): New functions.
11225 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
11226 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
11228 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
11230 * config/nios2/nios2.md (load_got_register): Initialize GOT
11231 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
11232 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
11234 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
11236 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
11237 preserverd by passthrough, do not propagate the type.
11239 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11241 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
11242 (mips_atomic_assign_expand_fenv): New function.
11243 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
11245 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11247 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
11248 (__builtin_mips_set_fcsr): Likewise.
11249 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
11250 MIPS_USI_FTYPE_VOID.
11251 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
11252 (mips16_expand_set_fcsr): Likewise.
11253 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
11254 (mips16_set_fcsr_stub): Likewise.
11255 (mips16_get_fcsr_one_only_stub): New class.
11256 (mips16_set_fcsr_one_only_stub): Likewise.
11257 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
11258 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
11259 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
11260 (hard_float): New availability predicate.
11261 (mips_builtins): Add get_fcsr and set_fcsr.
11262 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
11263 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
11264 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
11265 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
11266 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
11269 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11271 * config/mips/mips.c (mips_one_only_stub): New class.
11272 (mips_need_mips16_rdhwr_p): Replace with...
11273 (mips16_rdhwr_stub): ...this new variable.
11274 (mips16_stub_call_address): New function.
11275 (mips16_rdhwr_one_only_stub): New class.
11276 (mips_expand_thread_pointer): Use mips16_stub_call_address.
11277 (mips_output_mips16_rdhwr): Delete.
11278 (mips_finish_stub): New function.
11279 (mips_code_end): Use it to handle rdhwr stubs.
11281 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
11284 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
11285 when calculating size of integer atomic types.
11287 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
11289 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
11291 2014-02-01 Jakub Jelinek <jakub@redhat.com>
11293 PR tree-optimization/60003
11294 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
11295 * profile.c (branch_prob): Use gimple_call_builtin_p
11296 to check for BUILT_IN_SETJMP_RECEIVER.
11297 * tree-inline.c (copy_bb): Call notice_special_calls.
11299 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
11302 * lra-constraints.c (process_alt_operands): Update reload_sum only
11305 2014-01-31 Richard Henderson <rth@redhat.com>
11307 PR middle-end/60004
11308 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
11309 until after else_eh is processed.
11311 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
11313 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
11314 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
11315 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
11316 in smmintrin.h, remove them.
11317 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
11318 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
11319 * config/i386/i386.md (ROUND_SAE): Fix value.
11320 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
11321 (const48_operand): New.
11322 * config/i386/subst.md (round), (round_expand): Use
11323 const_4_or_8_to_11_operand.
11324 (round_saeonly), (round_saeonly_expand): Use const48_operand.
11326 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
11328 * config/i386/constraints.md (Yk): Swap meaning with k.
11329 * config/i386/i386.md (movhi_internal): Change Yk to k.
11330 (movqi_internal): Ditto.
11331 (*k<logic><mode>): Ditto.
11334 (kandn<mode>): Ditto.
11335 (*<code>hi_1): Ditto.
11336 (*<code>qi_1): Ditto.
11337 (kxnor<mode>): Ditto.
11338 (kortestzhi): Ditto.
11339 (kortestchi): Ditto.
11341 (*one_cmplhi2_1): Ditto.
11342 (*one_cmplqi2_1): Ditto.
11343 * config/i386/sse.md (): Change k to Yk.
11344 (avx512f_load<mode>_mask): Ditto.
11345 (avx512f_blendm<mode>): Ditto.
11346 (avx512f_store<mode>_mask): Ditto.
11347 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
11348 (avx512f_storedqu<mode>_mask): Ditto.
11349 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
11351 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
11352 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
11353 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
11354 (avx512f_maskcmp<mode>3): Ditto.
11355 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
11356 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
11357 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
11358 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
11359 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
11360 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
11361 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
11362 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
11363 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
11364 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
11365 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
11366 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
11367 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
11368 (vec_extract_lo_<mode>_maskm): Ditto.
11369 (vec_extract_hi_<mode>_maskm): Ditto.
11370 (avx512f_vternlog<mode>_mask): Ditto.
11371 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
11372 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
11373 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
11374 (avx512f_<code>v8div16qi2_mask): Ditto.
11375 (avx512f_<code>v8div16qi2_mask_store): Ditto.
11376 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
11377 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
11378 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
11379 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
11380 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
11381 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11382 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
11383 (*avx512pf_scatterpf<mode>df_mask): Ditto.
11384 (avx512cd_maskb_vec_dupv8di): Ditto.
11385 (avx512cd_maskw_vec_dupv16si): Ditto.
11386 (avx512f_vpermi2var<mode>3_maskz): Ditto.
11387 (avx512f_vpermi2var<mode>3_mask): Ditto.
11388 (avx512f_vpermi2var<mode>3_mask): Ditto.
11389 (avx512f_vpermt2var<mode>3_maskz): Ditto.
11390 (*avx512f_gathersi<mode>): Ditto.
11391 (*avx512f_gathersi<mode>_2): Ditto.
11392 (*avx512f_gatherdi<mode>): Ditto.
11393 (*avx512f_gatherdi<mode>_2): Ditto.
11394 (*avx512f_scattersi<mode>): Ditto.
11395 (*avx512f_scatterdi<mode>): Ditto.
11396 (avx512f_compress<mode>_mask): Ditto.
11397 (avx512f_compressstore<mode>_mask): Ditto.
11398 (avx512f_expand<mode>_mask): Ditto.
11399 * config/i386/subst.md (mask): Change k to Yk.
11400 (mask_scalar_merge): Ditto.
11403 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
11405 * doc/extend.texi (Vector Extensions): Document ?: in C++.
11407 2014-01-31 Richard Biener <rguenther@suse.de>
11409 PR middle-end/59990
11410 * builtins.c (fold_builtin_memory_op): Make sure to not
11411 use a floating-point mode or a boolean or enumeral type for
11412 the copy operation.
11414 2014-01-30 DJ Delorie <dj@redhat.com>
11416 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
11417 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
11418 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
11419 whenever main() has an epilogue.
11421 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11423 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
11424 unused variable "field".
11425 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
11426 (vsx_mergeh_<mode>): Likewise.
11427 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
11428 (altivec_vmrghh): Likewise.
11429 (altivec_vmrghw): Likewise.
11430 (altivec_vmrglb): Likewise.
11431 (altivec_vmrglh): Likewise.
11432 (altivec_vmrglw): Likewise.
11433 (altivec_vspltb): Add missing uses.
11434 (altivec_vsplth): Likewise.
11435 (altivec_vspltw): Likewise.
11436 (altivec_vspltsf): Likewise.
11438 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11441 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
11442 frame related instructions.
11444 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
11446 PR rtl-optimization/59959
11447 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
11448 any reload of register whose subreg is invalid.
11450 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11452 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
11453 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
11454 Add missing return type - void.
11456 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11458 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
11459 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
11460 remove element index adjustment for endian (now handled in vsx.md
11462 (altivec_expand_vec_perm_const): Use
11463 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
11464 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
11465 (vsx_xxspltw_<mode>): Adjust element index for little endian.
11466 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
11467 define_expand and a new define_insn *altivec_vspltb_internal;
11468 adjust for -maltivec=be on a little endian target.
11469 (altivec_vspltb_direct): New.
11470 (altivec_vsplth): Divide into a define_expand and a new
11471 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
11472 little endian target.
11473 (altivec_vsplth_direct): New.
11474 (altivec_vspltw): Divide into a define_expand and a new
11475 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
11476 little endian target.
11477 (altivec_vspltw_direct): New.
11478 (altivec_vspltsf): Divide into a define_expand and a new
11479 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
11480 a little endian target.
11482 2014-01-30 Richard Biener <rguenther@suse.de>
11484 PR tree-optimization/59993
11485 * tree-ssa-forwprop.c (associate_pointerplus): Check we
11486 can propagate form the earlier stmt and avoid the transform
11487 when the intermediate result is needed.
11489 2014-01-30 Alangi Derick <alangiderick@gmail.com>
11491 * README.Portability: Fix typo.
11493 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
11495 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
11496 comparison_operator with ordered_comparison_operator.
11498 2014-01-30 Nick Clifton <nickc@redhat.com>
11500 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
11501 Rename to mn10300_store_multiple_regs.
11502 * config/mn10300/mn10300.c: Likewise.
11503 * config/mn10300/mn10300.md (store_movm): Fix typo: call
11504 store_multiple_regs.
11505 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
11506 Call mn10300_store_multiple_regs.
11508 2014-01-30 Nick Clifton <nickc@redhat.com>
11509 DJ Delorie <dj@redhat.com>
11511 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
11512 %fp 2 to keep registers after it properly word-aligned.
11513 (rl78_alloc_physical_registers_umul): Handle the case where both
11514 input operands are the same.
11516 2014-01-30 Richard Biener <rguenther@suse.de>
11518 PR tree-optimization/59903
11519 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
11522 2014-01-30 Jason Merrill <jason@redhat.com>
11525 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
11528 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
11530 2014-01-30 Richard Biener <rguenther@suse.de>
11532 PR tree-optimization/59951
11533 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
11535 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
11538 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
11539 SFmode to DFmode case.
11541 2014-01-29 DJ Delorie <dj@redhat.com>
11543 * config/msp430/msp430.opt (-minrt): New.
