1 2014-05-31 Andreas Schwab <schwab@linux-m68k.org>
3 * config/m68k/m68k.md (beq0_di, bne0_di): Make the "o" constraint
4 a separate alternative where the scratch operand 2 is marked as
7 2014-05-31 Kugan Vivekanandarajah <kuganv@linaro.org>
9 * config/arm/arm.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New define.
10 (arm_builtins) : Add ARM_BUILTIN_GET_FPSCR and ARM_BUILTIN_SET_FPSCR.
11 (bdesc_2arg) : Add description for builtins __builtins_arm_set_fpscr
12 and __builtins_arm_get_fpscr.
13 (arm_init_builtins) : Initialize builtins __builtins_arm_set_fpscr and
14 __builtins_arm_get_fpscr.
15 (arm_expand_builtin) : Expand builtins __builtins_arm_set_fpscr and
16 __builtins_arm_ldfpscr.
17 (arm_atomic_assign_expand_fenv): New function.
18 * config/arm/vfp.md (set_fpscr): New pattern.
19 (get_fpscr) : Likewise.
20 * config/arm/unspecs.md (unspecv): Add VUNSPEC_GET_FPSCR and
22 * doc/extend.texi (AARCH64 Built-in Functions) : Document
23 __builtins_arm_set_fpscr, __builtins_arm_get_fpscr.
25 2014-05-30 Jakub Jelinek <jakub@redhat.com>
27 * asan.c (report_error_func): Add SLOW_P argument, use
28 BUILT_IN_ASAN_*_N if set.
29 (build_check_stmt): Likewise.
30 (instrument_derefs): If T has insufficient alignment,
31 force same handling as for odd sizes.
33 * sanitizer.def (BUILT_IN_ASAN_REPORT_LOAD_N,
34 BUILT_IN_ASAN_REPORT_STORE_N): New.
35 * asan.c (struct asan_mem_ref): Change access_size type to
37 (asan_mem_ref_init, asan_mem_ref_new, get_mem_refs_of_builtin_call,
38 update_mem_ref_hash_table): Likewise.
39 (asan_mem_ref_hasher::hash): Hash in a HWI.
40 (report_error_func): Change size_in_bytes argument to HWI.
41 Use *_N builtins if size_in_bytes is larger than 16 or not power of
43 (build_shadow_mem_access): New function.
44 (build_check_stmt): Use it. Change size_in_bytes argument to HWI.
45 Handle size_in_bytes not power of two or larger than 16.
46 (instrument_derefs): Don't give up if size_in_bytes is not
47 power of two or is larger than 16.
49 2014-05-30 Kai Tietz <ktietz@redhat.com>
52 * config/i386/i386.c (x86_output_mi_thunk): Add memory case
53 for sibling-tail-calls.
54 * config/i386/i386.md (sibcall_insn_operand): Add memory-constrain
56 * config/i386/predicates.md (sibcall_memory_operand): New predicate.
57 (sibcall_insn_operand): Add check for sibcall_memory_operand.
59 2014-05-30 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
61 * config/avr/avr-mcus.def: Change ATA6289 ISA to AVR4
62 * config/avr/avr-tables.opt: Regenerate.
63 * config/avr/t-multilib: Regenerate.
64 * doc/avr-mmcu.texi: Regenerate.
66 2014-05-30 Ian Lance Taylor <iant@google.com>
68 * config/i386/xmmintrin.h (_mm_pause): Move out of scope of pragma
71 2014-05-30 Tom de Vries <tom@codesourcery.com>
73 * config/i386/i386.c (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS):
76 2014-05-30 Tom de Vries <tom@codesourcery.com>
78 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
79 * lra.c (initialize_lra_reg_info_element): Add init of
80 actual_call_used_reg_set field.
81 (lra): Call lra_create_live_ranges before lra_inheritance for
83 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
85 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
86 instead of call_used_reg_set for -fuse-caller-save.
87 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
89 2014-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
91 * config/arm/thumb2.md (*thumb2_movhi_insn): Set type of movw
93 * config/arm/vfp.md (*thumb2_movsi_vfp): Likewise.
95 2014-05-30 Richard Sandiford <rdsandiford@googlemail.com>
97 * ira.c (ira_get_dup_out_num): Check for output operands at
98 the start of the loop. Handle cases where an included alternative
99 follows an excluded one.
101 2014-05-29 Mike Stump <mikestump@comcast.net>
104 * collect2.c (maybe_run_lto_and_relink): Be sure to always run
105 post ld passes when lto is used.
107 2014-05-29 Vladimir Makarov <vmakarov@redhat.com>
109 PR rtl-optimization/61325
110 * lra-constraints.c (process_address): Rename to
112 (process_address): New function.
114 2014-05-29 Alan Lawrence <alan.lawrence@arm.com>
116 * config/aarch64/aarch64-builtins.c (aarch64_types_binopv_qualifiers,
117 TYPES_BINOPV): New static data.
118 * config/aarch64/aarch64-simd-builtins.def (im_lane_bound): New builtin.
119 * config/aarch64/aarch64-simd.md (aarch64_ext, aarch64_im_lane_boundsi):
121 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_const_1): Match
123 (aarch64_evpc_ext): New function.
125 * config/aarch64/iterators.md (UNSPEC_EXT): New enum element.
127 * config/aarch64/arm_neon.h (vext_f32, vext_f64, vext_p8, vext_p16,
128 vext_s8, vext_s16, vext_s32, vext_s64, vext_u8, vext_u16, vext_u32,
129 vext_u64, vextq_f32, vextq_f64, vextq_p8, vextq_p16, vextq_s8,
130 vextq_s16, vextq_s32, vextq_s64, vextq_u8, vextq_u16, vextq_u32,
131 vextq_u64): Replace __asm with __builtin_shuffle and im_lane_boundsi.
133 2014-05-29 Tom de Vries <tom@codesourcery.com>
135 * rtl.h (BLOCK_SYMBOL_CHECK): Use SYMBOL_REF_FLAGS.
137 2014-05-29 Richard Earnshaw <rearnsha@arm.com>
138 Richard Sandiford <rdsandiford@googlemail.com>
140 * arm/iterators.md (shiftable_ops): New code iterator.
141 (t2_binop0, arith_shift_insn): New code attributes.
142 * arm/predicates.md (shift_nomul_operator): New predicate.
143 * arm/arm.md (insn_enabled): Delete.
144 (enabled): Remove insn_enabled test.
145 (*arith_shiftsi): Delete. Replace with ...
146 (*<arith_shift_insn>_multsi): ... new pattern.
147 (*<arith_shift_insn>_shiftsi): ... new pattern.
148 * config/arm/arm.c (arm_print_operand): Handle operand format 'b'.
150 2014-05-29 Radovan Obradovic <robradovic@mips.com>
151 Tom de Vries <tom@codesourcery.com>
153 * config/mips/mips.h (POST_CALL_TMP_REG): Define.
154 * config/mips/mips.c (mips_emit_call_insn): Add POST_CALL_TMP_REG
156 (mips_split_call): Use POST_CALL_TMP_REG.
157 (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): Redefine to true.
159 2014-05-29 Tom de Vries <tom@codesourcery.com>
161 * final.c (collect_fn_hard_reg_usage): Guard variable declaration
162 with #ifdef STACK_REGS.
164 2014-05-28 Jan Hubicka <hubicka@ucw.cz>
166 * varasm.c (get_variable_section): Walk aliases.
167 (place_block_symbol): Walk aliases.
169 2014-05-28 Tom de Vries <tom@codesourcery.com>
172 2014-05-28 Tom de Vries <tom@codesourcery.com>
174 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
175 * lra.c (initialize_lra_reg_info_element): Add init of
176 actual_call_used_reg_set field.
177 (lra): Call lra_create_live_ranges before lra_inheritance for
179 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
181 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
182 instead of call_used_reg_set for -fuse-caller-save.
183 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
185 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
187 * doc/md.texi: Document that the % constraint character must
188 be at the beginning of the string.
189 * genoutput.c (validate_insn_alternatives): Check that '=',
190 '+' and '%' only appear at the beginning of a constraint.
191 * ira.c (commutative_constraint_p): Delete.
192 (ira_get_dup_out_num): Expect the '%' commutativity marker to be
193 at the start of the string.
194 * config/alpha/alpha.md (*movmemdi_1, *clrmemdi_1): Remove
196 * config/arm/neon.md (bicdi3_neon): Likewise.
197 * config/iq2000/iq2000.md (addsi3_internal, subsi3_internal, sgt_si)
198 (slt_si, sltu_si): Likewise.
199 * config/vax/vax.md (sbcdi3): Likewise.
200 * config/h8300/h8300.md (*cmpstz): Remove duplicate '+'.
201 * config/arc/arc.md (mulsi_600, mulsidi_600, umulsidi_600)
202 (mul64): Move '%' to beginning of constraint.
203 * config/arm/arm.md (*xordi3_insn): Likewise.
204 * config/nds32/nds32.md (add<mode>3, mulsi3, andsi3, iorsi3)
207 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
209 * doc/md.texi: Document the restrictions on the "enabled" attribute.
211 2014-05-28 Jason Merrill <jason@redhat.com>
214 * cgraph.h (symtab_node::get_comdat_group_id): New.
215 * cgraphunit.c (analyze_functions): Call it.
216 * symtab.c (dump_symtab_node): Likewise.
217 * tree.c (decl_comdat_group_id): New.
218 * tree.h: Declare it.
219 * lto-streamer-out.c (write_symbol): Use it.
220 * trans-mem.c (ipa_tm_create_version_alias): Likewise.
222 2014-05-28 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
225 * wide-int.cc: Do not include longlong.h when compiling with clang.
227 2014-05-28 Richard Biener <rguenther@suse.de>
229 * tree-ssa-propagate.c (add_control_edge): Print less vertical space.
230 * tree-vrp.c (extract_range_from_ssa_name): Also copy VR_UNDEFINED.
231 (vrp_visit_assignment_or_call): Print less vertical space.
232 (vrp_visit_stmt): Likewise.
233 (vrp_visit_phi_node): Likewise. For a PHI argument with
234 VR_VARYING range consider recording it as copy.
236 2014-05-28 Richard Biener <rguenther@suse.de>
239 2014-05-28 Richard Biener <rguenther@suse.de>
241 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
243 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
245 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
246 sufficiently aligned and an offset is used at the same time.
247 (expand_expr_real_1): Likewise.
249 2014-05-28 Richard Biener <rguenther@suse.de>
252 * fold-const.c (fold_comparison): When folding
253 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
254 the sign of the remaining constant operand stays the same.
256 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
258 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
259 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
260 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
262 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES
264 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
265 (m32bit-doubles) Likewise.
266 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
267 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
270 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
272 * configure.ac ($gcc_cv_ld_clearcap): New test.
273 * configure: Regenerate.
274 * config.in: Regenerate.
275 * config/sol2.opt (mclear-hwcap): New option.
276 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
277 * config/sol2-clearcap.map: Moved here from
278 testsuite/gcc.target/i386/clearcap.map.
279 * config/sol2-clearcapv2.map: Move here from
280 gcc.target/i386/clearcapv2.map.
281 * config/t-sol2 (install): Depend on install-clearcap-map.
282 (install-clearcap-map): New target.
283 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
286 2014-05-28 Richard Biener <rguenther@suse.de>
288 * hwint.h (*_HALF_WIDE_INT*): Move to ...
289 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
290 ... here and remove the rest.
291 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
293 2014-05-28 Richard Biener <rguenther@suse.de>
295 PR tree-optimization/61335
296 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
297 new range fails, drop to varying.
299 2014-05-28 Tom de Vries <tom@codesourcery.com>
301 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
302 * lra.c (initialize_lra_reg_info_element): Add init of
303 actual_call_used_reg_set field.
304 (lra): Call lra_create_live_ranges before lra_inheritance for
306 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
308 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
309 instead of call_used_reg_set for -fuse-caller-save.
310 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
312 2014-05-28 Radovan Obradovic <robradovic@mips.com>
313 Tom de Vries <tom@codesourcery.com>
315 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save to
317 (@item -fuse-caller-save): New item.
319 2014-05-28 Radovan Obradovic <robradovic@mips.com>
320 Tom de Vries <tom@codesourcery.com>
322 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
323 OPT_fuse_caller_save.
325 2014-05-28 Radovan Obradovic <robradovic@mips.com>
326 Tom de Vries <tom@codesourcery.com>
328 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
329 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
330 get_call_reg_set_usage.
331 * resource.c (mark_set_resources, mark_target_live_regs): Use
332 get_call_reg_set_usage.
333 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
335 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
336 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
337 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
338 * ira-build.c (ira_create_allocno): Init
339 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
340 (create_cap_allocno, propagate_allocno_info)
341 (propagate_some_info_from_allocno)
342 (copy_info_to_removed_store_destinations): Handle
343 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
344 * ira-costs.c (ira_tune_allocno_costs): Use
345 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
347 2014-05-28 Radovan Obradovic <robradovic@mips.com>
348 Tom de Vries <tom@codesourcery.com>
350 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
351 and function_used_regs_valid fields.
352 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
353 find_all_hard_reg_sets.
354 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
355 (get_call_reg_set_usage): New function.
356 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
357 * regs.h (get_call_reg_set_usage): Declare.
359 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
362 * config/dbx.h (License): Add Runtime Library Exception.
363 * config/newlib-stdint.h (License): Same.
364 * config/rtems.h (License): Same
365 * config/initfini-array.h (License): Same
366 * config/v850/v850.h (License): Same.
367 * config/v850/v850-opts.h (License): Same
368 * config/v850/rtems.h (License): Same.
370 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
373 * doc/extend.texi (Local Labels): Note that label differences are
374 not supported for AVR.