11544 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
11546 (ENDFILE_SPEC): Likewise.
11548 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
11550 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
11551 (estimate_function_body_sizes): Use it.
11553 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
11556 * dwarf2out.c (is_cxx_auto): New.
11557 (is_base_type): Use it.
11558 (gen_type_die_with_usage): Likewise.
11560 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11562 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
11563 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
11564 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
11565 -maltivec=be with LE targets.
11566 (vsx_mergeh_<mode>): Likewise.
11567 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
11568 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
11569 (altivec_vmrghb): Replace with define_expand and new
11570 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
11571 (altivec_vmrghb_direct): New define_insn.
11572 (altivec_vmrghh): Replace with define_expand and new
11573 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
11574 (altivec_vmrghh_direct): New define_insn.
11575 (altivec_vmrghw): Replace with define_expand and new
11576 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
11577 (altivec_vmrghw_direct): New define_insn.
11578 (*altivec_vmrghsf): Adjust for endianness.
11579 (altivec_vmrglb): Replace with define_expand and new
11580 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
11581 (altivec_vmrglb_direct): New define_insn.
11582 (altivec_vmrglh): Replace with define_expand and new
11583 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
11584 (altivec_vmrglh_direct): New define_insn.
11585 (altivec_vmrglw): Replace with define_expand and new
11586 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
11587 (altivec_vmrglw_direct): New define_insn.
11588 (*altivec_vmrglsf): Adjust for endianness.
11589 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11590 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11591 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11592 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11593 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11594 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11595 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11596 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11598 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
11600 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
11601 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
11604 2014-01-29 Richard Biener <rguenther@suse.de>
11606 PR tree-optimization/58742
11607 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
11608 associate_pointerplus_align.
11609 (associate_pointerplus_diff): New function.
11610 (associate_pointerplus): Likewise. Call associate_pointerplus_align
11611 and associate_pointerplus_diff.
11613 2014-01-29 Richard Biener <rguenther@suse.de>
11615 * lto-streamer.h (LTO_major_version): Bump to 3.
11616 (LTO_minor_version): Reset to 0.
11618 2014-01-29 Renlin Li <Renlin.Li@arm.com>
11620 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
11621 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
11622 (arm_file_start): Generate correct asm header for armv7ve.
11623 * config/arm/bpabi.h: Add multilib support for armv7ve.
11624 * config/arm/driver-arm.c: Change the architectures of cortex-a7
11625 and cortex-a15 to armv7ve.
11626 * config/arm/t-aprofile: Add multilib support for armv7ve.
11627 * doc/invoke.texi: Document -march=armv7ve.
11629 2014-01-29 Richard Biener <rguenther@suse.de>
11631 PR tree-optimization/58742
11632 * tree-ssa-forwprop.c (associate_plusminus): Return true
11633 if we changed sth, defer EH cleanup to ...
11634 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
11635 (simplify_mult): New function.
11637 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11639 PR middle-end/59917
11640 PR tree-optimization/59920
11641 * tree.c (build_common_builtin_nodes): Remove
11642 __builtin_setjmp_dispatcher initialization.
11643 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
11644 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
11645 instead of gsi_after_labels + manually skipping debug stmts.
11646 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
11647 ignore bbs with IFN_ABNORMAL_DISPATCHER.
11648 * tree-inline.c (copy_edges_for_bb): Remove
11649 can_make_abnormal_goto argument, instead add abnormal_goto_dest
11650 argument. Ignore computed_goto_p stmts. Don't call
11651 make_abnormal_goto_edges. If a call might need abnormal edges
11652 for non-local gotos, see if it already has an edge to
11653 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
11654 with true argument, don't do anything then, otherwise add
11655 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
11656 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
11658 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
11659 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
11660 (lower_stmt): Don't set data->calls_builtin_setjmp.
11661 (lower_builtin_setjmp): Adjust comment.
11662 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
11663 * tree-cfg.c (found_computed_goto): Remove.
11664 (factor_computed_gotos): Remove.
11665 (make_goto_expr_edges): Return bool, true for computed gotos.
11666 Don't call make_abnormal_goto_edges.
11667 (build_gimple_cfg): Don't set found_computed_goto, don't call
11668 factor_computed_gotos.
11669 (computed_goto_p): No longer static.
11670 (make_blocks): Don't set found_computed_goto.
11671 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
11672 (make_edges): If make_goto_expr_edges returns true, push bb
11673 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
11674 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
11675 vector. Record mapping between bbs and OpenMP regions if there
11676 are any, adjust make_gimple_omp_edges caller. Call
11677 handle_abnormal_edges.
11678 (make_abnormal_goto_edges): Remove.
11679 * tree-cfg.h (make_abnormal_goto_edges): Remove.
11680 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
11681 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
11682 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
11683 * internal-fn.def (ABNORMAL_DISPATCHER): New.
11684 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
11685 filling *region also set *region_idx to (*region)->entry->index.
11688 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
11689 For REGs set ORIGINAL_REGNO.
11691 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
11693 * doc/md.texi: Mention that a target shouldn't implement
11694 vec_widen_(s|u)mul_even/odd pair if it is less efficient
11697 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11699 PR tree-optimization/59594
11700 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
11701 a copy of the datarefs vector rather than the vector itself.
11703 2014-01-28 Jason Merrill <jason@redhat.com>
11706 * dwarf2out.c (auto_die): New static.
11707 (gen_type_die_with_usage): Handle C++1y 'auto'.
11708 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
11711 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
11714 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
11715 (SPEC_X32): Likewise.
11716 (SPEC_64): Likewise.
11717 * config/i386/i386.c (ix86_option_override_internal): Turn off
11718 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
11720 (x86_file_start): Output .code16gcc for TARGET_16BIT.
11721 * config/i386/i386.h (TARGET_16BIT): New macro.
11722 (TARGET_16BIT_P): Likewise.
11723 * config/i386/i386.opt: Add m16.
11724 * doc/invoke.texi: Document -m16.
11726 2014-01-28 Jakub Jelinek <jakub@redhat.com>
11728 PR preprocessor/59935
11729 * input.c (location_get_source_line): Bail out on when line number
11730 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
11732 2014-01-28 Richard Biener <rguenther@suse.de>
11734 PR tree-optimization/58742
11735 * tree-ssa-forwprop.c (associate_plusminus): Handle
11736 pointer subtraction of the form (T)(P + A) - (T)P.
11738 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11740 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
11743 2014-01-28 Richard Biener <rguenther@suse.de>
11746 2014-01-28 Richard Biener <rguenther@suse.de>
11748 PR rtl-optimization/45364
11749 PR rtl-optimization/59890
11750 * var-tracking.c (local_get_addr_clear_given_value): Handle
11751 already cleared slot.
11752 (val_reset): Handle not allocated local_get_addr_cache.
11753 (vt_find_locations): Use post-order on the inverted CFG.
11755 2014-01-28 Richard Biener <rguenther@suse.de>
11757 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
11759 2014-01-28 Richard Biener <rguenther@suse.de>
11761 PR rtl-optimization/45364
11762 PR rtl-optimization/59890
11763 * var-tracking.c (local_get_addr_clear_given_value): Handle
11764 already cleared slot.
11765 (val_reset): Handle not allocated local_get_addr_cache.
11766 (vt_find_locations): Use post-order on the inverted CFG.
11768 2014-01-28 Alan Modra <amodra@gmail.com>
11770 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
11771 * configure.ac <recursive call for build != host>: Define
11772 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
11773 and LD_FOR_BUILD too.
11774 * configure: Regenerate.
11776 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
11778 * config/i386/i386.c (get_builtin_code_for_version): Separate
11779 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
11780 Broadwell from Haswell.
11782 2014-01-27 Steve Ellcey <sellcey@mips.com>
11784 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
11785 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
11786 * config/mips/mips.c (mips_option_override): Change setting
11788 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
11789 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
11790 Change from Mask to Var.
11792 2014-01-27 Jeff Law <law@redhat.com>
11794 * ipa-inline.c (inline_small_functions): Fix typo.
11796 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
11798 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
11799 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
11800 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
11801 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
11802 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
11803 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
11804 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
11805 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
11806 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
11807 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
11808 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
11809 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
11810 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
11811 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
11812 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
11813 (_mm512_storeu_epi64): Ditto.
11814 (_mm512_cmpge_epi32_mask): Ditto.
11815 (_mm512_cmpge_epu32_mask): Ditto.
11816 (_mm512_cmpge_epi64_mask): Ditto.
11817 (_mm512_cmpge_epu64_mask): Ditto.
11818 (_mm512_cmple_epi32_mask): Ditto.
11819 (_mm512_cmple_epu32_mask): Ditto.
11820 (_mm512_cmple_epi64_mask): Ditto.
11821 (_mm512_cmple_epu64_mask): Ditto.
11822 (_mm512_cmplt_epi32_mask): Ditto.
11823 (_mm512_cmplt_epu32_mask): Ditto.
11824 (_mm512_cmplt_epi64_mask): Ditto.
11825 (_mm512_cmplt_epu64_mask): Ditto.
11826 (_mm512_cmpneq_epi32_mask): Ditto.
11827 (_mm512_cmpneq_epu32_mask): Ditto.
11828 (_mm512_cmpneq_epi64_mask): Ditto.
11829 (_mm512_cmpneq_epu64_mask): Ditto.
11830 (_mm512_expand_pd): Ditto.
11831 (_mm512_expand_ps): Ditto.
11832 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
11833 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
11834 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
11835 * config/i386/i386.c (ix86_builtins): Add
11836 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
11837 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
11838 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
11839 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
11840 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
11841 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
11842 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
11843 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
11844 IX86_BUILTIN_PMOVUSQW512_MEM.
11845 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
11846 __builtin_ia32_pmovsqd512mem_mask,
11847 __builtin_ia32_pmovqd512mem_mask,
11848 __builtin_ia32_pmovusqw512mem_mask,
11849 __builtin_ia32_pmovsqw512mem_mask,
11850 __builtin_ia32_pmovqw512mem_mask,
11851 __builtin_ia32_pmovusdw512mem_mask,
11852 __builtin_ia32_pmovsdw512mem_mask,
11853 __builtin_ia32_pmovdw512mem_mask,
11854 __builtin_ia32_pmovqb512mem_mask,
11855 __builtin_ia32_pmovusqb512mem_mask,
11856 __builtin_ia32_pmovsqb512mem_mask,
11857 __builtin_ia32_pmovusdb512mem_mask,
11858 __builtin_ia32_pmovsdb512mem_mask,
11859 __builtin_ia32_pmovdb512mem_mask.
11860 (bdesc_args): Add __builtin_ia32_expanddf512,
11861 __builtin_ia32_expandsf512.
11862 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
11863 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
11864 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
11865 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
11866 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
11867 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
11868 (avx512f_<code>v8div16qi2_mask_store): This.
11869 (avx512f_expand<mode>): New.
11871 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
11873 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
11875 (_mm512_mask_prefetch_i64gather_pd): Ditto.
11876 (_mm512_prefetch_i32scatter_pd): Ditto.
11877 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
11878 (_mm512_prefetch_i64scatter_pd): Ditto.
11879 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
11880 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
11881 (_mm512_mask_prefetch_i64gather_ps): Ditto.
11882 (_mm512_prefetch_i32scatter_ps): Ditto.
11883 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
11884 (_mm512_prefetch_i64scatter_ps): Ditto.
11885 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
11886 * config/i386/i386-builtin-types.def: Define
11887 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
11888 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
11889 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
11890 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
11891 IX86_BUILTIN_SCATTERPFQPD.
11892 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
11893 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
11894 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
11895 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
11896 __builtin_ia32_scatterpfqps.
11897 (ix86_expand_builtin): Expand new built-ins.
11898 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
11899 fix memory access data type.
11900 (*avx512pf_gatherpf<mode>_mask): Ditto.
11901 (*avx512pf_gatherpf<mode>): Ditto.
11902 (avx512pf_scatterpf<mode>): Ditto.
11903 (*avx512pf_scatterpf<mode>_mask): Ditto.
11904 (*avx512pf_scatterpf<mode>): Ditto.
11905 (GATHER_SCATTER_SF_MEM_MODE): New.
11906 (avx512pf_gatherpf<mode>df): Ditto.
11907 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11908 (*avx512pf_scatterpf<mode>df): Ditto.
11910 2014-01-27 Jakub Jelinek <jakub@redhat.com>
11913 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
11914 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
11917 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11919 * common/config/arm/arm-common.c
11920 (arm_rewrite_mcpu): Handle multiple names.
11922 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11924 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11926 * gimple-builder.h (create_gimple_tmp): Delete.
11928 2014-01-27 Christian Bruel <christian.bruel@st.com>
11930 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
11933 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
11935 * config/pa/pa.md (call): Generate indirect long calls to non-local
11936 functions when outputing 32-bit code.
11937 (call_value): Likewise except for special call to buggy powf function.
11939 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
11940 portable runtime and PIC indirect calls.
11941 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
11942 and PIC call sequences. Use ldo instead of blr to set return register
11943 in PIC call sequence.
11945 2014-01-25 Walter Lee <walt@tilera.com>
11947 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
11948 avoid clobbering a live register.
11950 2014-01-25 Walter Lee <walt@tilera.com>
11952 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
11953 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
11954 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
11955 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
11957 2014-01-25 Walter Lee <walt@tilera.com>
11959 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
11960 arguments on even registers.
11961 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
11963 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
11964 (BIGGEST_ALIGNMENT): Ditto.
11965 (BIGGEST_FIELD_ALIGNMENT): Ditto.
11967 2014-01-25 Walter Lee <walt@tilera.com>
11969 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
11970 insns before bundling.
11971 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
11973 2014-01-25 Walter Lee <walt@tilera.com>
11975 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
11976 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
11977 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
11979 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11981 * config/mips/constraints.md (kl): Delete.
11982 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
11983 define expands, using...
11984 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
11985 instructions for MIPS16.
11986 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
11987 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
11989 2014-01-25 Walter Lee <walt@tilera.com>
11991 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
11995 2014-01-25 Walter Lee <walt@tilera.com>
11997 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
11998 (TARGET_EXPAND_TO_RTL_HOOK): Define.
12000 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
12002 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
12005 2014-01-25 Jakub Jelinek <jakub@redhat.com>
12007 * print-rtl.c (in_call_function_usage): New var.
12008 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
12009 EXPR_LIST mode as mode and not as reg note name.
12011 PR middle-end/59561
12012 * cfgloopmanip.c (copy_loop_info): If
12013 loop->warned_aggressive_loop_optimizations, make sure
12014 the flag is set in target loop too.
12016 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
12018 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
12020 * builtins.def: Likewise.
12021 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
12022 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
12023 * ira.c (ira_setup_eliminable_regset): Likewise.
12024 * omp-low.c (gate_expand_omp): Likewise.
12025 (execute_lower_omp): Likewise.
12026 (diagnose_sb_0): Likewise.
12027 (gate_diagnose_omp_blocks): Likewise.
12028 (simd_clone_clauses_extract): Likewise.
12031 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12033 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
12034 correction for little endian...
12035 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
12038 2014-01-24 Jeff Law <law@redhat.com>
12040 PR tree-optimization/59919
12041 * tree-vrp.c (find_assert_locations_1): Do not register asserts
12042 for non-returning calls.
12044 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
12046 * common/config/aarch64/aarch64-common.c
12047 (aarch64_rewrite_mcpu): Handle multiple names.
12048 * config/aarch64/aarch64.h
12049 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
12051 2014-01-24 Dodji Seketeli <dodji@redhat.com>
12053 * input.c (add_file_to_cache_tab): Handle the case where fopen
12056 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
12059 * config/i386/i386.md (pushsf splitter): Get stack adjustment
12060 from push operand if code of push isn't PRE_DEC.
12062 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12065 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
12066 -mquad-memory-atomic. Update -mquad-memory documentation to say
12067 it is only used for non-atomic loads/stores.
12069 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
12070 -mquad-memory or -mquad-memory-atomic switches.
12072 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
12073 -mquad-memory-atomic to ISA 2.07 support.
12075 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
12076 to separate support of normal quad word memory operations (ldq, stq)
12077 from the atomic quad word memory operations.
12079 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
12080 support to separate non-atomic quad word operations from atomic
12081 quad word operations. Disable non-atomic quad word operations in
12082 little endian mode so that we don't have to swap words after the
12083 load and before the store.
12084 (quad_load_store_p): Add comment about atomic quad word support.
12085 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
12086 options printed with -mdebug=reg.
12088 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
12089 -mquad-memory-atomic as the test for whether we have quad word
12090 atomic instructions.
12091 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
12092 or -mp8-vector are used, allow byte/half-word atomic operations.
12094 * config/rs6000/sync.md (load_lockedti): Insure that the address
12095 is a proper indexed or indirect address for the lqarx instruction.
12096 On little endian systems, swap the hi/lo registers after the lqarx
12098 (load_lockedpti): Use indexed_or_indirect_operand predicate to
12099 insure the address is valid for the lqarx instruction.
12100 (store_conditionalti): Insure that the address is a proper indexed
12101 or indirect address for the stqcrx. instruction. On little endian
12102 systems, swap the hi/lo registers before doing the stqcrx.
12104 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
12105 insure the address is valid for the stqcrx. instruction.
12107 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12108 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
12109 type of quad memory support is available.
12111 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
12113 PR regression/59915
12114 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
12115 there is a danger of looping.
12117 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
12119 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
12120 force flag_ira_loop_pressure if set via command line.
12122 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
12124 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
12125 (ashr_simd): New builtin handling DI mode.
12126 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
12127 (aarch64_sshr_simddi): New match pattern.
12128 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
12129 (vshrd_n_s64): Likewise.
12130 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
12132 2014-01-23 Nick Clifton <nickc@redhat.com>
12134 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
12135 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
12136 favour of mcu specific scripts.
12137 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
12140 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
12141 Alex Velenko <Alex.Velenko@arm.com>
12143 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
12144 (vaddv_s16): Likewise.
12145 (vaddv_s32): Likewise.
12146 (vaddv_u8): Likewise.
12147 (vaddv_u16): Likewise.
12148 (vaddv_u32): Likewise.
12149 (vaddvq_s8): Likewise.