376 2014-05-27 Andrew Pinski <apinski@cavium.com>
378 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
379 Use <w> for the register in assembly template.
380 (stack_protect_test): Use the mode of operands[0] for the
382 (stack_protect_test_<mode>): Use <w> for the register
383 in assembly template.
385 2014-05-27 DJ Delorie <dj@redhat.com>
387 * config/rx/rx.c (add_vector_labels): New.
388 (rx_output_function_prologue): Call it.
389 (rx_handle_func_attribute): Don't require empty arguments.
390 (rx_handle_vector_attribute): New.
391 (rx_attribute_table): Add "vector" attribute.
392 * doc/extend.texi (interrupt, vector): Document new/changed
393 RX-specific attributes.
395 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
397 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
399 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
400 predicate to detect a negative quotient.
402 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
404 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
405 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
406 Add X - Y CMP 0 to X CMP Y transformation.
407 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
409 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
411 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
414 2014-05-27 Steve Ellcey <sellcey@mips.com>
416 * config/mips/mips.c: Add include of cgraph.h.
418 2014-05-27 Richard Biener <rguenther@suse.de>
420 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
422 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
425 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
426 * config/arm/arm-cores.def (License): Same.
427 * config/arm/arm-opts.h (License): Same.
428 * config/arm/aout.h (License): Same.
429 * config/arm/bpabi.h (License): Same.
430 * config/arm/elf.h (License): Same.
431 * config/arm/linux-elf.h (License): Same.
432 * config/arm/linux-gas.h (License): Same.
433 * config/arm/netbsd-elf.h (License): Same.
434 * config/arm/uclinux-eabi.h (License): Same.
435 * config/arm/uclinux-elf.h (License): Same.
436 * config/arm/vxworks.h (License): Same.
438 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
440 * config/arm/neon.md (neon_bswap<mode>): New pattern.
441 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
442 (arm_init_neon_builtins): Handle NEON_BSWAP.
443 Define required type nodes.
444 (arm_expand_neon_builtin): Handle NEON_BSWAP.
445 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
446 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
447 * config/arm/iterators.md (VDQHSD): New mode iterator.
449 2014-05-27 Richard Biener <rguenther@suse.de>
451 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
452 Try using literal operands when comparing value-ranges failed.
454 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
456 * ira.c (commutative_operand): Adjust for change to recog_data.
457 [Missing from previous commit.]
459 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
461 * system.h (TEST_BIT): New macro.
462 * recog.h (alternative_mask): New type.
463 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
464 (recog_data_d): Replace alternative_enabled_p array with
465 enabled_alternatives.
466 (target_recog): New structure.
467 (default_target_recog, this_target_recog): Declare.
468 (get_enabled_alternatives, recog_init): Likewise.
469 * recog.c (default_target_recog, this_target_recog): New variables.
470 (get_enabled_alternatives): New function.
471 (extract_insn): Use it.
472 (recog_init): New function.
473 (preprocess_constraints, constrain_operands): Adjust for change to
475 * postreload.c (reload_cse_simplify_operands): Likewise.
476 * reload.c (find_reloads): Likewise.
477 * ira-costs.c (record_reg_classes): Likewise.
478 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
479 all alternatives after a disabled one would be skipped.
480 (ira_implicitly_set_insn_hard_regs): Likewise.
481 * ira.c (ira_setup_alts): Adjust for change to recog_data.
482 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
483 with enabled_alternatives.
484 * lra.c (free_insn_recog_data): Update accordingly.
485 (lra_update_insn_recog_data): Likewise.
486 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
487 * lra-constraints.c (process_alt_operands): Likewise. Handle
488 only_alternative as part of the enabled mask.
489 * target-globals.h (this_target_recog): Declare.
490 (target_globals): Add a recog field.
491 (restore_target_globals): Restore this_target_recog.
492 * target-globals.c: Include recog.h.
493 (default_target_globals): Initialize recog field.
494 (save_target_globals): Likewise.
495 * reginfo.c (reinit_regs): Call recog_init.
496 * toplev.c (backend_init_target): Likewise.
498 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
500 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
501 rather than any named insn's code.
503 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
506 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
507 * config/arm/arm-cores.def (License): Same.
509 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
511 * tree.h (decl_comdat_group): Declare.
512 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
513 * tree.c (decl_comdat_group): Here.
515 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
517 PR rtl-optimization/61222
518 * combine.c (simplify_shift_const_1): When moving a PLUS outside
519 the shift, truncate the PLUS operand to the result mode.
521 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
524 * config/i386/i386.c (ix86_rtx_costs)
525 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
528 2014-05-26 Martin Jambor <mjambor@suse.cz>
530 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
533 2014-05-26 Richard Biener <rguenther@suse.de>
535 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
536 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
537 Provide specializations.
538 (wi::int_traits <HOST_WIDE_INT>,
539 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
541 2014-05-26 Alan Modra <amodra@gmail.com>
544 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
545 params and return a bool. Remove dead code. Update comment.
546 Assert we have a const_int source. Remove bogus code from
547 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
548 handling of constants > 2G and reg_equal note, from..
549 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
550 return value. Update comment. If we can, use a new pseudo
551 for intermediate calculations.
552 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
554 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
555 call to rs6000_emit_set_const in splitter.
556 (movdi_internal64+2, +3): Likewise.
558 2014-05-26 Richard Biener <rguenther@suse.de>
560 * system.h: Define __STDC_FORMAT_MACROS before
561 including inttypes.h.
562 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
563 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
564 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
565 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
566 HOST_WIDEST_INT_C): Remove.
567 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
568 if C99 inttypes.h is not available.
569 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
570 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
571 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
572 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
573 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
574 (struct output_info): Likewise.
575 (print_statistics): Adjust.
576 (dump_bitmap_statistics): Likewise.
577 * bt-load.c (migrate_btr_defs): Print with PRId64.
578 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
579 (MAX_SAFE_MULTIPLIER): Adjust.
580 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
581 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
582 dump_cgraph_node): Likewise.
583 * final.c (dump_basic_block_info): Likewise.
584 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
585 * gcov.c (format_gcov): Likewise.
586 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
588 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
589 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
590 (inline_small_functions, dump_overall_stats, dump_inline_stats):
591 Use PRId64 for dumping.
592 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
593 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
594 (add_allocno_hard_regs): Adjust.
595 * loop-doloop.c (doloop_modify): Print using PRId64.
596 * loop-iv.c (inverse): Compute in uint64_t.
597 (determine_max_iter, iv_number_of_iterations): Likewise.
598 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
600 * lto-streamer-out.c (write_symbol): Use uint64_t.
601 * mcf.c (CAP_INFINITY): Use int64_t maximum.
602 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
603 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
604 * modulo-sched.c (const_iteration_count): Use int64_t.
605 (sms_schedule): Dump using PRId64.
606 * predict.c (dump_prediction): Likewise.
607 * pretty-print.h (pp_widest_integer): Remove.
608 * profile.c (get_working_sets, is_edge_inconsistent,
609 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
610 * tree-pretty-print.c (pp_double_int): Remove case handling
611 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
612 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
614 (pass_optimize_bswap::execute): Remove restriction on hosts.
615 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
616 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
617 * tree.c (widest_int_cst_value): Remove.
618 * tree.h (widest_int_cst_value): Likewise.
619 * value-prof.c (dump_histogram_value): Print using PRId64.
620 * gengtype.c (main): Also inject int64_t.
621 * ggc-page.c (struct max_alignment): Use int64_t.
622 * alloc-pool.c (struct allocation_object_def): Likewise.
623 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
625 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
626 * doc/tm.texi: Regenerated.
627 * gengtype-lex.l (IWORD): Handle [u]int64_t.
628 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
629 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
630 mmix_output_register_setting): Use [u]int64_t in prototypes.
631 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
632 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
633 mmix_output_octa, mmix_output_shifted_value): Adjust.
634 (mmix_intval): Adjust. Remove unreachable case.
635 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
637 2014-05-26 Richard Biener <rguenther@suse.de>
639 * configure.ac: Drop __int64 type check. Insist that we
640 found uint64_t and int64_t.
641 * hwint.h (HOST_BITS_PER___INT64): Remove.
642 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
643 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
644 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
645 (HOST_WIDEST_FAST_INT): Remove __int64 case.
646 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
647 for dst_q_src_df_rms_cdt.
648 * configure: Regenerate.
649 * config.in: Likewise.
651 2014-05-26 Michael Tautschnig <mt@debian.org>
654 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
655 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
657 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
659 PR rtl-optimization/61278
660 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
662 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
664 PR rtl-optimization/61220
665 Part of PR rtl-optimization/61225
666 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
667 insn; skip split_edge for a block with only one successor.
669 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
671 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
674 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
676 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
677 (update_vtable_references): New function.
678 (function_and_variable_visibility): Rewrite also vtable initializers.
679 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
681 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
683 * ggc.h (ggc_grow): New function.
684 * ggc-none.c (ggc_grow): New function.
685 * ggc-page.c (ggc_grow): Likewise.
687 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
689 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
690 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
691 comdat_can_be_unshared_p, cgraph_externally_visible_p,
692 varpool_externally_visible_p, can_replace_by_local_alias,
693 update_visibility_by_resolution_info, function_and_variable_visibility,
694 pass_data_ipa_function_and_variable_visibility,
695 make_pass_ipa_function_and_variable_visibility,
696 whole_program_function_and_variable_visibility,
697 pass_data_ipa_whole_program_visibility,
698 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
699 * cgraph.h (cgraph_local_node_p): Declare.
700 * ipa-visibility.c: New file.
701 * Makefile.in (OBJS): Add ipa-visiblity.o
703 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
705 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
706 that var decl is available.
708 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
710 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
712 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
713 (find_decls_types_r): Do not walk COMDAT_GROUP.
714 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
715 * varasm.c (make_decl_one_only): Use set_comdat_group;
716 create node if needed.
717 * ipa-inline-transform.c (save_inline_function_body): Update
718 way we decl->symtab mapping.
719 * symtab.c (symtab_hash, hash_node, eq_node
720 symtab_insert_node_to_hashtable): Remove.
721 (symtab_register_node): Update.
722 (symtab_unregister_node): Update.
723 (symtab_get_node): Reimplement as inline function.
724 (symtab_add_to_same_comdat_group): Update.
725 (symtab_dissolve_same_comdat_group_list): Update.
726 (dump_symtab_base): Update.
727 (verify_symtab_base): Update.
728 (symtab_make_decl_local): Update.
729 (fixup_same_cpp_alias_visibility): Update.
730 (symtab_nonoverwritable_alias): Update.
731 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
732 * ipa.c (update_visibility_by_resolution_info): UPdate.
733 * bb-reorder.c: Include cgraph.h
734 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
736 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
737 * cgraph.c (cgraph_get_create_node): Update.
738 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
740 (symtab_get_node): Make inline.
741 (symtab_insert_node_to_hashtable): Remove.
742 (symtab_can_be_discarded): Update.
743 (decl_comdat_group): New function.
744 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
746 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
748 (read_comdat_group): New function.
749 (input_node, input_varpool_node): Use it.
750 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
752 * mips.c (mips_start_unique_function): Likewise.
753 (ix86_code_end): Likewise.
754 (rs6000_code_end): Likweise.
755 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
757 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
759 * gengtype-state.c (fatal_reading_state): Bring offline.
760 * optabs.c (widening_optab_handler): Bring offline.
761 * optabs.h (widening_optab_handler): Likewise.
762 * final.c (get_attr_length_1): Likewise.
764 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
766 * sched-int.h (sd_iterator_cond): Manually tail recurse.
768 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
770 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
771 (ppc440-compare): Include shift with dot.
772 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
773 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
775 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
777 (e6500_sfx2): Include it.
778 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
779 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
780 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
781 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
782 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
783 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
784 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
785 *lshiftrt_internal1le, *lshiftrt_internal1be,
786 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
787 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
788 *rotldi3_internal10le, *rotldi3_internal10be,
789 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
790 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
791 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
792 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
793 define_insns): Use type "shift" in the appropriate alternatives.
795 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
797 * config/rs6000/rs6000.md (type): Add "logical". Delete
799 (dot): Adjust comment.
800 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
801 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
802 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
803 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
804 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
805 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
806 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
807 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
809 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
810 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
811 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
812 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
813 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
814 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
815 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
816 * config/rs6000/8540.md (ppc8540_su): Adjust.
817 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
818 cell-cmp-microcoded): Adjust.
819 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
820 * config/rs6000/e500mc.md (e500mc_su): Adjust.
821 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
822 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
823 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
824 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
825 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
826 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
827 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
829 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
830 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
831 Adjust. Adjust comment.
832 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
833 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
835 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
837 * config/rs6000/rs6000.md (type): Add "add".
838 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
839 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
840 define_insns): Use it.
841 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
843 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
844 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
845 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
846 * config/rs6000/601.md (ppc601-integer): Adjust.
847 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
848 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
849 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
850 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
851 * config/rs6000/8540.md (ppc8540_su): Adjust.
852 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
853 cell-cmp-microcoded): Adjust.
854 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
855 * config/rs6000/e500mc.md (e500mc_su): Adjust.
856 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
857 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
858 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
859 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
860 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
861 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
862 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
864 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
865 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
866 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
867 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
869 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
871 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
872 "delayed_compare", "var_delayed_compare".
873 (var_shift): New attribute.
874 (cell_micro): Adjust.
875 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
876 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
877 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
878 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
879 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
880 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
881 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
882 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
883 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
884 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
885 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
886 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
887 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
888 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
889 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
890 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
891 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
892 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
893 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
894 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
895 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
896 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
897 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
898 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
899 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
901 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
902 * config/rs6000/440.md (ppc440-integer): Adjust.