12150 (vaddvq_s16): Likewise.
12151 (vaddvq_s32): Likewise.
12152 (vaddvq_s64): Likewise.
12153 (vaddvq_u8): Likewise.
12154 (vaddvq_u16): Likewise.
12155 (vaddvq_u32): Likewise.
12156 (vaddvq_u64): Likewise.
12157 (vaddv_f32): Likewise.
12158 (vaddvq_f32): Likewise.
12159 (vaddvq_f64): Likewise.
12160 (vmaxv_f32): Likewise.
12161 (vmaxv_s8): Likewise.
12162 (vmaxv_s16): Likewise.
12163 (vmaxv_s32): Likewise.
12164 (vmaxv_u8): Likewise.
12165 (vmaxv_u16): Likewise.
12166 (vmaxv_u32): Likewise.
12167 (vmaxvq_f32): Likewise.
12168 (vmaxvq_f64): Likewise.
12169 (vmaxvq_s8): Likewise.
12170 (vmaxvq_s16): Likewise.
12171 (vmaxvq_s32): Likewise.
12172 (vmaxvq_u8): Likewise.
12173 (vmaxvq_u16): Likewise.
12174 (vmaxvq_u32): Likewise.
12175 (vmaxnmv_f32): Likewise.
12176 (vmaxnmvq_f32): Likewise.
12177 (vmaxnmvq_f64): Likewise.
12178 (vminv_f32): Likewise.
12179 (vminv_s8): Likewise.
12180 (vminv_s16): Likewise.
12181 (vminv_s32): Likewise.
12182 (vminv_u8): Likewise.
12183 (vminv_u16): Likewise.
12184 (vminv_u32): Likewise.
12185 (vminvq_f32): Likewise.
12186 (vminvq_f64): Likewise.
12187 (vminvq_s8): Likewise.
12188 (vminvq_s16): Likewise.
12189 (vminvq_s32): Likewise.
12190 (vminvq_u8): Likewise.
12191 (vminvq_u16): Likewise.
12192 (vminvq_u32): Likewise.
12193 (vminnmv_f32): Likewise.
12194 (vminnmvq_f32): Likewise.
12195 (vminnmvq_f64): Likewise.
12197 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
12199 * config/aarch64/aarch64-simd.md
12200 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
12201 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
12202 (*aarch64_mul3_elt<mode>): Likewise.
12203 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
12204 (*aarch64_mul3_elt_to_64v2df): Likewise.
12205 (*aarch64_mla_elt<mode>): Likewise.
12206 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
12207 (*aarch64_mls_elt<mode>): Likewise.
12208 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
12209 (*aarch64_fma4_elt<mode>): Likewise.
12210 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
12211 (*aarch64_fma4_elt_to_64v2df): Likewise.
12212 (*aarch64_fnma4_elt<mode>): Likewise.
12213 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
12214 (*aarch64_fnma4_elt_to_64v2df): Likewise.
12215 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
12216 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
12217 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
12218 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
12219 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
12220 (aarch64_sqdmull_lane<mode>_internal): Likewise.
12221 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
12223 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
12225 * config/aarch64/aarch64-simd.md
12226 (aarch64_be_checked_get_lane<mode>): New define_expand.
12227 * config/aarch64/aarch64-simd-builtins.def
12228 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
12229 New builtin definition.
12230 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
12231 Use new safe be builtin.
12233 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
12235 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
12237 (aarch64_be_st1<mode>): Likewise.
12238 (aarch_ld1<VALL:mode>): Define_expand modified.
12239 (aarch_st1<VALL:mode>): Likewise.
12240 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
12241 (UNSPEC_ST1): Likewise.
12243 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
12245 * config/microblaze/microblaze.md: Add trap insn and attribute
12247 2014-01-23 Dodji Seketeli <dodji@redhat.com>
12249 PR preprocessor/58580
12250 * input.h (location_get_source_line): Take an additional line_size
12252 (void diagnostics_file_cache_fini): Declare new function.
12253 * input.c (struct fcache): New type.
12254 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
12255 New static constants.
12256 (diagnostic_file_cache_init, total_lines_num)
12257 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
12258 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
12259 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
12260 (get_next_line, read_next_line, goto_next_line, read_line_num):
12261 New static function definitions.
12262 (diagnostic_file_cache_fini): New function.
12263 (location_get_source_line): Take an additional output line_len
12264 parameter. Re-write using lookup_or_add_file_to_cache_tab and
12266 * diagnostic.c (diagnostic_finish): Call
12267 diagnostic_file_cache_fini.
12268 (adjust_line): Take an additional input parameter for the length
12269 of the line, rather than calculating it with strlen.
12270 (diagnostic_show_locus): Adjust the use of
12271 location_get_source_line and adjust_line with respect to their new
12272 signature. While displaying a line now, do not stop at the first
12273 null byte. Rather, display the zero byte as a space and keep
12274 going until we reach the size of the line.
12275 * Makefile.in: Add vec.o to OBJS-libcommon
12277 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
12278 Ilya Tocar <ilya.tocar@intel.com>
12280 * config/i386/avx512fintrin.h (_mm512_kmov): New.
12281 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
12282 (__builtin_ia32_kmov16): Ditto.
12283 * config/i386/i386.md (UNSPEC_KMOV): New.
12286 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
12288 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
12289 (_mm512_storeu_si512): Ditto.
12291 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
12294 * rtl.h (get_referenced_operands): Declare.
12295 * recog.c (get_referenced_operands): New function.
12296 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
12297 operands have been referenced when recording LO_SUM references.
12299 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
12301 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
12303 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
12305 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
12306 Enable for generic and recent AMD targets.
12308 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
12310 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
12311 ARG_SIZE note when adjustment was eliminated.
12313 2014-01-22 Jeff Law <law@redhat.com>
12315 PR tree-optimization/59597
12316 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
12317 in file. Accept new argument REGISTERING and use it to modify
12318 dump output appropriately.
12319 (register_jump_thread): Corresponding changes.
12320 (mark_threaded_blocks): Reinstate code to cancel unprofitable
12321 thread paths involving joiner blocks. Add code to dump cancelled
12322 jump threading paths.
12324 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
12326 PR rtl-optimization/59477
12327 * lra-constraints.c (inherit_in_ebb): Process call for living hard
12328 regs. Update reloads_num and potential_reload_hard_regs for all insns.
12330 2014-01-22 Tom Tromey <tromey@redhat.com>
12332 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
12334 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
12336 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
12338 PR rtl-optimization/59896
12339 * lra-constraints.c (process_alt_operands): Check unused note for
12340 matched operands of insn with no output reloads.
12342 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
12344 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
12345 (mips_move_from_gpr_cost): Likewise.
12347 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
12349 PR rtl-optimization/59858
12350 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
12351 ira_class_hard_regs_num.
12352 (process_alt_operands): Increase reject for dying matched operand.
12354 2014-01-21 Jakub Jelinek <jakub@redhat.com>
12357 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
12358 smaller than size, perform several stores or loads and stores
12359 at dst + count - size to store or copy all of size bytes, rather
12360 than just last modesize bytes.
12362 2014-01-20 DJ Delorie <dj@redhat.com>
12364 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
12365 that CLOBBERs are REGs before propogating their values.
12367 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
12369 PR middle-end/59789
12370 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
12371 (cgraph_inline_failed_type): New function.
12372 * cgraph.h (DEFCIFCODE): Add type.
12373 (cgraph_inline_failed_type_t): New enum.
12374 (cgraph_inline_failed_type): New prototype.
12375 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
12376 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
12377 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
12378 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
12379 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
12380 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
12381 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
12382 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
12383 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
12384 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
12385 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
12386 OPTIMIZATION_MISMATCH.
12387 * tree-inline.c (expand_call_inline): Emit errors during
12388 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
12390 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12393 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
12394 mode attribute in insn output.
12396 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
12398 * output.h (output_constant): Delete.
12399 * varasm.c (output_constant): Make private.
12401 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
12403 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
12405 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12407 PR middle-end/59860
12408 * tree.h (fold_builtin_strcat): New prototype.
12409 * builtins.c (fold_builtin_strcat): No longer static. Add len
12410 argument, if non-NULL, don't call c_strlen. Optimize
12411 directly into __builtin_memcpy instead of __builtin_strcpy.
12412 (fold_builtin_2): Adjust fold_builtin_strcat caller.
12413 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
12415 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12417 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12418 for SImode_address_operand operands, having only a REG argument.
12420 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
12422 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
12423 loader name using mbig-endian.
12424 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
12426 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12428 * doc/invoke.texi (-march): Clarify documentation for AArch64.
12429 (-mtune): Likewise.
12432 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
12434 * config/aarch64/aarch64-protos.h
12435 (aarch64_cannot_change_mode_class_ptr): Declare.
12436 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
12437 aarch64_cannot_change_mode_class_ptr): New.
12438 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
12439 backend hook aarch64_cannot_change_mode_class.
12441 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12443 * common/config/aarch64/aarch64-common.c
12444 (aarch64_handle_option): Don't handle any option order logic here.
12445 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
12446 selected_cpu, warn on architecture version mismatch.
12447 (aarch64_override_options): Fix parsing order for option strings.