903 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
904 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
905 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
906 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
907 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
908 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
909 * config/rs6000/8540.md (ppc8540_su): Adjust.
910 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
911 cell-cmp-microcoded): Adjust.
912 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
913 * config/rs6000/e500mc.md (e500mc_su): Adjust.
914 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
915 e500mc64_delayed): Adjust.
916 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
917 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
918 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
919 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
920 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
921 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
922 power6-delayed-compare, power6-var-delayed-compare): Adjust.
923 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
924 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
926 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
927 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
929 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
931 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
932 (bits): New mode_attr.
933 (idiv_ldiv): Delete mode_attr.
934 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
935 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
936 rs6000_adjust_priority, is_nonpipeline_insn,
937 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
939 * config/rs6000/40x.md (ppc403-idiv): Adjust.
940 * config/rs6000/440.md (ppc440-idiv): Adjust.
941 * config/rs6000/476.md (ppc476-idiv): Adjust.
942 * config/rs6000/601.md (ppc601-idiv): Adjust.
943 * config/rs6000/603.md (ppc603-idiv): Adjust.
944 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
945 ppc620-ldiv): Adjust.
946 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
947 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
948 * config/rs6000/8540.md (ppc8540_divide): Adjust.
949 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
950 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
951 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
952 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
953 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
954 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
955 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
956 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
957 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
958 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
959 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
960 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
961 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
962 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
963 * config/rs6000/titan.md (titan_fxu_div): Adjust.
965 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
967 * config/rs6000/rs6000.md (type): Delete "insert_word",
968 "insert_dword". Add "insert".
969 (size): Update comment.
970 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
971 insn_must_be_first_in_group): Adjust.
972 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
973 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
974 *insvsi_internal6, insvdi_internal): Adjust.
976 * config/rs6000/40x.md (ppc403-integer): Adjust.
977 * config/rs6000/440.md (ppc440-integer): Adjust.
978 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
979 * config/rs6000/601.md (ppc601-integer): Adjust.
980 * config/rs6000/603.md (ppc603-integer): Adjust.
981 * config/rs6000/6xx.md (ppc604-integer): Adjust.
982 * config/rs6000/7450.md (ppc7450-integer): Adjust.
983 * config/rs6000/7xx.md (ppc750-integer): Adjust.
984 * config/rs6000/8540.md (ppc8540_su): Adjust.
985 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
986 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
987 * config/rs6000/e500mc.md (e500mc_su): Adjust.
988 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
989 * config/rs6000/e5500.md (e5500_sfx): Adjust.
990 * config/rs6000/e6500.md (e6500_sfx): Adjust.
991 * config/rs6000/mpc.md (mpccore-integer): Adjust.
992 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
993 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
994 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
995 * config/rs6000/power7.md (power7-integer): Adjust.
996 * config/rs6000/power8.md (power8-1cyc): Adjust.
997 * config/rs6000/rs64.md (rs64a-integer): Adjust.
998 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1000 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1002 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
1003 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
1004 (size): New attribute.
1005 (dot): New attribute.
1006 (cell_micro): Adjust.
1007 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
1008 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
1009 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
1010 umuldi3_highpart): Adjust.
1011 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1012 rs6000_adjust_priority, is_nonpipeline_insn,
1013 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1015 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
1016 ppc405-imul3): Adjust.
1017 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
1018 * config/rs6000/476.md (ppc476-imul): Adjust.
1019 * config/rs6000/601.md (ppc601-imul): Adjust.
1020 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
1021 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
1022 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
1023 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
1024 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
1026 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
1027 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
1028 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
1030 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
1031 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
1032 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
1033 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
1034 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
1035 * config/rs6000/mpc.md (mpccore-imul): Adjust.
1036 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
1037 power4-lmul, power4-imul, power4-imul3): Adjust.
1038 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
1039 power5-lmul, power5-imul, power5-imul3): Adjust.
1040 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
1041 power6-lmul, power6-imul, power6-imul3): Adjust.
1042 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
1043 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
1045 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
1046 rs64a-lmul): Adjust.
1047 * config/rs6000/titan.md (titan_imul): Adjust.
1049 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1051 * config/rs6000/rs6000.md (type): Add new value "halfmul".
1052 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
1053 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
1054 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
1055 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
1056 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
1057 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
1058 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
1059 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
1060 * config/rs6000/titan.md: Delete nonsensical comment.
1061 (titan_imul): Add type imul3.
1062 (titan_mulhw): Remove type imul3; add type halfmul.
1064 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1066 * config/rs6000/rs6000.md (type): Reorder, reformat.
1068 2014-05-23 Martin Jambor <mjambor@suse.cz>
1070 PR tree-optimization/53787
1071 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
1072 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
1073 analysis_done, update all uses.
1074 * ipa-prop.c: Include domwalk.h
1075 (param_analysis_info): Removed.
1076 (param_aa_status): New type.
1077 (ipa_bb_info): Likewise.
1078 (func_body_info): Likewise.
1079 (ipa_get_bb_info): New function.
1080 (aa_overwalked): Likewise.
1081 (find_dominating_aa_status): Likewise.
1082 (parm_bb_aa_status_for_bb): Likewise.
1083 (parm_preserved_before_stmt_p): Changed to use new param AA info.
1084 (load_from_unmodified_param): Accept func_body_info as a parameter
1085 instead of parms_ainfo.
1086 (parm_ref_data_preserved_p): Changed to use new param AA info.
1087 (parm_ref_data_pass_through_p): Likewise.
1088 (ipa_load_from_parm_agg_1): Likewise. Update callers.
1089 (compute_complex_assign_jump_func): Changed to use new param AA info.
1090 (compute_complex_ancestor_jump_func): Likewise.
1091 (ipa_compute_jump_functions_for_edge): Likewise.
1092 (ipa_compute_jump_functions): Removed.
1093 (ipa_compute_jump_functions_for_bb): New function.
1094 (ipa_analyze_indirect_call_uses): Likewise, moved variable
1096 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
1097 and info, moved variable declarations down.
1098 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
1100 (ipa_analyze_stmt_uses): Likewise.
1101 (ipa_analyze_params_uses): Removed.
1102 (ipa_analyze_params_uses_in_bb): New function.
1103 (ipa_analyze_controlled_uses): Likewise.
1104 (free_ipa_bb_info): Likewise.
1105 (analysis_dom_walker): New class.
1106 (ipa_analyze_node): Handle node-specific forbidden analysis,
1107 initialize and free func_body_info, use dominator walker.
1108 (ipcp_modif_dom_walker): New class.
1109 (ipcp_transform_function): Create and free func_body_info, use
1110 ipcp_modif_dom_walker, moved a lot of functionality there.
1112 2014-05-23 Marek Polacek <polacek@redhat.com>
1113 Jakub Jelinek <jakub@redhat.com>
1115 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
1116 * gcc.c (sanitize_spec_function): Likewise.
1117 * convert.c (convert_to_integer): Include "ubsan.h". Add
1118 floating-point to integer instrumentation.
1119 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
1120 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
1121 SANITIZE_NONDEFAULT.
1122 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
1123 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
1124 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
1125 * ubsan.c: Include "realmpfr.h" and "dfp.h".
1126 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
1127 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
1128 float/double/long double.
1129 (ubsan_instrument_float_cast): New function.
1130 * ubsan.h (ubsan_instrument_float_cast): Declare.
1132 2014-05-23 Jiong Wang <jiong.wang@arm.com>
1134 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
1136 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
1137 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
1138 Adjust for tailcalling through registers.
1139 * config/aarch64/aarch64.h (enum reg_class): New caller save
1141 (REG_CLASS_NAMES): Likewise.
1142 (REG_CLASS_CONTENTS): Likewise.
1143 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
1144 Allow tailcalling without decls.
1146 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1148 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1149 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
1151 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
1152 gsi, and variables v_* to v*.
1154 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
1156 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
1158 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1160 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
1161 * omp-low.c: Update accordingly.
1163 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
1164 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
1165 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
1166 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
1167 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
1168 GF_OMP_TARGET_KIND_UPDATE.
1170 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1171 Explicitly enumerate the expected region types.
1173 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
1176 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
1177 documentation; the old documentation didn't clearly state the
1178 constraints on the contents of the pointed-to storage.
1180 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1182 Fix bootstrap error on ia64
1183 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
1184 Return default value.
1186 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
1188 PR tree-optimization/54733
1189 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
1191 (find_bswap_or_nop_load): New.
1192 (find_bswap_1): Renamed to ...
1193 (find_bswap_or_nop_1): This. Also add support for memory source.
1194 (find_bswap): Renamed to ...
1195 (find_bswap_or_nop): This. Also add support for memory source and
1196 detection of bitwise operations equivalent to load in host endianness.
1197 (execute_optimize_bswap): Likewise. Also move its leading comment back
1198 in place and split statement transformation into ...
1199 (bswap_replace): This.
1201 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1203 PR rtl-optimization/61215
1204 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
1205 simplify_gen_subreg until final substitution.
1207 2014-05-23 Alan Modra <amodra@gmail.com>
1210 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
1211 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
1212 Use "Y" constraint rather than "m".
1214 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1216 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
1218 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
1219 New function declaration.
1220 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
1221 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
1222 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
1223 (aarch64_init_builtins) : Initialize builtins
1224 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1225 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1226 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
1227 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
1228 and __builtins_aarch64_set_fpsr.
1229 (aarch64_atomic_assign_expand_fenv): New function.
1230 * config/aarch64/aarch64.md (set_fpcr): New pattern.
1231 (get_fpcr) : Likewise.
1232 (set_fpsr) : Likewise.
1233 (get_fpsr) : Likewise.
1234 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
1235 and UNSPECV_SET_FPSR.
1236 * doc/extend.texi (AARCH64 Built-in Functions) : Document
1237 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1238 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1240 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1242 PR rtl-optimization/60969
1243 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
1244 constraints. Set up mem cost for NO_REGS case.
1246 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
1248 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
1250 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
1252 * config/darwin.c: Include "lto-section-names.h".
1253 (LTO_SEGMENT_NAME): Don't define.
1254 * config/i386/winnt.c: Include "lto-section-names.h".
1255 * lto-streamer.c: Include "lto-section-names.h".
1256 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
1257 * lto-wrapper.c: Include "lto-section-names.h".
1258 (LTO_SECTION_NAME_PREFIX): Don't define.
1259 * lto-section-names.h: New file.
1260 * cgraphunit.c: Include "lto-section-names.h".
1262 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
1264 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
1266 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
1269 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
1271 2014-05-22 Nick Clifton <nickc@redhat.com>
1273 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
1275 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
1277 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
1278 -> (T)A transformation to integer types.
1280 2014-05-22 Teresa Johnson <tejohnson@google.com>
1282 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
1283 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
1284 (gcov_rewrite): Use gcov_nonruntime_assert.
1286 (gcov_write_words): Ditto.
1287 (gcov_write_length): Ditto.
1288 (gcov_read_words): Use gcov_nonruntime_assert, and remove
1289 gcc_assert from IN_LIBGCOV code.
1290 (gcov_read_summary): Use gcov_error to flag profile corruption.
1291 (gcov_sync): Use gcov_nonruntime_assert.
1292 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
1293 (gcov_histo_index): Use gcov_nonruntime_assert.
1294 (static void gcov_histogram_merge): Ditto.
1295 (compute_working_sets): Ditto.
1296 * gcov-io.h (gcov_nonruntime_assert): Define.
1297 (gcov_error): Define for !IN_LIBGCOV
1299 2014-05-22 Richard Biener <rguenther@suse.de>
1301 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1302 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
1303 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
1304 and deallocation site.
1305 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1306 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
1307 passing through the incoming points-to set.
1308 (handle_lhs_call): Use flags argument instead of recomputing it.
1309 (find_func_aliases_for_call): Call handle_lhs_call with proper
1312 2014-05-22 Jakub Jelinek <jakub@redhat.com>
1314 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
1315 all padding bits in REAL_VALUE_TYPE are cleared.
1317 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1319 Cleanup and improve multipass_dfa_lookahead_guard
1320 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
1321 (core2i7_first_cycle_multipass_begin,)
1322 (core2i7_first_cycle_multipass_issue,)
1323 (core2i7_first_cycle_multipass_backtrack): Update signature.
1324 * config/ia64/ia64.c
1325 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
1326 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
1327 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
1329 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
1330 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
1332 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
1334 * doc/tm.texi: Regenerate.
1336 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
1337 * haifa-sched.c (ready_try): Make signed to allow negative values.
1338 (rebug_ready_list_1): Update.
1339 (choose_ready): Simplify.
1340 (sched_extend_ready_list): Update.
1342 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1344 Remove IA64 speculation tweaking flags
1345 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
1346 speculation tuning flags.
1347 (msched-prefer-non-data-spec-insns,)
1348 (msched-prefer-non-control-spec-insns): Obsolete options.
1349 * haifa-sched.c (choose_ready): Remove handling of
1350 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1351 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
1352 and PREFER_NON_DATA_SPEC.
1353 * sel-sched.c (process_spec_exprs): Remove handling of
1354 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1356 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1358 Improve scheduling debug output
1359 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
1360 (advance_one_cycle): Update.
1361 (schedule_insn, queue_to_ready): Add debug printouts.
1362 (debug_ready_list_1): New static function.
1363 (debug_ready_list): Update.
1364 (max_issue): Add debug printouts.
1365 (dump_insn_stream): New static function.
1366 (schedule_block): Use it. Also better indent printouts.
1368 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1370 Fix sched_insn debug counter
1371 * haifa-sched.c (schedule_insn): Update.
1372 (struct haifa_saved_data): Add nonscheduled_insns_begin.
1373 (save_backtrack_point, restore_backtrack_point): Update.
1374 (first_nonscheduled_insn): New static function.
1375 (queue_to_ready, choose_ready): Use it.
1376 (schedule_block): Init nonscheduled_insns_begin.
1377 (sched_emit_insn): Update.
1380 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
1382 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
1384 (aarch64_secondary_reload) : LikeWise.
1385 (aarch64_class_max_nregs) : Remove CORE_REGS.
1386 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
1387 (REG_CLASS_NAMES) : Likewise.
1388 (REG_CLASS_CONTENTS) : LikeWise.
1389 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
1391 2014-05-21 Guozhi Wei <carrot@google.com>
1394 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
1396 (vqdmulhq_n_s16): Likewise.
1398 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
1400 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
1402 2014-05-21 Marek Polacek <polacek@redhat.com>
1405 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
1407 2014-05-21 Martin Jambor <mjambor@suse.cz>
1409 * doc/invoke.texi (Optimize Options): Document parameters
1410 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
1411 ipa-cp-array-index-hint-bonus.
1413 2014-05-21 Mark Wielaard <mjw@redhat.com>
1416 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
1417 version >= 3 or not strict DWARF.
1418 * langhooks.h (struct lang_hooks_for_types): Add
1419 enum_underlying_base_type.
1420 * langhooks.c (lhd_enum_underlying_base_type): New function.
1421 * gcc/langhooks.h (struct lang_hooks_for_types): Add
1422 enum_underlying_base_type.
1423 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
1424 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
1425 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
1427 2014-05-21 Richard Biener <rguenther@suse.de>
1429 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
1431 2014-05-21 John Marino <gnugcc@marino.st>
1433 * config.gcc (*-*-dragonfly*): New target.
1434 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
1435 * configure: Regenerate.
1436 * config/dragonfly-stdint.h: New.
1437 * config/dragonfly.h: New.
1438 * config/dragonfly.opt: New.
1439 * config/i386/dragonfly.h: New.
1440 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
1442 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1444 * tree.def (VOID_CST): New.
1445 * tree-core.h (TI_VOID): New.
1446 * tree.h (void_node): New.
1447 * tree.c (tree_node_structure_for_code, tree_code_size)
1448 (iterative_hash_expr): Handle VOID_CST.
1449 (build_common_tree_nodes): Initialize void_node.
1451 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
1453 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
1455 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
1457 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
1460 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
1461 flag_reorder_blocks_and_partition.
1462 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
1464 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
1467 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
1469 (*addc_r_t): Add new insn_and_split.
1471 2014-05-21 Jakub Jelinek <jakub@redhat.com>
1474 * omp-low.c (handle_simd_reference): New function.
1475 (lower_rec_input_clauses): Use it. Defer adding reference
1476 initialization even for reduction without placeholder if in simd,
1477 handle it properly later on.
1479 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1481 PR tree-optimization/60899
1482 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
1483 assume all static symbols will have definition wile parsing and
1484 check the do have definition later in compilation; check that
1485 variable referring symbol will be output before concluding that
1486 reference is safe; be conservative for referring local statics;
1487 be more precise about when comdat is output in other partition.
1489 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1492 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
1494 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
1495 (ipa_inline): Loop inline_to_all_callers until no more aliases
1498 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1500 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
1501 set writeonly flag only for vars actually written to.
1503 2014-05-20 Dehao Chen <dehao@google.com>
1505 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
1506 and callee count to get clone count.
1507 * tree-inline.c (expand_call_inline): Use callee count instead of bb
1510 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
1512 PR rtl-optimization/61243
1513 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
1515 2014-05-20 Xinliang David Li <davidxl@google.com>
1517 * cgraphunit.c (walk_polymorphic_call_targets): Add
1518 dbgcnt and fopt-info support.
1519 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1520 * ipa-devirt.c (ipa_devirt): Ditto.
1521 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1522 * ipa.c (walk_polymorphic_call_targets): Ditto.
1523 * gimple-fold.c (fold_gimple_assign): Ditto.
1524 (gimple_fold_call): Ditto.
1525 * dbgcnt.def: New counter.
1527 2014-05-20 DJ Delorie <dj@redhat.com>
1529 * config/msp430/msp430.md (split): Don't allow subregs when
1530 splitting SImode adds.
1531 (andneghi): Fix subtraction logic.
1532 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1534 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1536 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1538 * except.c (switch_to_exception_section, resolve_unique_section,
1539 get_named_text_section, default_function_rodata_section,
1540 align_variable, get_block_for_decl, default_section_type_flags):
1541 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1542 * symtab.c (symtab_add_to_same_comdat_group,
1543 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1544 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1546 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1547 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1548 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1549 (c6x_function_in_section_p): Likewise.
1550 * config/darwin.c (machopic_select_section): Likewise.
1551 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1552 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1553 * config/mep/mep.c (mep_select_section): LIkewise.
1554 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1556 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1558 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1559 EH region of calls to pure functions that can throw an exception.
1560 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1561 (copy_reference_ops_from_call): Also copy the EH region of the call if
1562 it can throw an exception.
1564 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1566 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1567 nested VEC_SELECTs that are inverses of each other.
1569 2014-05-20 Richard Biener <rguenther@suse.de>
1571 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1572 (extract_and_process_scc_for_name): not here.
1573 (cond_dom_walker::before_dom_children): Only process
1574 stmts that end the BB in interesting ways.
1575 (run_scc_vn): Mark param uses as visited.
1577 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1579 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1582 2014-05-20 Nick Clifton <nickc@redhat.com>
1584 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1585 (msp430_gimplify_va_arg_expr): New function.
1586 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
1588 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
1589 operand 0 in order to prevent confusion about the number of
1592 2014-05-20 Richard Biener <rguenther@suse.de>
1594 PR tree-optimization/61221
1595 * tree-ssa-pre.c (el_to_update): Remove.
1596 (eliminate_dom_walker::before_dom_children): Handle released
1597 VDEFs by value-numbering them to the associated VUSE. Update
1598 stmt immediately for substituted call address.
1599 (eliminate): Remove delayed stmt updating code.
1600 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
1601 possibly late re-numbered vuses.
1602 (vn_reference_lookup_2): Adjust.
1603 (vn_reference_lookup_pieces): Likewise.
1604 (vn_reference_lookup): Likewise.
1606 2014-05-20 Richard Biener <rguenther@suse.de>
1608 * config.gcc: Remove need_64bit_hwint.
1609 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
1610 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
1612 * config.in: Regenerate.
1613 * configure: Likewise.
1615 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
1617 * doc/extend.texi: Create Label Attributes section,
1618 move all label attributes into it and reference it.
1620 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
1622 * arm.c (thumb1_reorg): When scanning backwards skip anything
1623 that's not a proper insn.
1625 2014-05-19 Richard Biener <rguenther@suse.de>
1627 PR tree-optimization/61221
1628 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
1629 Do nothing for unreachable blocks.
1630 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
1631 Improve unreachability detection.
1633 2014-05-19 Richard Biener <rguenther@suse.de>
1635 PR tree-optimization/61209
1636 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
1638 2014-05-19 Nick Clifton <nickc@redhat.com>
1640 * except.c (init_eh): Fix computation of builtin setjmp buffer
1641 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
1643 2014-05-19 Richard Biener <rguenther@suse.de>
1645 PR tree-optimization/61184
1646 * tree-vrp.c (is_negative_overflow_infinity): Use
1647 TREE_OVERFLOW_P and do that check first.
1648 (is_positive_overflow_infinity): Likewise.
1649 (is_overflow_infinity): Likewise.
1650 (vrp_operand_equal_p): Properly treat operands with
1651 differing overflow as not equal.
1653 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
1655 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
1656 shift simplification where it was intended.
1658 2014-05-19 Christian Bruel <christian.bruel@st.com>
1661 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
1663 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
1666 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
1669 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1671 * reg-notes.def (CROSSING_JUMP): Likewise.
1672 * rtl.h (rtx_def): Update comment for jump flag.
1673 (CROSSING_JUMP_P): Define.
1674 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
1675 of a REG_CROSSING_JUMP note.
1676 * cfghooks.c (tidy_fallthru_edges): Likewise.
1677 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
1678 * emit-rtl.c (try_split): Likewise.
1679 * haifa-sched.c (sched_create_recovery_edges): Likewise.
1680 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
1681 * jump.c (redirect_jump_2): Likewise.
1682 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
1683 (relax_delay_slots): Likewise.
1684 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
1685 (bbit_di): Likewise.
1686 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
1687 * config/sh/sh.md (jump_compact): Likewise.
1688 * bb-reorder.c (rotate_loop): Likewise.
1689 (pass_duplicate_computed_gotos::execute): Likewise.
1690 (add_reg_crossing_jump_notes): Rename to...
1691 (update_crossing_jump_flags): ...this.
1692 (pass_partition_blocks::execute): Update accordingly.
1694 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1696 * tree.h: Remove extraneous template <>.
1698 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1700 * ipa.c (symtab_remove_unreachable_nodes): Remove
1701 symbol from comdat group if its body was eliminated.
1702 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
1703 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
1704 (symtab_unregister_node): ... this one.
1705 (verify_symtab_base): More strict checking of comdats.
1706 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
1708 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1710 * tree-pass.h (make_pass_ipa_comdats): New pass.
1711 * timevar.def (TV_IPA_COMDATS): New timevar.
1712 * passes.def (pass_ipa_comdats): Add.
1713 * Makefile.in (OBJS): Add ipa-comdats.o
1714 * ipa-comdats.c: New file.
1716 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1718 * ipa.c (update_visibility_by_resolution_info): New function.
1719 (function_and_variable_visibility): Use it.
1721 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1723 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
1725 (FOR_EACH_DEFINED_SYMBOL): New macro.
1726 (varpool_first_static_initializer, varpool_next_static_initializer,
1727 varpool_first_defined_variable, varpool_next_defined_variable):
1729 (symtab_in_same_comdat_p): Correctly deal with inline functions.
1731 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1733 * ggc-page.c (ggc_handle_finalizers): Add comment.
1735 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1737 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
1738 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
1739 (ggc_internal_cleared_alloc): Likewise.
1740 * ggc-page.c (finalizer): New class.
1741 (vec_finalizer): Likewise.
1742 (globals::finalizers): New member.
1743 (globals::vec_finalizers): Likewise.
1744 (ggc_internal_alloc): Record the finalizer if any for the block being
1746 (ggc_handle_finalizers): New function.
1747 (ggc_collect): Call ggc_handle_finalizers.
1748 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
1750 (ggc_internal_cleared_alloc): Likewise.
1751 (finalize): New function.
1752 (need_finalization_p): Likewise.
1753 (ggc_alloc): Install the type's destructor as the finalizer if it
1755 (ggc_cleared_alloc): Likewise.
1756 (ggc_vec_alloc): Likewise.
1757 (ggc_cleared_vec_alloc): Likewise.
1759 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1761 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
1763 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1765 * alias.c (record_alias_subset): Adjust.
1766 * bitmap.c (bitmap_element_allocate): Likewise.
1767 (bitmap_gc_alloc_stat): Likewise.
1768 * cfg.c (init_flow): Likewise.
1769 (alloc_block): Likewise.
1770 (unchecked_make_edge): Likewise.
1771 * cfgloop.c (alloc_loop): Likewise.
1772 (flow_loops_find): Likewise.
1773 (rescan_loop_exit): Likewise.
1774 * cfgrtl.c (init_rtl_bb_info): Likewise.
1775 * cgraph.c (insert_new_cgraph_node_version): Likewise.
1776 (cgraph_allocate_node): Likewise.
1777 (cgraph_create_edge_1): Likewise.
1778 (cgraph_allocate_init_indirect_info): Likewise.
1779 * cgraphclones.c (cgraph_clone_edge): Likewise.
1780 * cgraphunit.c (add_asm_node): Likewise.
1781 (init_lowered_empty_function): Likewise.
1782 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
1783 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
1784 (alpha_use_linkage): Likewise.
1785 * config/arc/arc.c (arc_init_machine_status): Likewise.
1786 * config/arm/arm.c (arm_init_machine_status): Likewise.
1787 * config/avr/avr.c (avr_init_machine_status): Likewise.
1788 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
1789 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
1790 * config/cris/cris.c (cris_init_machine_status): Likewise.
1791 * config/darwin.c (machopic_indirection_name): Likewise.
1792 (darwin_build_constant_cfstring): Likewise.
1793 (darwin_enter_string_into_cfstring_table): Likewise.
1794 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
1795 * config/frv/frv.c (frv_init_machine_status): Likewise.
1796 * config/i386/i386.c (get_dllimport_decl): Likewise.
1797 (ix86_init_machine_status): Likewise.
1798 (assign_386_stack_local): Likewise.
1799 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
1800 (i386_pe_maybe_record_exported_symbol): Likewise.
1801 (i386_pe_record_stub): Likewise.
1802 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
1803 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
1804 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
1805 (m32c_note_pragma_address): Likewise.
1806 * config/mep/mep.c (mep_init_machine_status): Likewise.
1807 (mep_note_pragma_flag): Likewise.
1808 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
1809 (mips16_local_alias): Likewise.
1810 (mips_init_machine_status): Likewise.
1811 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
1812 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
1813 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
1814 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
1815 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
1816 * config/pa/pa.c (pa_init_machine_status): Likewise.
1817 (pa_get_deferred_plabel): Likewise.
1818 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
1819 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
1820 (rs6000_init_machine_status): Likewise.
1821 (output_toc): Likewise.
1822 * config/s390/s390.c (s390_init_machine_status): Likewise.
1823 * config/score/score.c (score_output_external): Likewise.
1824 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
1825 * config/spu/spu.c (spu_init_machine_status): Likewise.