12449 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12450 Iain Sandoe <iain@codesourcery.com>
12453 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
12454 warning. Amend comment to reflect current functionality.
12456 2014-01-20 Richard Biener <rguenther@suse.de>
12458 PR middle-end/59860
12459 * builtins.c (fold_builtin_strcat): Remove case better handled
12460 by tree-ssa-strlen.c.
12462 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
12464 * config/aarch64/aarch64.opt
12465 (mcpu, march, mtune): Make case-insensitive.
12467 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12470 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12471 if operands[1] is a REG or ZERO_EXTEND of a REG.
12473 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
12475 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
12477 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
12479 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
12480 long non-pic millicode calls.
12482 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12484 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
12486 2014-01-19 Kito Cheng <kito@0xlab.org>
12488 * builtins.c (expand_movstr): Check movstr expand done or fail.
12490 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12491 H.J. Lu <hongjiu.lu@intel.com>
12494 * config/i386/i386.md (*lea<mode>): Zero-extend return register
12495 to DImode for zero-extended addresses.
12497 2014-01-19 Jakub Jelinek <jakub@redhat.com>
12499 PR rtl-optimization/57763
12500 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
12501 on the new indirect jump_insn and increment LABEL_NUSES (label).
12503 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
12507 * config.gcc (x86_archs): New variable.
12508 (x86_64_archs): Likewise.
12509 (x86_cpus): Likewise.
12510 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
12511 --with-arch/--with-cpu= options.
12512 Support --with-arch=/--with-cpu={nehalem,westmere,
12513 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
12515 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12517 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
12518 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
12520 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12522 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
12524 2014-01-18 Jakub Jelinek <jakub@redhat.com>
12527 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
12528 clear cpp_get_options (parse_in)->warn_unused_macros for
12529 ix86_target_macros_internal with cpp_define.
12531 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
12533 * jump.c (delete_related_insns): Keep (use (insn))s.
12534 * reorg.c (redundant_insn): Check for barriers too.
12536 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12538 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
12540 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
12542 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
12543 call to $$dyncall when TARGET_LONG_CALLS is true.
12545 2014-01-17 Jeff Law <law@redhat.com>
12547 * ree.c (combine_set_extension): Temporarily disable test for
12548 changing number of hard registers.
12550 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12552 PR middle-end/58125
12553 * ipa-inline-analysis.c (inline_free_summary):
12554 Do not free summary of aliases.
12556 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12558 PR middle-end/59706
12559 * gimplify.c (gimplify_expr): Use create_tmp_var
12560 instead of create_tmp_var_raw. If cond doesn't have
12561 integral type, don't add the IFN_ANNOTATE builtin at all.
12563 2014-01-17 Martin Jambor <mjambor@suse.cz>
12566 * ipa-cp.c (prev_edge_clone): New variable.
12567 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
12568 Also resize prev_edge_clone vector.
12569 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
12570 (ipcp_edge_removal_hook): New function.
12571 (ipcp_driver): Register ipcp_edge_removal_hook.
12573 2014-01-17 Andrew Pinski <apinski@cavium.com>
12574 Steve Ellcey <sellcey@mips.com>
12577 * config/mips/mips.c (mips_print_operand): Check operand mode instead
12580 2014-01-17 Jeff Law <law@redhat.com>
12582 PR middle-end/57904
12583 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
12584 so that pass_ccp runs first.
12586 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12588 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
12589 (ix86_adjust_cost): Use !TARGET_XXX.
12590 (do_reorder_for_imul): Likewise.
12591 (swap_top_of_ready_list): Likewise.
12592 (ix86_sched_reorder): Likewise.
12594 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12596 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
12597 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
12598 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
12599 (intel_memset): New. Duplicate slm_memset.
12600 (intel_cost): New. Duplicate slm_cost.
12601 (m_INTEL): New macro.
12602 (processor_target_table): Add "intel".
12603 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
12604 with PROCESSOR_INTEL for "intel".
12605 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
12606 PROCESSOR_SILVERMONT.
12607 (ix86_issue_rate): Likewise.
12608 (ix86_adjust_cost): Likewise.
12609 (ia32_multipass_dfa_lookahead): Likewise.
12610 (swap_top_of_ready_list): Likewise.
12611 (ix86_sched_reorder): Likewise.
12612 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
12613 instead of TARGET_OPT_AGU.
12614 * config/i386/i386.h (TARGET_INTEL): New.
12615 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
12616 (processor_type): Add PROCESSOR_INTEL.
12617 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
12618 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
12620 2014-01-17 Marek Polacek <polacek@redhat.com>
12623 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
12626 2014-01-17 Richard Biener <rguenther@suse.de>
12628 PR tree-optimization/46590
12629 * opts.c (default_options_table): Add entries for
12630 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
12631 all enabled at -O1 but not for -Og.
12632 * common.opt (fbranch-count-reg): Remove Init(1).
12633 (fmove-loop-invariants): Likewise.
12634 (ftree-pta): Likewise.
12636 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12638 * config/i386/i386.c (ix86_data_alignment): For compatibility with
12639 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
12640 decls to at least the GCC 4.8 used alignments.
12643 * tree-nested.c (convert_nonlocal_reference_stmt,
12644 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
12645 of GIMPLE_BIND stmts, adjust associated decls.
12647 2014-01-17 Richard Biener <rguenther@suse.de>
12649 PR tree-optimization/46590
12650 * vec.h (vec<>::bseach): New member function implementing
12651 binary search according to C89 bsearch.
12652 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
12653 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
12654 bitmap pointer again. Make accesses_in_loop a flat array.
12655 (mem_ref_obstack): New global.
12656 (outermost_indep_loop): Adjust for mem_ref->stored changes.
12657 (mark_ref_stored): Likewise.
12658 (ref_indep_loop_p_2): Likewise.
12659 (set_ref_stored_in_loop): New helper function.
12660 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
12661 (memref_free): Adjust.
12662 (record_mem_ref_loc): Simplify.
12663 (gather_mem_refs_stmt): Adjust.
12664 (sort_locs_in_loop_postorder_cmp): New function.
12665 (analyze_memory_references): Sort accesses_in_loop after
12666 loop postorder number.
12667 (find_ref_loc_in_loop_cmp): New function.
12668 (for_all_locs_in_loop): Find relevant cluster of locs in
12669 accesses_in_loop and iterate without recursion.
12670 (execute_sm): Avoid uninit warning.
12671 (struct ref_always_accessed): Simplify.
12672 (ref_always_accessed::operator ()): Likewise.
12673 (ref_always_accessed_p): Likewise.
12674 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
12675 loop postorder numbers here.
12676 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
12679 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12682 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
12683 on decls for which assemble_alias has been called.
12685 2014-01-17 Nick Clifton <nickc@redhat.com>
12687 * config/msp430/msp430.opt: (mcpu): New option.
12688 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
12689 (msp430_option_override): Parse target_cpu. If the MCU name
12690 matches a generic string, clear target_mcu.
12691 (msp430_attr): Allow numeric interrupt values up to 63.
12692 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
12693 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
12695 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
12697 * config/msp430/msp430.md (popm): Use %J rather than %I.
12698 (addsi3): Use msp430_nonimmediate_operand for operand 2.
12699 (addhi_cy_i): Use immediate_operand for operand 2.
12700 * doc/invoke.texi: Document -mcpu option.
12702 2014-01-17 Richard Biener <rguenther@suse.de>
12704 PR rtl-optimization/38518
12705 * df.h (df_analyze_loop): Declare.
12706 * df-core.c: Include cfgloop.h.
12707 (df_analyze_1): Split out main part of df_analyze.
12708 (df_analyze): Adjust.
12709 (loop_inverted_post_order_compute): New function.
12710 (loop_post_order_compute): Likewise.
12711 (df_analyze_loop): New function avoiding whole-function
12712 postorder computes.
12713 * loop-invariant.c (find_defs): Use df_analyze_loop.
12714 (find_invariants): Adjust.
12715 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
12717 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
12719 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
12720 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
12722 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
12724 * ipa-ref.c (ipa_remove_stmt_references): Fix references
12725 traversal when removing references.
12727 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
12730 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
12732 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
12734 PR middle-end/56791
12735 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
12736 pushing a reload for an autoinc when we had previously reloaded an
12737 inner part of the address.
12739 2014-01-16 Jakub Jelinek <jakub@redhat.com>
12741 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
12743 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
12744 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
12745 when not giving up or versioning for alias only because of
12747 (vect_analyze_data_ref_dependences): Set to true.
12748 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
12750 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
12751 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
12754 PR middle-end/58344
12755 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
12758 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
12759 operand 0 predicate for gathers, use a new pseudo as subtarget.
12761 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12763 PR middle-end/59609
12764 * lra-constraints.c (process_alt_operands): Add printing debug info.
12765 Check absence of input/output reloads for matched operands too.
12767 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12769 PR rtl-optimization/59835
12770 * ira.c (ira_init_register_move_cost): Increase cost for
12773 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
12775 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
12777 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
12780 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
12781 non-register objects. Use gen_(high/low)part more consistently.
12784 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12787 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
12788 endian support, remove tests for WORDS_BIG_ENDIAN.
12789 (p8_mfvsrd_3_<mode>): Likewise.
12790 (reload_gpr_from_vsx<mode>): Likewise.
12791 (reload_gpr_from_vsxsf): Likewise.
12792 (p8_mfvsrd_4_disf): Likewise.
12794 2014-01-16 Richard Biener <rguenther@suse.de>
12796 PR rtl-optimization/46590
12797 * lcm.c (compute_antinout_edge): Use postorder iteration.
12798 (compute_laterin): Use inverted postorder iteration.
12800 2014-01-16 Nick Clifton <nickc@redhat.com>
12802 PR middle-end/28865
12803 * varasm.c (output_constant): Return the number of bytes actually
12805 (output_constructor_array_range): Update the field size with the
12806 number of bytes emitted by output_constant.
12807 (output_constructor_regular_field): Likewise. Also do not
12808 complain if the total number of bytes emitted is now greater
12809 than the expected fieldpos.
12810 * output.h (output_constant): Update prototype and descriptive comment.
12812 2014-01-16 Marek Polacek <polacek@redhat.com>
12814 PR middle-end/59827
12815 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
12816 it is error_mark_node.
12818 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
12820 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
12821 VALID_AVX256_REG_OR_OI_MODE.
12823 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
12825 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
12826 current procedure should be profiled.
12828 2014-01-15 Andrew Pinski <apinski@cavium.com>
12830 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
12831 of moving from/to the STACK_REG register class.
12833 2014-01-15 Richard Henderson <rth@redhat.com>
12836 * reginfo.c (global_regs_decl): Globalize.
12837 * rtl.h (global_regs_decl): Declare.
12838 * ira.c (do_reload): Diagnose frame_pointer_needed and it
12839 reserved via global_regs.
12841 2014-01-15 Teresa Johnson <tejohnson@google.com>
12843 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
12845 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12847 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
12848 and vmulosh rather than call gen_vec_widen_smult_*.
12849 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
12850 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
12851 (vec_widen_smult_even_v16qi): Likewise.
12852 (vec_widen_umult_even_v8hi): Likewise.
12853 (vec_widen_smult_even_v8hi): Likewise.
12854 (vec_widen_umult_odd_v16qi): Likewise.
12855 (vec_widen_smult_odd_v16qi): Likewise.
12856 (vec_widen_umult_odd_v8hi): Likewise.
12857 (vec_widen_smult_odd_v8hi): Likewise.
12858 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
12859 vmuloub rather than call gen_vec_widen_umult_*.
12860 (vec_widen_umult_lo_v16qi): Likewise.
12861 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
12862 vmulosb rather than call gen_vec_widen_smult_*.
12863 (vec_widen_smult_lo_v16qi): Likewise.
12864 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
12865 rather than call gen_vec_widen_umult_*.
12866 (vec_widen_umult_lo_v8hi): Likewise.
12867 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
12868 rather than call gen_vec_widen_smult_*.
12869 (vec_widen_smult_lo_v8hi): Likewise.
12871 2014-01-15 Jeff Law <law@redhat.com>
12873 PR tree-optimization/59747
12874 * ree.c (find_and_remove_re): Properly handle case where a second
12875 eliminated extension requires widening a copy created for elimination
12876 of a prior extension.
12877 (combine_set_extension): Ensure that the number of hard regs needed
12878 for a destination register does not change when we widen it.
12880 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
12882 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
12883 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
12884 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
12885 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
12886 (avr-*-rtems*): Likewise.
12887 (bfin*-rtems*): Likewise.
12888 (moxie-*-rtems*): Likewise.
12889 (h8300-*-rtems*): Likewise.
12890 (i[34567]86-*-rtems*): Likewise.
12891 (lm32-*-rtems*): Likewise.
12892 (m32r-*-rtems*): Likewise.
12893 (m68k-*-rtems*): Likewise.
12894 (microblaze*-*-rtems*): Likewise.
12895 (mips*-*-rtems*): Likewise.
12896 (powerpc-*-rtems*): Likewise.
12897 (sh-*-rtems*): Likewise.
12898 (sparc-*-rtems*): Likewise.
12899 (sparc64-*-rtems*): Likewise.
12900 (v850-*-rtems*): Likewise.
12901 (m32c-*-rtems*): Likewise.
12903 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
12905 PR rtl-optimization/59511
12906 * ira.c (ira_init_register_move_cost): Use memory costs for some
12907 cases of register move cost calculations.
12908 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
12909 instead of BB frequency.
12910 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
12911 * lra-assigns.c (find_hard_regno_for): Ditto.
12913 2014-01-15 Richard Biener <rguenther@suse.de>
12915 PR tree-optimization/59822
12916 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
12917 (vectorizable_load): Use it to hoist defs of uses of invariant
12918 loads out of the loop.
12920 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12921 Kugan Vivekanandarajah <kuganv@linaro.org>
12924 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
12927 2014-01-15 Richard Biener <rguenther@suse.de>
12929 PR rtl-optimization/59802
12930 * lcm.c (compute_available): Use inverted postorder to seed
12931 the initial worklist.
12933 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12936 * config/s390/s390.c (s390_preferred_reload_class): Don't return
12937 ADDR_REGS for invalid symrefs in non-PIC code.
12939 2014-01-15 Jakub Jelinek <jakub@redhat.com>
12942 * builtins.c (determine_block_size): Initialize *probable_max_size
12943 even if len_rtx is CONST_INT.
12945 2014-01-14 Andrew Pinski <apinski@cavium.com>
12947 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
12948 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
12949 (cortexa53_tunings): Likewise.
12950 (aarch64_sched_issue_rate): New function.
12951 (TARGET_SCHED_ISSUE_RATE): Define.
12953 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12955 * ira-costs.c (find_costs_and_classes): Add missed
12956 ira_init_register_move_cost_if_necessary.
12958 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12961 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
12963 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
12966 * config/i386/i386.c (type_natural_mode): Add a bool parameter
12967 to indicate if type is used for function return value. Warn ABI
12968 change if the vector mode isn't available for function return value.
12969 (ix86_function_arg_advance): Pass false to type_natural_mode.
12970 (ix86_function_arg): Likewise.
12971 (ix86_gimplify_va_arg): Likewise.
12972 (function_arg_32): Don't warn ABI change.
12973 (ix86_function_value): Pass true to type_natural_mode.
12974 (ix86_return_in_memory): Likewise.
12975 (ix86_struct_value_rtx): Removed.
12976 (TARGET_STRUCT_VALUE_RTX): Likewise.
12978 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12980 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
12981 converting a conditional jump into a conditional return.
12983 2014-01-14 Richard Biener <rguenther@suse.de>
12985 PR tree-optimization/58921
12986 PR tree-optimization/59006
12987 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
12988 hoisting invariant stmts.
12989 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
12990 invariant loads on the preheader edge if possible.
12992 2014-01-14 Joey Ye <joey.ye@arm.com>
12994 * doc/plugin.texi (Building GCC plugins): Update to C++.
12996 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
12998 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
12999 (_mm_rcp28_round_ss): Ditto.
13000 (_mm_rsqrt28_round_sd): Ditto.
13001 (_mm_rsqrt28_round_ss): Ditto.
13002 (_mm_rcp28_sd): Ditto.
13003 (_mm_rcp28_ss): Ditto.
13004 (_mm_rsqrt28_sd): Ditto.
13005 (_mm_rsqrt28_ss): Ditto.
13006 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
13007 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
13008 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
13009 (IX86_BUILTIN_RCP28SD): Ditto.
13010 (IX86_BUILTIN_RCP28SS): Ditto.
13011 (IX86_BUILTIN_RSQRT28SD): Ditto.
13012 (IX86_BUILTIN_RSQRT28SS): Ditto.
13013 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
13014 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
13015 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
13016 (ix86_expand_special_args_builtin): Expand new FTYPE.
13017 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
13018 (srcp14<mode>): Make insn unary.
13019 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
13020 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
13021 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
13022 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
13023 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
13024 Fix rounding: make it SAE only.
13025 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
13027 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
13029 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
13030 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
13031 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
13032 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
13033 (round_saeonly_mask_scalar_operand4): Ditto.
13034 (round_saeonly_mask_scalar_op3): Ditto.
13035 (round_saeonly_mask_scalar_op4): Ditto.
13037 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13039 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
13040 Implement -maltivec=be for vec_insert and vec_extract.
13042 2014-01-10 DJ Delorie <dj@redhat.com>
13044 * config/msp430/msp430.md (call_internal): Don't allow memory
13045 references with SP as the base register.
13046 (call_value_internal): Likewise.
13047 * config/msp430/constraints.md (Yc): New. For memory references
13048 that don't use SP as a base register.
13050 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
13051 "an integer without a # prefix"
13052 * config/msp430/msp430.md (epilogue_helper): Use it.
13054 2014-01-13 Jakub Jelinek <jakub@redhat.com>
13057 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
13058 AVX512F gather builtins.
13059 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
13060 on gather decls with INTEGER_TYPE masktype.
13061 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
13062 directly into the builtin rather than hoisting it before loop.
13064 PR tree-optimization/59387
13065 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
13066 (scev_const_prop): If folded_casts and type has undefined overflow,
13067 use force_gimple_operand instead of force_gimple_operand_gsi and
13068 for each added stmt if it is assign with
13069 arith_code_with_undefined_signed_overflow, call
13070 rewrite_to_defined_overflow.