1826 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
1827 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
1828 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
1829 * coverage.c (coverage_end_function): Likewise.
1830 * dbxout.c (dbxout_init): Likewise.
1831 * doc/gty.texi: Don't mention variable_size attribute.
1832 * dwarf2cfi.c (new_cfi): Adjust.
1833 (new_cfi_row): Likewise.
1834 (copy_cfi_row): Likewise.
1835 (create_cie_data): Likewise.
1836 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
1837 (new_loc_descr): Likewise.
1838 (find_AT_string_in_table): Likewise.
1839 (add_addr_table_entry): Likewise.
1840 (new_die): Likewise.
1841 (add_var_loc_to_decl): Likewise.
1842 (clone_die): Likewise.
1843 (clone_as_declaration): Likewise.
1844 (break_out_comdat_types): Likewise.
1845 (new_loc_list): Likewise.
1846 (add_loc_descr_to_each): Likewise.
1847 (add_location_or_const_value_attribute): Likewise.
1848 (add_linkage_name): Likewise.
1849 (lookup_filename): Likewise.
1850 (dwarf2out_var_location): Likewise.
1851 (new_line_info_table): Likewise.
1852 (dwarf2out_init): Likewise.
1853 (mem_loc_descriptor): Likewise.
1854 (loc_descriptor): Likewise.
1855 (add_const_value_attribute): Likewise.
1856 (tree_add_const_value_attribute): Likewise.
1857 (comp_dir_string): Likewise.
1858 (dwarf2out_vms_debug_main_pointer): Likewise.
1859 (string_cst_pool_decl): Likewise.
1860 * emit-rtl.c (set_mem_attrs): Likewise.
1861 (get_reg_attrs): Likewise.
1862 (start_sequence): Likewise.
1863 (init_emit): Likewise.
1864 (init_emit_regs): Likewise.
1865 * except.c (init_eh_for_function): Likewise.
1866 (gen_eh_region): Likewise.
1867 (gen_eh_region_catch): Likewise.
1868 (gen_eh_landing_pad): Likewise.
1869 (add_call_site): Likewise.
1870 * function.c (add_frame_space): Likewise.
1871 (insert_temp_slot_address): Likewise.
1872 (assign_stack_temp_for_type): Likewise.
1873 (get_hard_reg_initial_val): Likewise.
1874 (allocate_struct_function): Likewise.
1875 (prepare_function_start): Likewise.
1876 (types_used_by_var_decl_insert): Likewise.
1877 * gengtype.c (variable_size_p): Remove function.
1878 (enum alloc_quantity): Remove enum.
1879 (write_typed_alloc_def): Remove function.
1880 (write_typed_struct_alloc_def): Likewise.
1881 (write_typed_typedef_alloc_def): Likewise.
1882 (write_typed_alloc_defns): Likewise.
1884 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
1885 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
1886 * ggc.h (ggc_alloc): new function.
1887 (ggc_cleared_alloc): Likewise.
1888 (ggc_vec_alloc): Template on type of vector element, and remove
1889 element size argument.
1890 (ggc_cleared_vec_alloc): Likewise.
1891 * gimple.c (gimple_build_omp_for): Adjust.
1892 (gimple_copy): Likewise.
1893 * ipa-cp.c (get_replacement_map): Likewise.
1894 (find_aggregate_values_for_callers_subset): Likewise.
1895 (known_aggs_to_agg_replacement_list): Likewise.
1896 * ipa-devirt.c (get_odr_type): Likewise.
1897 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
1898 (read_agg_replacement_chain): Likewise.
1899 * loop-iv.c (get_simple_loop_desc): Likewise.
1900 * lto-cgraph.c (input_node_opt_summary): Likewise.
1901 * lto-section-in.c (lto_new_in_decl_state): Likewise.
1902 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
1903 (input_eh_region): Likewise.
1904 (input_eh_lp): Likewise.
1905 (input_cfg): Likewise.
1906 * optabs.c (set_optab_libfunc): Likewise.
1907 (init_tree_optimization_optabs): Likewise.
1908 (set_conv_libfunc): Likewise.
1909 * passes.c (do_per_function_toporder): Likewise.
1910 * rtl.h: Don't use variable_size gty attribute.
1911 * sese.c (if_region_set_false_region): Adjust.
1912 * stringpool.c (gt_pch_save_stringpool): Likewise.
1913 * target-globals.c (save_target_globals): Likewise.
1914 * toplev.c (general_init): Likewise.
1915 * trans-mem.c (record_tm_replacement): Likewise.
1916 (split_bb_make_tm_edge): Likewise.
1917 * tree-cfg.c (move_sese_region_to_fn): Likewise.
1918 * tree-data-ref.h (lambda_vector_new): Likewise.
1919 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
1920 * tree-iterator.c (tsi_link_before): Likewise.
1921 (tsi_link_after): Likewise.
1922 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
1923 * tree-ssa-loop-niter.c (record_estimate): Likewise.
1924 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
1925 * tree-ssa-operands.h: Don't use variable_size gty attribute.
1926 * tree-ssa.c (init_tree_ssa): Adjust.
1927 * tree-ssanames.c (set_range_info): Likewise.
1928 (get_ptr_info): Likewise.
1929 (duplicate_ssa_name_ptr_info): Likewise.
1930 (duplicate_ssa_name_range_info): Likewise.
1931 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
1932 (unpack_ts_fixed_cst_value_fields): Likewise.
1933 * tree.c (build_fixed): Likewise.
1934 (build_real): Likewise.
1935 (build_string): Likewise.
1936 (decl_priority_info): Likewise.
1937 (decl_debug_expr_insert): Likewise.
1938 (decl_value_expr_insert): Likewise.
1939 (decl_debug_args_insert): Likewise.
1940 (type_hash_add): Likewise.
1941 (build_omp_clause): Likewise.
1942 * ubsan.c (decl_for_type_insert): Likewise.
1943 * varasm.c (get_unnamed_section): Likewise.
1944 (get_noswitch_section): Likewise.
1945 (get_section): Likewise.
1946 (get_block_for_section): Likewise.
1947 (create_block_symbol): Likewise.
1948 (build_constant_desc): Likewise.
1949 (create_constant_pool): Likewise.
1950 (force_const_mem): Likewise.
1951 (record_tm_clone_pair): Likewise.
1952 * varpool.c (varpool_create_empty_node): Likewise.
1954 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1956 * dwarf2out.c (tree_add_const_value_attribute): Call
1957 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
1958 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
1959 instead of ggc_internal_<x>alloc_stat.
1960 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
1961 (ggc_realloc): Likewise.
1962 * ggc-none.c (ggc_internal_alloc): Likewise.
1963 (ggc_internal_cleared_alloc): Likewise.
1964 * ggc-page.c: Likewise.
1965 * ggc.h (ggc_internal_alloc_stat): Likewise.
1966 (ggc_internal_alloc): Remove macro.
1967 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
1968 (ggc_internal_cleared_alloc): Remove macro.
1969 (GGC_RESIZEVEC): Adjust.
1970 (ggc_resizevar): Remove macro.
1971 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
1972 (ggc_internal_cleared_vec_alloc_stat): Likewise.
1973 (ggc_internal_vec_cleared_alloc): Remove macro.
1974 (ggc_alloc_atomic_stat): Drop _stat suffix.
1975 (ggc_alloc_atomic): Remove macro.
1976 (ggc_alloc_cleared_atomic): Remove macro.
1977 (ggc_alloc_string_stat): Drop _stat suffix.
1978 (ggc_alloc_string): Remove macro.
1979 (ggc_alloc_rtx_def_stat): Adjust.
1980 (ggc_alloc_tree_node_stat): Likewise.
1981 (ggc_alloc_cleared_tree_node_stat): Likewise.
1982 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
1983 (ggc_alloc_cleared_simd_clone_stat): Likewise.
1984 * gimple.c (gimple_build_omp_for): Likewise.
1985 (gimple_copy): Likewise.
1986 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
1987 * toplev.c (realloc_for_line_map): Adjust.
1988 * tree-data-ref.h (lambda_vector_new): Likewise.
1989 * tree-phinodes.c (allocate_phi_node): Likewise.
1990 * tree.c (grow_tree_vec_stat): Likewise.
1991 * vec.h (va_gc::reserve): Adjust.
1993 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
1995 * config/microblaze/microblaze.c (break_handler): New Declaration.
1996 (microblaze_break_function_p,microblaze_is_break_handler): New.
1997 (compute_frame_size): Use microblaze_break_function_p.
1998 Add the test of break_handler.
1999 (microblaze_function_prologue) : Add the test of variable
2000 break_handler. Check the fnname by BREAK_HANDLER_NAME.
2001 (microblaze_function_epilogue) : Add the test of break_handler.
2002 (microblaze_globalize_label) : Add the test of break_handler.
2003 Check the name by BREAK_HANDLER_NAME.
2005 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
2007 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
2008 microblaze_is_break_handler test.
2009 (call_internal1,call_value_intern): Use microblaze_break_function_p.
2010 Use SYMBOL_REF_DECL.
2012 * config/microblaze/microblaze-protos.h
2013 (microblaze_break_function_p,microblaze_is_break_handler):
2016 * doc/extend.texi (MicroBlaze break_handler Functions): Document
2017 new MicroBlaze break_handler functions.
2019 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2021 * doc/extend.texi (Size of an asm): Move node text according
2022 to its @menu entry position.
2024 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
2026 PR tree-optimization/61140
2027 PR tree-optimization/61150
2028 PR tree-optimization/61197
2029 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
2031 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2033 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
2035 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
2037 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
2038 __SIZEOF_INT128__ is defined.
2040 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2042 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
2043 (rs6000_delegitimize_address): Use it.
2045 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2047 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
2048 inplace argument. Store the new address in the original MEM when true.
2049 * emit-rtl.c (change_address_1): Likewise.
2050 (adjust_address_1, adjust_automodify_address_1, offset_address):
2052 * rtl.h (plus_constant): Add an inplace argument.
2053 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
2054 when true. Avoid generating (plus X (const_int 0)).
2055 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
2056 in-place. Pass true to plus_constant.
2057 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
2059 2014-05-16 Dehao Chen <dehao@google.com>
2061 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
2063 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2066 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
2068 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
2070 2014-05-16 Dehao Chen <dehao@google.com>
2072 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
2073 optimize_function_for_size_p.
2074 * regs.h (REG_FREQ_FROM_BB): Likewise.
2076 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2079 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
2080 negt_reg_operand cases.
2081 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
2083 * config/sh/predicates.md (cbranch_treg_value): Simplify.
2085 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2087 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
2090 2014-05-16 David Malcolm <dmalcolm@redhat.com>
2093 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2095 * tree-cfg.c (dump_function_to_file): Dump the return type of
2096 functions, in a line to itself before the function body, mimicking
2097 the layout of a C function.
2099 2014-05-16 Dehao Chen <dehao@google.com>
2101 * cfghooks.c (make_forwarder_block): Use direct computation to
2102 get fall-through edge's count and frequency.
2104 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
2106 * config/arc/arc.c (arc_init): Fix typo in error message.
2107 * config/i386/i386.c (ix86_expand_builtin): Likewise.
2108 (split_stack_prologue_scratch_regno): Likewise.
2109 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
2110 word from error message.
2112 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
2114 * ira-costs.c: Fix typo in comment.
2116 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
2118 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
2120 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
2122 * varpool.c (dump_varpool_node): Dump write-only flag.
2123 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
2125 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
2126 write-only variables.
2127 * ipa.c (process_references): New function.
2128 (set_readonly_bit): New function.
2129 (set_writeonly_bit): New function.
2130 (clear_addressable_bit): New function.
2131 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
2132 fix handling of aliases.
2133 * cgraph.h (struct varpool_node): Add writeonly flag.
2135 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
2137 PR rtl-optimization/60969
2138 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
2139 Calculate costs for this case.
2141 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
2143 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
2144 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
2146 2014-05-16 Richard Biener <rguenther@suse.de>
2148 PR tree-optimization/61194
2149 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
2150 bool patterns ending in a COND_EXPR.
2152 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2154 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
2156 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2158 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
2159 where we were unable to cost an RTX.
2161 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2163 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
2166 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2167 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2169 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
2171 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2172 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2174 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
2175 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
2177 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2178 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2180 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
2183 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2184 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2186 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2189 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2190 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2192 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
2193 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
2195 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2196 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2198 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2201 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2202 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2204 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
2205 ZERO_EXTEND and SIGN_EXTEND better.
2207 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2208 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2210 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
2213 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2214 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2216 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
2217 costs when costing loads and stores to memory.
2219 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2220 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
2222 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
2225 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2227 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
2229 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2230 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2232 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
2234 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
2236 (aarch64_rtx_mult_cost): New.
2237 (aarch64_rtx_costs): Use it, refactor as appropriate.
2239 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2240 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2242 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
2243 emit instructions, return number of instructions which would
2245 (aarch64_add_constant): Update call to aarch64_build_constant.
2246 (aarch64_output_mi_thunk): Likewise.
2247 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
2250 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2252 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
2253 (TARGET_RTX_COSTS): Call it.
2255 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2257 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
2258 (cortexa57_vector_cost): Likewise.
2259 (cortexa57_tunings): Use them.
2261 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2263 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
2264 (cpu_addrcost_table): Use it.
2265 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
2266 (aarch64_address_cost): Rewrite using aarch64_classify_address,
2269 2014-05-16 Richard Biener <rguenther@suse.de>
2271 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
2272 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
2273 (visit_phi): Ignore edges marked as not executable.
2274 (class cond_dom_walker): New.
2275 (cond_dom_walker::before_dom_children): Value-number
2276 control statements and mark successor edges as not
2277 executable if possible.