13071 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
13072 gimple-fold.h instead.
13073 (arith_code_with_undefined_signed_overflow,
13074 rewrite_to_defined_overflow): Moved to ...
13075 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
13076 rewrite_to_defined_overflow): ... here. No longer static.
13077 Include gimplify-me.h.
13078 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
13079 rewrite_to_defined_overflow): New prototypes.
13081 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13083 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
13085 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
13087 * builtins.c (get_object_alignment_2): Minor tweak.
13088 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
13090 2014-01-13 Christian Bruel <christian.bruel@st.com>
13092 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
13093 optimized non constant lengths.
13095 2014-01-13 Jakub Jelinek <jakub@redhat.com>
13098 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
13099 load as __atomic_load_N if possible.
13101 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
13103 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
13105 (rs6000_expand_builtin): Adjust call.
13107 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
13110 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
13111 * config/rs6000/rs6000.c: Include target-globals.h.
13112 (rs6000_set_current_function): Instead of doing target_reinit
13113 unconditionally, use save_target_globals_default_opts and
13114 restore_target_globals.
13116 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
13118 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
13119 (rs6000_expand_builtin): Handle mffs and mtfsf.
13120 (rs6000_init_builtins): Define mffs and mtfsf.
13121 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
13122 (rs6000_mffs): New pattern.
13123 (rs6000_mtfsf): New pattern.
13125 2014-01-11 Bin Cheng <bin.cheng@arm.com>
13127 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
13128 Start narrowing with START. Apply candidate-use pair
13129 and check overall cost in narrowing.
13130 (iv_ca_prune): Pass new argument.
13132 2014-01-10 Jeff Law <law@redhat.com>
13134 PR middle-end/59743
13135 * ree.c (combine_reaching_defs): Ensure the defining statement
13136 occurs before the extension when optimizing extensions with
13137 different source and destination hard registers.
13139 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
13142 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
13143 vtables into the type inheritance graph.
13145 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13147 PR rtl-optimization/59754
13148 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
13149 modes in the REGNO != REGNO case.
13151 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13153 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
13155 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13157 PR tree-optimization/59745
13158 * tree-predcom.c (tree_predictive_commoning_loop): Call
13159 free_affine_expand_cache if giving up because components is NULL.
13161 * target-globals.c (save_target_globals): Allocate < 4KB structs using
13162 GC in payload of target_globals struct instead of allocating them on
13163 the heap and the larger structs separately using GC.
13164 * target-globals.h (struct target_globals): Make regs, hard_regs,
13165 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
13166 of GTY((skip)) and change type to void *.
13167 (reset_target_globals): Cast loads from those fields to corresponding
13170 2014-01-10 Steve Ellcey <sellcey@mips.com>
13173 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
13174 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
13175 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
13177 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
13180 * aarch64-modes.def (CC_Zmode): New flags mode.
13181 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
13182 represents an equality.
13183 (aarch64_get_condition_code): Handle CC_Zmode.
13184 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
13186 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13188 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
13189 extraction in good case.
13191 2014-01-10 Richard Biener <rguenther@suse.de>
13193 PR tree-optimization/59374
13194 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
13195 checking after SLP discovery. Mark stmts not participating
13196 in any SLP instance properly.
13198 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13200 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
13201 when handling a SET rtx.
13203 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13205 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
13206 (cortex-a57): Likewise.
13207 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
13209 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13211 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
13212 non-iwmmxt builtins.
13214 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
13218 * ipa-devirt.c record_target_from_binfo): Take as argument
13219 stack of binfos and lookup matching one for virtual inheritance.
13220 (possible_polymorphic_call_targets_1): Update.
13222 2014-01-10 Huacai Chen <chenhc@lemote.com>
13224 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
13225 kernel strings for Loongson-2E/2F/3A.
13227 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13229 PR middle-end/59670
13230 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
13231 is_gimple_call before calling gimple_call_internal_p.
13233 2014-01-09 Steve Ellcey <sellcey@mips.com>
13235 * Makefile.in (TREE_FLOW_H): Remove.
13236 (TREE_SSA_H): Add file names from tree-flow.h.
13237 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
13238 * tree.h: Remove tree-flow.h reference.
13239 * hash-table.h: Remove tree-flow.h reference.
13240 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
13241 reference with tree-ssa-loop.h.
13243 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13245 * doc/invoke.texi: Add -maltivec={be,le} options, and document
13246 default element-order behavior for -maltivec.
13247 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
13248 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
13249 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
13250 when targeting big endian, at least for now.
13251 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
13253 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13255 PR middle-end/47735
13256 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
13257 var satisfies use_register_for_decl, just take into account type
13258 alignment, rather than decl alignment.
13260 PR tree-optimization/59622
13261 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
13262 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
13263 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
13264 Don't devirtualize for inplace at all. For targets.length () == 1,
13265 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
13267 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
13269 * config/i386/i386.md (cpu): Remove the unused btver1.
13271 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
13273 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
13275 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13278 * tree-core.h (struct target_globals): New forward declaration.
13279 (struct tree_target_option): Add globals field.
13280 * tree.h (TREE_TARGET_GLOBALS): Define.
13281 (prepare_target_option_nodes_for_pch): New prototype.
13282 * target-globals.h (struct target_globals): Define even if
13283 !SWITCHABLE_TARGET.
13284 * tree.c (prepare_target_option_node_for_pch,
13285 prepare_target_option_nodes_for_pch): New functions.
13286 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
13287 * config/i386/i386.c: Include target-globals.h.
13288 (ix86_set_current_function): Instead of doing target_reinit
13289 unconditionally, use save_target_globals_default_opts and
13290 restore_target_globals.
13292 2014-01-09 Richard Biener <rguenther@suse.de>
13294 PR tree-optimization/59715
13295 * tree-cfg.h (split_critical_edges): Declare.
13296 * tree-cfg.c (split_critical_edges): Export.
13297 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
13299 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
13301 * cfgexpand.c (expand_stack_vars): Optionally disable
13302 asan stack protection.
13303 (expand_used_vars): Likewise.
13304 (partition_stack_vars): Likewise.
13305 * asan.c (asan_emit_stack_protection): Optionally disable
13306 after return stack usage.
13307 (instrument_derefs): Optionally disable memory access instrumentation.
13308 (instrument_builtin_call): Likewise.
13309 (instrument_strlen_call): Likewise.
13310 (asan_protect_global): Optionally disable global variables protection.
13311 * doc/invoke.texi: Added doc for new options.
13312 * params.def: Added new options.
13313 * params.h: Likewise.
13315 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13317 PR rtl-optimization/59724
13318 * ifcvt.c (cond_exec_process_if_block): Don't call
13319 flow_find_head_matching_sequence with 0 longest_match.
13320 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
13321 non-active insns if !stop_after.
13322 (try_head_merge_bb): Revert 2014-01-07 changes.
13324 2014-01-08 Jeff Law <law@redhat.com>
13326 * ree.c (get_sub_rtx): New function, extracted from...
13327 (merge_def_and_ext): Here.
13328 (combine_reaching_defs): Use get_sub_rtx.
13330 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
13332 * cgraph.h (varpool_variable_node): Do not choke on null node.
13334 2014-01-08 Catherine Moore <clm@codesourcery.com>
13336 * config/mips/mips.md (simple_return): Attempt to use JRC
13338 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
13340 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13342 PR rtl-optimization/59137
13343 * reorg.c (steal_delay_list_from_target): Call update_block for
13345 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
13347 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13349 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
13350 two duplicate entries.
13352 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13355 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
13357 * config/mips/mips.c (mips_truncated_op_cost): New function.
13358 (mips_rtx_costs): Adjust test for BADDU.
13359 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
13361 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
13363 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
13364 (*baddu_si): ...this new pattern.
13366 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13369 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
13371 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
13373 PR middle-end/57748
13374 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
13376 (expand_expr, expand_normal): Adjust.
13377 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
13378 inner_reference_p. Use inner_reference_p to expand inner references.
13379 (store_expr): Adjust.
13380 * cfgexpand.c (expand_call_stmt): Adjust.
13382 2014-01-08 Rong Xu <xur@google.com>
13384 * gcov-io.c (gcov_var): Move from gcov-io.h.
13385 (gcov_position): Ditto.
13386 (gcov_is_error): Ditto.
13387 (gcov_rewrite): Ditto.
13388 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
13389 only part to libgcc/libgcov.h.
13391 2014-01-08 Marek Polacek <polacek@redhat.com>
13393 PR middle-end/59669
13394 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
13396 2014-01-08 Marek Polacek <polacek@redhat.com>
13399 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
13401 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13403 PR rtl-optimization/59649
13404 * stor-layout.c (get_mode_bounds): For BImode return
13405 0 and STORE_FLAG_VALUE.
13407 2014-01-08 Richard Biener <rguenther@suse.de>
13409 PR middle-end/59630
13410 * gimple.h (is_gimple_builtin_call): Remove.
13411 (gimple_builtin_call_types_compatible_p): New.
13412 (gimple_call_builtin_p): New overload.
13413 * gimple.c (is_gimple_builtin_call): Remove.
13414 (validate_call): Rename to ...
13415 (gimple_builtin_call_types_compatible_p): ... this and export. Also
13416 check return types.