2278 (run_scc_vn): First walk all control statements in
2279 dominator order, marking edges as not executable.
2280 * tree-inline.c (copy_edges_for_bb): Be not confused
2281 about random edge flags.
2283 2014-05-16 Richard Biener <rguenther@suse.de>
2285 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
2287 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
2290 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
2291 (__TM_simple_begin): Use it.
2292 (__TM_begin): Likewise.
2294 2014-05-15 Martin Jambor <mjambor@suse.cz>
2297 * ipa-prop.c (update_indirect_edges_after_inlining): Check
2298 type_preserved flag when the indirect edge is polymorphic.
2300 2014-05-15 Martin Jambor <mjambor@suse.cz>
2302 PR tree-optimization/61090
2303 * tree-sra.c (sra_modify_expr): Pass the current gsi to
2304 build_ref_for_model.
2306 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2308 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
2309 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
2311 2014-05-15 Jakub Jelinek <jakub@redhat.com>
2313 PR tree-optimization/61158
2314 * fold-const.c (fold_binary_loc): If X is zero-extended and
2315 shiftc >= prec, make sure zerobits is all ones instead of
2316 invoking undefined behavior.
2318 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2320 * regcprop.h: New file.
2321 * regcprop.c (skip_debug_insn_p): New decl.
2322 (replace_oldest_value_reg): Check skip_debug_insn_p.
2323 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
2324 * shrink-wrap.c: Include regcprop.h.
2325 (prepare_shrink_wrap): Call
2326 copyprop_hardreg_forward_bb_without_debug_insn.
2328 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2330 * shrink-wrap.h: Update comment.
2331 * shrink-wrap.c: Update comment.
2332 (next_block_for_reg): Rename to live_edge_for_reg.
2333 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
2334 (move_insn_for_shrink_wrap): Split live_edge.
2335 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
2337 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2339 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
2341 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
2342 * config/sparc/sparc.md (fptype_ut699): New attribute.
2343 (in_branch_delay): Return false if -mfix-ut699 is specified and
2344 fptype_ut699 is set to single.
2345 (truncdfsf2): Add fptype_ut699 attribute.
2346 (fix_truncdfsi2): Likewise.
2347 (floatsisf2): Change fptype attribute.
2348 (fix_truncsfsi2): Likewise.
2349 (negtf2_notv9): Delete.
2350 (negtf2_v9): Likewise.
2351 (negtf2_hq): New instruction.
2352 (negtf2): New instruction and splitter.
2353 (negdf2_notv9): Rewrite.
2354 (abstf2_notv9): Delete.
2355 (abstf2_hq_v9): Likewise.
2356 (abstf2_v9): Likewise.
2357 (abstf2_hq): New instruction.
2358 (abstf2): New instruction and splitter.
2359 (absdf2_notv9): Rewrite.
2361 2014-05-14 Cary Coutant <ccoutant@google.com>
2364 * opts.c (common_handle_option): Don't special-case "-g".
2365 (set_debug_level): Default to at least level 2 with "-g".
2367 2014-05-14 DJ Delorie <dj@redhat.com>
2369 * config/msp430/msp430.c (msp430_builtin): Add
2370 MSP430_BUILTIN_DELAY_CYCLES.
2371 (msp430_init_builtins): Register void __delay_cycles(long long).
2372 (msp430_builtin_decl): Add it.
2373 (cg_magic_constant): New.
2374 (msp430_expand_delay_cycles): New.
2375 (msp430_expand_builtin): Call it.
2376 (msp430_print_operand_raw): Change integer printing from "int" to
2378 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
2379 (delay_cycles_start): New.
2380 (delay_cycles_end): New.
2381 (delay_cycles_32): New.
2382 (delay_cycles_32x): New.
2383 (delay_cycles_16): New.
2384 (delay_cycles_16x): New.
2385 (delay_cycles_2): New.
2386 (delay_cycles_1): New.
2387 * doc/extend.texi: Document __delay_cycles().
2389 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
2391 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
2392 length attribute computation.
2394 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
2397 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
2399 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
2402 * config/sparc/sparc.md: Fix types of low and high in DI constant
2403 splitter. Use gen_int_mode in some other splitters.
2405 2014-05-14 Martin Jambor <mjambor@suse.cz>
2408 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
2410 2014-05-14 James Norris <jnorris@codesourcery.com>
2412 * omp-low.c (expand_parallel_call): Remove shadow variable.
2413 (expand_omp_taskreg): Likewise.
2415 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
2417 * common/config/i386/i386-common.c
2418 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
2419 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
2420 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
2421 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
2422 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
2423 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
2424 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
2425 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
2426 xsavecintrin.h, xsavesintrin.h.
2427 (x86_64-*-*): Ditto.
2428 * config/i386/clflushoptintrin.h: New.
2429 * config/i386/xsavecintrin.h: Ditto.
2430 * config/i386/xsavesintrin.h: Ditto.
2431 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
2432 (bit_XSAVES): Ditto.
2433 (bit_XSAVES): Ditto.
2434 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
2435 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
2437 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2438 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
2439 OPTION_MASK_ISA_XSAVES.
2440 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
2442 (PTA_CLFLUSHOPT) Define.
2443 (PTA_XSAVEC): Ditto.
2444 (PTA_XSAVES): Ditto.
2445 (ix86_option_override_internal): Handle new options.
2446 (ix86_valid_target_attribute_inner_p): Ditto.
2447 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
2448 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
2449 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
2450 (bdesc_special_args): Add __builtin_ia32_xsaves,
2451 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
2452 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
2453 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
2454 (ix86_expand_builtin): Handle new builtins.
2455 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
2456 (TARGET_CLFLUSHOPT_P): Ditto.
2457 (TARGET_XSAVEC): Ditto.
2458 (TARGET_XSAVEC_P): Ditto.
2459 (TARGET_XSAVES): Ditto.
2460 (TARGET_XSAVES_P): Ditto.
2461 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
2462 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
2463 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
2465 (ANY_XRSTOR64): Ditto.
2467 (xrstor): Change into <xrstor>.
2468 (xrstor_rex64): Change into <xrstor>_rex64.
2469 (xrstor64): Change into <xrstor>64
2471 * config/i386/i386.opt (mclflushopt): New.
2474 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
2476 * doc/invoke.texi: Document new options.
2478 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2480 PR rtl-optimization/60866
2481 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
2482 Default it to -1. Pass it down to init_simplejump_data.
2483 (init_simplejump_data): New parameter old_seqno. Pass it down
2484 to get_seqno_for_a_jump.
2485 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
2486 initializing new jump seqno as a last resort. Add comment.
2487 (sel_redirect_edge_and_branch): Save old seqno of the conditional
2488 jump and pass it down to sel_init_new_insn.
2489 (sel_redirect_edge_and_branch_force): Likewise.
2491 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
2493 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
2494 shifted values to avoid build warning.
2496 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2498 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
2499 * cfgrtl.c (rtl_merge_blocks): Fix comment.
2500 (cfg_layout_merge_blocks): Likewise.
2501 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
2503 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2505 PR rtl-optimization/60901
2506 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
2507 bb predecessor belongs to the same scheduling region. Adjust comment.
2509 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
2511 * doc/sourcebuild.texi: (dfp_hw): Document.
2512 (p8vector_hw): Likewise.
2513 (powerpc_eabi_ok): Likewise.
2514 (powerpc_elfv2): Likewise.
2515 (powerpc_htm_ok): Likewise.
2516 (ppc_recip_hw): Likewise.
2519 2014-05-13 Cary Coutant <ccoutant@google.com>
2521 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2523 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2525 * gengtype-parse.c (require3): Eliminate in favor of...
2527 (require_template_declaration): Update to support optional single *
2530 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2531 (create_user_defined_type): Handle a single level of explicit
2532 pointerness within template arguments.
2533 (struct write_types_data): Add field "kind".
2534 (filter_type_name): Handle "*" character.
2535 (write_user_func_for_structure_ptr): Require a write_types_data
2536 rather than just a prefix string, so that we can look up the kind
2537 of the wtd and use it as an index into wrote_user_func_for_ptr,
2538 ensuring that such functions are written at most once. Support
2539 subclasses by invoking the marking function of the ultimate base class.
2540 (write_user_func_for_structure_body): Require a write_types_data
2541 rather than just a prefix string, so that we can pass this to
2542 write_user_func_for_structure_ptr.
2543 (write_func_for_structure): Likewise.
2544 (ggc_wtd): Add initializer of new "kind" field.
2545 (pch_wtd): Likewise.
2547 * gengtype.h (enum write_types_kinds): New.
2548 (struct type): Add field wrote_user_func_for_ptr to the "s"
2551 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2553 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2554 instead of const_binop.
2555 (fold_binary_loc): Likewise.
2557 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2559 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2560 calculation to match get_ref_base_and_extent.
2562 2014-05-13 Catherine Moore <clm@codesourcery.com>
2563 Sandra Loosemore <sandra@codesourcery.com>
2565 * configure.ac: Fix assembly for explicit JALR relocation check.
2566 * configure: Regenerate.
2568 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2570 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2571 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2572 Remove associated type declarations and initialisations.
2573 (arm_expand_neon_builtin): Likewise.
2574 (neon_emit_pair_result_insn): Delete.
2575 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2576 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2577 (neon_vzip<mode>): Likewise.
2578 (neon_vuzp<mode>): Likewise.
2580 2014-05-13 Richard Biener <rguenther@suse.de>
2583 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2584 it needs revisiting whether the call still may be tail-called.
2586 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2588 * rtl.def (SYMBOL_REF): Remove middle "0" field.
2589 * rtl.h (block_symbol): Reduce number of fields to 2.
2590 (rtx_def): Add u2.symbol_ref_flags.
2591 (SYMBOL_REF_FLAGS): Use it.
2592 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
2593 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
2594 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
2595 Lower index of SYMBOL_REF_DATA.
2596 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
2597 Print SYMBOL_REF_FLAGS at the same time.
2598 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
2600 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2602 * rtl.def (VAR_LOCATION): Remove "i" field.
2603 * rtl.h (rtx_def): Add u2.var_location_status.
2604 (PAT_VAR_LOCATION_STATUS): Use it.
2605 (gen_rtx_VAR_LOCATION): Declare.
2606 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
2607 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
2608 * var-tracking.c (emit_note_insn_var_location): Remove casts.
2610 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2612 * rtl.def (scratch): Fix outdated comment and remove "0" field.
2613 * gengtype.c (adjust_field_rtx_def): Update accordingly.
2615 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2617 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
2618 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
2619 * rtl.h (rtx_def): Add insn_uid to u2 field.
2620 (RTX_FLAG_CHECK8): Delete in favor of...
2621 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
2622 (INSN_DELETED_P): Update accordingly.
2623 (INSN_UID): Use u2.insn_uid.
2624 (INSN_CHAIN_CODE_P): Define.
2625 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
2626 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
2627 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
2628 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
2629 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
2630 indices accordingly.
2631 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
2632 Update indices for insn-chain rtxes.
2633 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
2634 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
2635 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
2636 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
2637 * combine.c (try_combine): Likewise.
2638 * ira.c (setup_prohibited_mode_move_regs): Likewise.
2640 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2642 * rtl.def (REG): Remove middle field.
2643 * rtl.h (rtx_def): Add orignal_regno to u2.
2644 (ORIGINAL_REGNO): Use it instead of field 1.
2645 (REG_ATTRS): Lower field index accordingly.
2646 * gengtype.c (adjust_field_rtx_def): Remove handling of
2647 ORIGINAL_REGNO. Move REG_ATTRS index down.
2648 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
2649 code that prints the REGNO.
2651 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2653 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
2656 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2658 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
2660 2014-05-13 Bin Cheng <bin.cheng@arm.com>
2662 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
2663 (alloc_iv): Lower base expressions containing ADDR_EXPR.
2665 2014-05-13 Ian Bolton <ian.bolton@arm.com>
2667 * config/aarch64/aarch64-protos.h
2668 (aarch64_hard_regno_caller_save_mode): New prototype.
2669 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
2671 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
2673 2014-05-13 Christian Bruel <christian.bruel@st.com>
2675 * target.def (mode_switching): New hook vector.
2676 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
2677 (mode_exit, modepriority_to_mode): Likewise.
2678 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
2679 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2680 * target.h: Include tm.h and hard-reg-set.h.
2681 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
2682 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
2683 * doc/tm.texi Regenerate.
2684 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2685 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2686 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
2687 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
2688 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2689 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2690 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
2691 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
2692 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
2693 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
2694 (ix86_emit_mode_set): Hookify.
2695 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
2697 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2698 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
2699 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
2700 (epiphany_mode_priority_to_mode): Remove declaration.
2701 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
2702 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
2703 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
2705 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
2706 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
2707 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
2709 2014-05-13 Jakub Jelinek <jakub@redhat.com>
2712 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
2713 is const0_rtx, return immediately. Don't test count == 0 when
2716 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2718 * Makefile.in: add shrink-wrap.o.
2719 * config/i386/i386.c: include "shrink-wrap.h"
2720 * function.c: Likewise.
2721 (requires_stack_frame_p, next_block_for_reg,
2722 move_insn_for_shrink_wrap, prepare_shrink_wrap,
2723 dup_block_and_redirect): Move to shrink-wrap.c
2724 (thread_prologue_and_epilogue_insns): Extract three code segments
2725 as functions in shrink-wrap.c
2726 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
2728 * shrink-wrap.c: New file.
2729 * shrink-wrap.h: New file.
2731 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2733 * doc/extend.texi: Reflect current numbers of pragmas. Remove
2734 reference to Solaris.
2736 2014-05-12 Mike Stump <mikestump@comcast.net>
2739 * genattrtab.c (filename): Add.
2740 (convert_set_attr_alternative): Improve error message.
2741 (check_defs): Restore read_md_filename for error messages.
2742 (gen_insn): Save filename.
2744 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
2746 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
2747 -fno-local-ivars and -fivar-visibility.
2748 * c-family/c.opt: Make -Wshadow also implicitly enable
2751 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2753 * doc/tm.texi: Remove reference to deleted macro.
2754 * doc/tm.texi.in: Likewise.
2756 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
2759 * config/avr/avr.c (avr_out_store_psi): Use correct constant
2762 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
2765 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
2766 * config/arm/aout.h (License): Same.
2767 * config/arm/bpabi.h (License): Same.
2768 * config/arm/elf.h (License): Same.
2769 * config/arm/linux-elf.h (License): Same.
2770 * config/arm/linux-gas.h (License): Same.
2771 * config/arm/netbsd-elf.h (License): Same.
2772 * config/arm/uclinux-eabi.h (License): Same.
2773 * config/arm/uclinux-elf.h (License): Same.
2774 * config/arm/vxworks.h (License): Same.
2776 2014-05-11 Jakub Jelinek <jakub@redhat.com>
2778 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
2779 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
2780 number of operands to 3.
2781 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
2782 * tree-nested.c (convert_nonlocal_omp_clauses,
2783 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
2784 * gimplify.c (gimplify_scan_omp_clauses): Handle
2785 OMP_CLAUSE_LINEAR_STMT.
2786 * omp-low.c (lower_rec_input_clauses): Fix typo.
2787 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
2788 cast between Fortran boolean_type_node and C _Bool if
2791 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
2793 PR tree-optimization/61136
2794 * wide-int.h (multiple_of_p): Define a version that doesn't return
2796 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
2797 integer_zerop/const_binop pair.
2798 (multiple_of_p): Likewise, converting both operands to widest_int
2801 2014-05-09 Teresa Johnson <tejohnson@google.com>
2803 * cgraphunit.c (analyze_functions): Use correct dump file.
2805 2014-05-09 Florian Weimer <fweimer@redhat.com>
2807 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
2809 (stack_protect_return_slot_p): New function.
2810 (expand_used_vars): Call stack_protect_decl_p and
2811 stack_protect_return_slot_p for -fstack-protector-strong.
2813 2014-05-09 David Wohlferd <LimeGreenSocks@yahoo.com>
2814 Andrew Haley <aph@redhat.com>
2815 Richard Sandiford <rdsandiford@googlemail.com>
2817 * doc/extend.texi: Rewrite inline asm page / re-org asm-related
2820 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
2823 * fold-const.c (fold_binary_loc): Changed width of mask.
2825 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2827 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
2828 unsigned int initializers for regno_in, regno_out.
2830 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2833 * config/avr/avr.md (cc): Add new attribute set_vzn.
2834 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
2835 Set cc insn attribute to set_vzn instead of set_zn for alternatives
2836 with INC, DEC or NEG.
2837 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
2838 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
2839 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
2841 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2844 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2846 * wide-int.cc (UTItype): Define.
2847 (UDWtype): Define for appropriate W_TYPE_SIZE.
2849 2014-05-09 Richard Biener <rguenther@suse.de>
2851 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
2852 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
2853 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
2854 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
2855 ssa_propagate): Adjust.
2857 2014-05-08 Jeff Law <law@redhat.com>
2859 PR tree-optimization/61009
2860 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
2861 tri-state rather than a boolean. When a block is too big to
2862 thread through, inform caller via negative return value.
2863 (thread_across_edge): If a block was too big for normal threading,
2864 then it's too big for a joiner too, so remove temporary equivalences
2865 and return immediately.
2867 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
2868 Matthias Klose <doko@ubuntu.com>
2871 * optc-gen.awk: Fix option handling for -Wunused-parameter.
2873 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2876 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
2878 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2881 * config/alpha/alpha.c: Include gimple-iterator.h.
2882 (alpha_gimple_fold_builtin): New function. Move
2883 ALPHA_BUILTIN_UMULH folding from ...
2884 (alpha_fold_builtin): ... here.
2885 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
2887 2014-05-08 Wei Mi <wmi@google.com>
2890 * config/i386/i386.c (ix86_compute_frame_layout): Update
2891 preferred_stack_boundary for call, expanded from tls descriptor.
2892 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
2893 to depend on SP register.
2894 (*tls_local_dynamic_base_32_gnu): Ditto.
2895 (*tls_local_dynamic_32_once): Ditto.
2896 (tls_global_dynamic_64_<mode>): Set
2897 ix86_tls_descriptor_calls_expanded_in_cfun.
2898 (tls_local_dynamic_base_64_<mode>): Ditto.
2899 (tls_global_dynamic_32): Set
2900 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
2901 to depend on SP register.
2902 (tls_local_dynamic_base_32): Ditto.
2904 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2906 * config/arm/arm_neon.h: Update comment.
2907 * config/arm/neon-docgen.ml: Delete.
2908 * config/arm/neon-gen.ml: Delete.
2909 * doc/arm-neon-intrinsics.texi: Update comment.
2911 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2913 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
2915 (vand, vorr, veor, vorn, vbic): Remove.
2916 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
2918 (neon_vsub_unspec): Likewise.
2919 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
2921 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2923 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
2924 (vadd_s16): Likewise.
2925 (vadd_s32): Likewise.
2926 (vadd_f32): Likewise.
2927 (vadd_u8): Likewise.
2928 (vadd_u16): Likewise.
2929 (vadd_u32): Likewise.
2930 (vadd_s64): Likewise.
2931 (vadd_u64): Likewise.
2932 (vaddq_s8): Likewise.
2933 (vaddq_s16): Likewise.
2934 (vaddq_s32): Likewise.
2935 (vaddq_s64): Likewise.
2936 (vaddq_f32): Likewise.
2937 (vaddq_u8): Likewise.
2938 (vaddq_u16): Likewise.
2939 (vaddq_u32): Likewise.
2940 (vaddq_u64): Likewise.
2941 (vmul_s8): Likewise.
2942 (vmul_s16): Likewise.
2943 (vmul_s32): Likewise.
2944 (vmul_f32): Likewise.
2945 (vmul_u8): Likewise.
2946 (vmul_u16): Likewise.
2947 (vmul_u32): Likewise.
2948 (vmul_p8): Likewise.
2949 (vmulq_s8): Likewise.
2950 (vmulq_s16): Likewise.
2951 (vmulq_s32): Likewise.
2952 (vmulq_f32): Likewise.
2953 (vmulq_u8): Likewise.
2954 (vmulq_u16): Likewise.
2955 (vmulq_u32): Likewise.
2956 (vsub_s8): Likewise.
2957 (vsub_s16): Likewise.
2958 (vsub_s32): Likewise.
2959 (vsub_f32): Likewise.
2960 (vsub_u8): Likewise.
2961 (vsub_u16): Likewise.
2962 (vsub_u32): Likewise.
2963 (vsub_s64): Likewise.
2964 (vsub_u64): Likewise.
2965 (vsubq_s8): Likewise.
2966 (vsubq_s16): Likewise.
2967 (vsubq_s32): Likewise.
2968 (vsubq_s64): Likewise.
2969 (vsubq_f32): Likewise.
2970 (vsubq_u8): Likewise.
2971 (vsubq_u16): Likewise.
2972 (vsubq_u32): Likewise.
2973 (vsubq_u64): Likewise.
2974 (vand_s8): Likewise.
2975 (vand_s16): Likewise.
2976 (vand_s32): Likewise.
2977 (vand_u8): Likewise.
2978 (vand_u16): Likewise.
2979 (vand_u32): Likewise.
2980 (vand_s64): Likewise.
2981 (vand_u64): Likewise.
2982 (vandq_s8): Likewise.
2983 (vandq_s16): Likewise.
2984 (vandq_s32): Likewise.
2985 (vandq_s64): Likewise.
2986 (vandq_u8): Likewise.
2987 (vandq_u16): Likewise.
2988 (vandq_u32): Likewise.
2989 (vandq_u64): Likewise.
2990 (vorr_s8): Likewise.
2991 (vorr_s16): Likewise.
2992 (vorr_s32): Likewise.
2993 (vorr_u8): Likewise.
2994 (vorr_u16): Likewise.
2995 (vorr_u32): Likewise.
2996 (vorr_s64): Likewise.
2997 (vorr_u64): Likewise.
2998 (vorrq_s8): Likewise.
2999 (vorrq_s16): Likewise.
3000 (vorrq_s32): Likewise.
3001 (vorrq_s64): Likewise.
3002 (vorrq_u8): Likewise.
3003 (vorrq_u16): Likewise.
3004 (vorrq_u32): Likewise.
3005 (vorrq_u64): Likewise.
3006 (veor_s8): Likewise.
3007 (veor_s16): Likewise.
3008 (veor_s32): Likewise.
3009 (veor_u8): Likewise.
3010 (veor_u16): Likewise.
3011 (veor_u32): Likewise.
3012 (veor_s64): Likewise.
3013 (veor_u64): Likewise.
3014 (veorq_s8): Likewise.
3015 (veorq_s16): Likewise.
3016 (veorq_s32): Likewise.
3017 (veorq_s64): Likewise.
3018 (veorq_u8): Likewise.
3019 (veorq_u16): Likewise.
3020 (veorq_u32): Likewise.
3021 (veorq_u64): Likewise.
3022 (vbic_s8): Likewise.
3023 (vbic_s16): Likewise.
3024 (vbic_s32): Likewise.
3025 (vbic_u8): Likewise.
3026 (vbic_u16): Likewise.
3027 (vbic_u32): Likewise.
3028 (vbic_s64): Likewise.
3029 (vbic_u64): Likewise.
3030 (vbicq_s8): Likewise.
3031 (vbicq_s16): Likewise.
3032 (vbicq_s32): Likewise.
3033 (vbicq_s64): Likewise.
3034 (vbicq_u8): Likewise.
3035 (vbicq_u16): Likewise.
3036 (vbicq_u32): Likewise.
3037 (vbicq_u64): Likewise.
3038 (vorn_s8): Likewise.
3039 (vorn_s16): Likewise.
3040 (vorn_s32): Likewise.
3041 (vorn_u8): Likewise.
3042 (vorn_u16): Likewise.
3043 (vorn_u32): Likewise.
3044 (vorn_s64): Likewise.
3045 (vorn_u64): Likewise.
3046 (vornq_s8): Likewise.
3047 (vornq_s16): Likewise.
3048 (vornq_s32): Likewise.
3049 (vornq_s64): Likewise.
3050 (vornq_u8): Likewise.
3051 (vornq_u16): Likewise.
3052 (vornq_u32): Likewise.
3053 (vornq_u64): Likewise.
3055 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3057 * wide-int.cc (UTItype): Define.
3058 (UDWtype): Define for appropriate W_TYPE_SIZE.
3060 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
3062 PR tree-optimization/59100
3063 * tree-ssa-phiopt.c: Include tree-inline.h.
3064 (neutral_element_p, absorbing_element_p): New functions.
3065 (value_replacement): Handle conditional binary operations with a
3066 neutral or absorbing element.
3068 2014-05-08 Richard Biener <rguenther@suse.de>
3070 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
3071 folding the expression.
3072 (valueize_expr): Remove.
3073 (visit_reference_op_load): Do not valueize the result of
3075 (simplify_binary_expression): Likewise.
3076 (simplify_unary_expression): Likewise.
3078 2014-05-08 Richard Biener <rguenther@suse.de>
3080 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
3081 looking at TYPE_ARG_TYPES.
3083 2014-05-08 Richard Biener <rguenther@suse.de>
3085 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
3086 pointer propagation special-case.
3088 2014-05-08 Bin Cheng <bin.cheng@arm.com>
3090 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
3091 core part of address expressions.
3093 2014-05-08 Alan Modra <amodra@gmail.com>
3096 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
3097 loads and stores when -mno-strict-align at any alignment.
3098 (expand_block_clear): Similarly. Also correct calculation of
3101 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
3104 * tree-complex.c (expand_complex_move): Keep line info when expanding
3106 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
3107 of complex expression. Use new argument to display correct location
3108 for values coming from phi statement.
3109 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
3110 (warn_uninitialized_phi): Pass location of phi argument to
3112 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
3113 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
3115 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
3117 * config/rs6000/predicates.md (indexed_address_mem): New.
3118 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
3119 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
3120 fpstore_ux, fpstore_u.
3121 (sign_extend, indexed, update): New.
3122 (cell_micro): Adjust.
3123 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
3124 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
3125 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
3126 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
3127 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
3128 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
3129 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
3130 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
3131 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
3132 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
3133 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
3134 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
3135 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
3136 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
3137 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
3139 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
3140 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
3141 *vsx_extract_<mode>_store): Adjust.
3142 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
3143 is_cracked_insn, insn_must_be_first_in_group,
3144 insn_must_be_last_in_group): Adjust.
3146 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
3148 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
3149 ppc440-fpstore): Adjust.
3150 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
3151 ppc476-fpstore): Adjust.
3152 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
3153 ppc601-fpstore): Adjust.
3154 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
3156 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
3158 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
3159 ppc7450-fpstore): Adjust.
3160 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
3161 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
3162 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
3164 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
3165 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
3166 cell-fpstore, cell-fpstore-update): Adjust.
3167 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
3168 ppce300c3_store, ppce300c3_fpstore): Adjust.
3169 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
3170 e500mc_fpstore): Adjust.
3171 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
3172 e500mc64_store, e500mc64_fpstore): Adjust.
3173 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
3174 e5500_fpstore): Adjust.
3175 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
3176 e6500_fpstore): Adjust.