13417 (validate_type): New static function.
13418 (gimple_call_builtin_p): New overload and adjust.
13419 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
13420 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
13421 (gimple_fold_stmt_to_constant_1): Likewise.
13422 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
13424 2014-01-08 Richard Biener <rguenther@suse.de>
13426 PR middle-end/59471
13427 * gimplify.c (gimplify_expr): Gimplify register-register type
13428 VIEW_CONVERT_EXPRs to separate stmts.
13430 2014-01-07 Jeff Law <law@redhat.com>
13432 PR middle-end/53623
13433 * ree.c (combine_set_extension): Handle case where source
13434 and destination registers in an extension insn are different.
13435 (combine_reaching_defs): Allow source and destination registers
13436 in extension to be different under limited circumstances.
13437 (add_removable_extension): Remove restriction that the
13438 source and destination registers in the extension are the same.
13439 (find_and_remove_re): Emit a copy from the extension's
13440 destination to its source after the defining insn if
13441 the source and destination registers are different.
13443 PR middle-end/59285
13444 * ifcvt.c (merge_if_block): If we are merging a block with more than
13445 one successor with a block with no successors, remove any BARRIER
13446 after the second block.
13448 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
13450 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
13452 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
13455 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
13456 for 14-bit register offsets when INT14_OK_STRICT is false.
13458 2014-01-07 Roland Stigge <stigge@antcom.de>
13459 Michael Meissner <meissner@linux.vnet.ibm.com>
13462 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
13463 Only check TFmode for SPE constants. Don't check TImode or TDmode.
13465 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
13467 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
13470 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
13472 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
13473 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
13474 rtx is const0_rtx or not.
13476 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
13479 * target-globals.c (save_target_globals): Remove this_fn_optab
13481 * toplev.c: Include optabs.h.
13482 (target_reinit): Temporarily restore the global options if another
13483 set of options are in force.
13485 2014-01-07 Jakub Jelinek <jakub@redhat.com>
13487 PR rtl-optimization/58668
13488 * cfgcleanup.c (flow_find_cross_jump): Don't count
13489 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
13490 to determine what is counted.
13491 (flow_find_head_matching_sequence): Use active_insn_p to determine
13493 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
13495 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
13496 determine what is counted.
13498 PR tree-optimization/59643
13499 * tree-predcom.c (split_data_refs_to_components): If one dr is
13500 read and one write, determine_offset fails and the write isn't
13501 in the bad component, just put the read into the bad component.
13503 2014-01-07 Mike Stump <mikestump@comcast.net>
13504 Jakub Jelinek <jakub@redhat.com>
13507 * tree-core.h (struct tree_optimization_option): Change optabs
13508 type from unsigned char * to void *.
13509 * optabs.c (init_tree_optimization_optabs): Adjust
13510 TREE_OPTIMIZATION_OPTABS initialization.
13512 2014-01-06 Jakub Jelinek <jakub@redhat.com>
13515 * config/i386/i386.h (struct machine_function): Add
13516 no_drap_save_restore field.
13517 * config/i386/i386.c (ix86_save_reg): Use
13518 !cfun->machine->no_drap_save_restore instead of
13519 crtl->stack_realign_needed.
13520 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
13521 this function clears frame_pointer_needed. Set
13522 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
13523 and DRAP reg is needed.
13525 2014-01-06 Marek Polacek <polacek@redhat.com>
13528 * doc/implement-c.texi: Mention that other integer types are
13529 permitted as bit-field types in strictly conforming mode.
13531 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
13533 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
13534 is newly allocated.
13536 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
13538 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
13540 2014-01-06 Martin Jambor <mjambor@suse.cz>
13543 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
13545 * ipa-prop.c (ipa_print_node_params): Fix indentation.
13547 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
13551 * var-tracking.c (add_stores): Preserve the value of the source even if
13552 we don't record the store.
13554 2014-01-06 Terry Guo <terry.guo@arm.com>
13556 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
13558 2014-01-05 Iain Sandoe <iain@codesourcery.com>
13561 * config/darwin.c (darwin_function_section): Adjust return values to
13562 correspond to optimisation changes made in r206070.
13564 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
13566 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
13567 from prefetch_block tune setting.
13568 (nocona_cost): Correct size of prefetch block to 64.
13570 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
13572 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
13573 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
13574 used to save the static chain register in the computation of the offset
13575 from which the FP registers need to be restored.
13577 2014-01-04 Jakub Jelinek <jakub@redhat.com>
13579 PR tree-optimization/59519
13580 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
13581 ICE if get_current_def (current_new_name) is already non-NULL, as long
13582 as it is a phi result of some other phi in *new_exit_bb that has
13585 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
13586 or vmovdqu* for misaligned_operand.
13587 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
13588 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
13589 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
13590 aligned_mem for AVX512F masked aligned load and store builtins and for
13591 non-temporal moves.
13593 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
13595 PR tree-optimization/59651
13596 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13597 Address range for negative step should be added by TYPE_SIZE_UNIT.
13599 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
13601 * config/m68k/m68k.c (handle_move_double): Handle pushes with
13602 overlapping registers also for registers other than the stack pointer.
13604 2014-01-03 Marek Polacek <polacek@redhat.com>
13607 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
13610 2014-01-03 Jakub Jelinek <jakub@redhat.com>
13613 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
13616 * config/i386/i386.md (MODE_SIZE): New mode attribute.
13617 (push splitter): Use <P:MODE_SIZE> instead of
13618 GET_MODE_SIZE (<P:MODE>mode).
13619 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
13620 (mov -1, reg peephole2): Likewise.
13621 * config/i386/sse.md (*mov<mode>_internal,
13622 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
13623 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
13624 *<code><mode>3, *andnot<mode>3<mask_name>,
13625 <mask_codefor><code><mode>3<mask_name>): Likewise.
13626 * config/i386/subst.md (mask_mode512bit_condition,
13627 sd_mask_mode512bit_condition): Likewise.
13629 2014-01-02 Xinliang David Li <davidxl@google.com>
13631 PR tree-optimization/59303
13632 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
13633 (dump_predicates): Better output format.
13634 (pred_equal_p): New function.
13635 (is_neq_relop_p): Ditto.
13636 (is_neq_zero_form_p): Ditto.
13637 (pred_expr_equal_p): Ditto.
13638 (pred_neg_p): Ditto.
13639 (simplify_pred): Ditto.
13640 (simplify_preds_2): Ditto.
13641 (simplify_preds_3): Ditto.
13642 (simplify_preds_4): Ditto.
13643 (simplify_preds): Ditto.
13644 (push_pred): Ditto.
13645 (push_to_worklist): Ditto.
13646 (get_pred_info_from_cmp): Ditto.
13647 (is_degenerated_phi): Ditto.
13648 (normalize_one_pred_1): Ditto.
13649 (normalize_one_pred): Ditto.
13650 (normalize_one_pred_chain): Ditto.
13651 (normalize_preds): Ditto.
13652 (normalize_cond_1): Remove function.
13653 (normalize_cond): Ditto.
13654 (is_gcond_subset_of): Ditto.
13655 (is_subset_of_any): Ditto.
13656 (is_or_set_subset_of): Ditto.
13657 (is_and_set_subset_of): Ditto.
13658 (is_norm_cond_subset_of): Ditto.
13659 (pred_chain_length_cmp): Ditto.
13660 (convert_control_dep_chain_into_preds): Type change.
13661 (find_predicates): Ditto.
13662 (find_def_preds): Ditto.
13663 (destroy_predicates_vecs): Ditto.
13664 (find_matching_predicates_in_rest_chains): Ditto.
13665 (use_pred_not_overlap_with_undef_path_pred): Ditto.
13666 (is_pred_expr_subset): Ditto.
13667 (is_pred_chain_subset_of): Ditto.
13668 (is_included_in): Ditto.
13669 (is_superset_of): Ditto.
13671 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13673 Update copyright years.
13675 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13677 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
13678 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
13679 config/arc/arc.md, config/arc/arc.opt,
13680 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
13681 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
13682 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
13683 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
13684 config/linux-protos.h, config/linux.c, config/winnt-c.c,
13685 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
13686 vtable-verify.c, vtable-verify.h: Use the standard form for the
13689 2014-01-02 Tobias Burnus <burnus@net-b.de>
13691 * gcc.c (process_command): Update copyright notice dates.
13692 * gcov-dump.c: Ditto.
13694 * doc/cpp.texi: Bump @copying's copyright year.
13695 * doc/cppinternals.texi: Ditto.
13696 * doc/gcc.texi: Ditto.
13697 * doc/gccint.texi: Ditto.
13698 * doc/gcov.texi: Ditto.
13699 * doc/install.texi: Ditto.
13700 * doc/invoke.texi: Ditto.
13702 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13704 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
13706 2014-01-01 Jakub Jelinek <jakub@redhat.com>
13708 * config/i386/sse.md (*mov<mode>_internal): Guard
13709 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
13711 PR rtl-optimization/59647
13712 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
13713 new_rtx into UNSIGNED_FLOAT rtxes.
13715 Copyright (C) 2014 Free Software Foundation, Inc.
13717 Copying and distribution of this file, with or without modification,
13718 are permitted in any medium without royalty provided the copyright
13719 notice and this notice are preserved.