3177 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
3179 * config/rs6000/power4.md (power4-load, power4-load-ext,
3180 power4-load-ext-update, power4-load-ext-update-indexed,
3181 power4-load-update-indexed, power4-load-update, power4-fpload,
3182 power4-fpload-update, power4-store, power4-store-update,
3183 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
3185 * config/rs6000/power5.md (power5-load, power5-load-ext,
3186 power5-load-ext-update, power5-load-ext-update-indexed,
3187 power5-load-update-indexed, power5-load-update, power5-fpload,
3188 power5-fpload-update, power5-store, power5-store-update,
3189 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
3191 * config/rs6000/power6.md (power6-load, power6-load-ext,
3192 power6-load-update, power6-load-update-indexed,
3193 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
3194 power6-fpload-update, power6-store, power6-store-update,
3195 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
3197 * config/rs6000/power7.md (power7-load, power7-load-ext,
3198 power7-load-update, power7-load-update-indexed,
3199 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
3200 power7-fpload-update, power7-store, power7-store-update,
3201 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
3203 * config/rs6000/power8.md (power8-load, power8-load-update,
3204 power8-load-ext, power8-load-ext-update, power8-fpload,
3205 power8-fpload-update, power8-store, power8-store-update-indexed,
3206 power8-fpstore, power8-fpstore-update): Adjust.
3207 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
3209 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
3210 titan_lsu_store, titan_lsu_fpstore): Adjust.
3211 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
3213 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
3216 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
3217 unrolled byte insns. Emit address increments after move insns.
3219 2014-05-07 David Malcolm <dmalcolm@redhat.com>
3221 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
3222 const_gimple, rather than a gimple.
3223 (gimple_call_builtin_p): Likewise, for the three variants.
3225 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
3226 (gimple_call_builtin_p): Likewise, for the three variants.
3228 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3230 PR tree-optimization/61095
3231 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
3233 2014-05-07 Richard Biener <rguenther@suse.de>
3235 PR tree-optimization/61034
3236 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
3237 (maybe_skip_until): Use translate to take into account
3238 lattices when trying to do disambiguations.
3239 (get_continuation_for_phi_1): Likewise.
3240 (get_continuation_for_phi): Adjust for added translate arguments.
3241 (walk_non_aliased_vuses): Likewise.
3242 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
3243 (walk_non_aliased_vuses): Likewise.
3244 (call_may_clobber_ref_p_1): Declare.
3245 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
3246 calls. Stop early if we are only supposed to disambiguate.
3247 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
3249 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
3251 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
3252 Emit an error when the function has arguments.
3254 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3256 * cfgloop.h (unswitch_loops): Remove.
3257 * doc/passes.texi: Remove references to loop-unswitch.c
3258 * timevar.def (TV_LOOP_UNSWITCH): Remove.
3260 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
3262 * tree-vect-data-refs.c (vect_grouped_load_supported): New
3263 check for loads group of length 3.
3264 (vect_permute_load_chain): New permutations for loads group of
3266 * tree-vect-stmts.c (vect_model_load_cost): Change cost
3267 of vec_perm_shuffle for the new permutations.
3269 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
3271 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
3272 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
3273 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
3274 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
3275 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
3276 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
3277 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
3278 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
3280 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3282 * loop-unswitch.c: Delete.
3284 2014-05-07 Richard Biener <rguenther@suse.de>
3286 * config.gcc: Always set need_64bit_hwint to yes.
3288 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
3290 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
3291 of using optimize_size.
3293 2014-05-06 Mike Stump <mikestump@comcast.net>
3295 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
3297 2014-05-06 Joseph Myers <joseph@codesourcery.com>
3299 * config/i386/sse.md (*mov<mode>_internal)
3300 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
3301 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
3302 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
3303 (*<code><mode>3, *andnot<mode>3<mask_name>)
3304 (<mask_codefor><code><mode>3<mask_name>): Only consider
3305 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
3307 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3310 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3312 * lra-constraints.c (valid_address_p): Move earlier in file.
3313 Add a constraint argument to the address_info version.
3314 (satisfies_memory_constraint_p): New function.
3315 (satisfies_address_constraint_p): Likewise.
3316 (process_alt_operands, curr_insn_transform): Use them.
3317 (process_address): Pass the constraint to valid_address_p when
3318 checking address operands.
3320 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
3322 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
3323 to their respective blocks. Fix inadvertent use of "node".
3325 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3327 * emit-rtl.c (init_derived_machine_modes): New functionm, split
3329 (init_emit_once): ...here.
3330 * rtl.h (init_derived_machine_modes): Declare.
3331 * toplev.c (do_compile): Call it even if no_backend.
3333 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
3334 Mike Stump <mikestump@comcast.net>
3335 Richard Sandiford <rdsandiford@googlemail.com>
3336 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3338 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
3339 (rtx_equal_for_memref_p): Update comment.
3340 (adjust_offset_for_component_ref): Use wide-int interfaces.
3341 * builtins.c (get_object_alignment_2): Likewise.
3342 (c_readstr): Likewise.
3343 (target_char_cast): Add comment.
3344 (determine_block_size): Use wide-int interfaces.
3345 (expand_builtin_signbit): Likewise.
3346 (fold_builtin_int_roundingfn): Likewise.
3347 (fold_builtin_bitop): Likewise.
3348 (fold_builtin_bswap): Likewise.
3349 (fold_builtin_logarithm): Use signop.
3350 (fold_builtin_pow): Likewise.
3351 (fold_builtin_memory_op): Use wide-int interfaces.
3352 (fold_builtin_object_size): Likewise.
3353 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
3354 nb_iterations_estimate.
3355 (record_niter_bound): Use wide-int interfaces.
3356 (get_estimated_loop_iterations_int): Likewise.
3357 (get_estimated_loop_iterations): Likewise.
3358 (get_max_loop_iterations): Likewise.
3359 * cfgloop.h: Include wide-int.h.
3360 (struct nb_iter_bound): Change bound to widest_int.
3361 (struct loop): Change nb_iterations_upper_bound and
3362 nb_iterations_estimate to widest_int.
3363 (record_niter_bound): Switch to use widest_int.
3364 (get_estimated_loop_iterations): Likewise.
3365 (get_max_loop_iterations): Likewise.
3366 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
3367 update for wide-int.
3368 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
3369 * combine.c (try_combine): Likewise.
3370 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
3371 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
3373 (aarch64_float_const_representable_p): Likewise.
3374 * config/arc/arc.c: Include wide-int.h.
3375 (arc_can_use_doloop_p): Use wide-int interfaces.
3376 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
3377 (vfp3_const_double_index): Likewise.
3378 * config/avr/avr.c (avr_out_round): Likewise.
3379 (avr_fold_builtin): Likewise.
3380 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
3381 (bfin_can_use_doloop_p): Likewise.
3382 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
3383 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
3384 * config/i386/i386.c: Include wide-int.h.
3385 (ix86_data_alignment): Use wide-int interfaces.
3386 (ix86_local_alignment): Likewise.
3387 (ix86_emit_swsqrtsf): Update real_from_integer.
3388 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
3389 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
3390 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
3391 (zero_constant): Likewise.
3392 (input_operand): Likewise.
3393 (splat_input_operand): Likewise.
3394 (non_logical_cint_operand): Change const_double to const_wide_int.
3395 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
3396 (easy_altivec_constant): Remove comment.
3397 (paired_expand_vector_init): Use CONSTANT_P.
3398 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
3399 (rs6000_emit_move): Update checks.
3400 (rs6000_aggregate_candidate): Use wide-int interfaces.
3401 (rs6000_expand_ternop_builtin): Likewise.
3402 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
3403 (rs6000_assemble_integer): Likewise.
3404 (rs6000_hash_constant): Likewise.
3405 (output_toc): Likewise.
3406 (rs6000_rtx_costs): Likewise.
3407 (rs6000_emit_swrsqrt); Update call to real_from_integer.
3408 * config/rs6000/rs6000-c.c: Include wide-int.h.
3409 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
3410 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
3411 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
3412 Handle CONST_WIDE_INT.
3413 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
3414 Use tree_fits_uhwi_p.
3415 * config/sparc/sparc.c: Include wide-int.h.
3416 (sparc_fold_builtin): Use wide-int interfaces.
3417 * config/vax/vax.c: Include wide-int.h.
3418 (vax_float_literal): Use real_from_integer.
3419 * coretypes.h (struct hwivec_def): New.
3421 (const_hwivec): New.
3422 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
3423 (equiv_constant): Handle CONST_WIDE_INT.
3424 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
3425 (cselib_hash_rtx): Handle CONST_WIDE_INT.
3426 * dbxout.c (stabstr_U): Use wide-int interfaces.
3427 (dbxout_type): Update to use cst_fits_shwi_p.
3428 * defaults.h (LOG2_BITS_PER_UNIT): Define.
3429 (TARGET_SUPPORTS_WIDE_INT): Add default.
3430 * dfp.c: Include wide-int.h.
3431 (decimal_real_to_integer2): Use wide-int interfaces and rename to
3432 decimal_real_to_integer.
3433 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
3434 decimal_real_to_integer.
3435 * doc/generic.texi (Constant expressions): Update for wide_int.
3436 * doc/rtl.texi (const_double): Likewise.
3437 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
3438 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
3439 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
3440 (REAL_VALUE_FROM_INT): Remove.
3441 (TARGET_SUPPORTS_WIDE_INT): New.
3442 * doc/tm.texi: Regenerate.
3443 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
3444 * double-int.h: Include wide-int.h.
3445 (struct wi::int_traits): New.
3446 * dwarf2out.c (get_full_len): New.
3447 (dw_val_equal_p): Add case dw_val_class_wide_int.
3448 (size_of_loc_descr): Likewise.
3449 (output_loc_operands): Likewise.
3450 (insert_double): Remove.
3451 (insert_wide_int): New.
3453 (print_die): Add case dw_val_class_wide_int.
3454 (attr_checksum): Likewise.
3455 (attr_checksum_ordered): Likewise.
3456 (same_dw_val_p): Likewise.
3457 (size_of_die): Likewise.
3458 (value_format): Likewise.
3459 (output_die): Likewise.
3460 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
3462 (clz_loc_descriptor): Use wide-int interfaces.
3463 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
3464 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
3465 (round_up_to_align): Use wide-int interfaces.
3466 (field_byte_offset): Likewise.
3467 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
3468 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
3469 CONST_DOUBLE handling. Use wide-int interfaces.
3470 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
3471 (gen_enumeration_type_die): Use add_AT_wide.
3472 (hash_loc_operands): Add case dw_val_class_wide_int.
3473 (compare_loc_operands): Likewise.
3474 * dwarf2out.h: Include wide-int.h.
3475 (wide_int_ptr): New.
3476 (enum dw_val_class): Add dw_val_class_wide_int.
3477 (struct dw_val_struct): Add val_wide.
3478 * emit-rtl.c (const_wide_int_htab): New.
3479 (const_wide_int_htab_hash): New.
3480 (const_wide_int_htab_eq): New.
3481 (lookup_const_wide_int): New.
3482 (const_double_htab_hash): Use wide-int interfaces.
3483 (const_double_htab_eq): Likewise.
3484 (rtx_to_double_int): Conditionally compile for wide-int.
3485 (immed_double_int_const): Rename to immed_wide_int_const and
3486 update for wide-int.
3487 (immed_double_const): Conditionally compile for wide-int.
3488 (init_emit_once): Use wide-int interfaces.
3489 * explow.c (plus_constant): Likewise.
3490 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
3491 (lshift_value): Use wide-int interfaces.
3492 (expand_mult): Likewise.
3493 (choose_multiplier): Likewise.
3494 (expand_smod_pow2): Likewise.
3495 (make_tree): Likewise.
3496 * expr.c (convert_modes): Consolidate handling of constants.
3497 Use wide-int interfaces.
3498 (emit_group_load_1): Add note.
3499 (store_expr): Update comment.
3500 (get_inner_reference): Use wide-int interfaces.
3501 (expand_constructor): Update comment.
3502 (expand_expr_real_2): Use wide-int interfaces.
3503 (expand_expr_real_1): Likewise.
3504 (reduce_to_bit_field_precision): Likewise.
3505 (const_vector_from_tree): Likewise.
3506 * final.c: Include wide-int-print.h.
3507 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
3508 * fixed-value.c: Include wide-int.h.
3509 (fixed_from_string): Use wide-int interfaces.
3510 (fixed_to_decimal): Likewise.
3511 (fixed_convert_from_real): Likewise.
3512 (real_convert_from_fixed): Likewise.
3513 * fold-const.h (mem_ref_offset): Return an offset_int.
3514 (div_if_zero_remainder): Remove code parameter.
3515 * fold-const.c (div_if_zero_remainder): Remove code parameter.
3516 Use wide-int interfaces.
3517 (may_negate_without_overflow_p): Use wide-int interfaces.
3518 (negate_expr_p): Likewise.
3519 (fold_negate_expr): Likewise.
3520 (int_const_binop_1): Likewise.
3521 (const_binop): Likewise.
3522 (fold_convert_const_int_from_int): Likewise.
3523 (fold_convert_const_int_from_real): Likewise.
3524 (fold_convert_const_int_from_fixed): Likewise.
3525 (fold_convert_const_fixed_from_int): Likewise.
3526 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
3527 (sign_bit_p): Use wide-int interfaces.
3528 (make_range_step): Likewise.
3529 (build_range_check): Likewise. Pass an integer of the correct type
3530 instead of using integer_one_node.
3531 (range_predecessor): Pass an integer of the correct type instead
3532 of using integer_one_node.
3533 (range_successor): Likewise.
3534 (merge_ranges): Likewise.
3535 (unextend): Use wide-int interfaces.
3536 (extract_muldiv_1): Likewise.
3537 (fold_div_compare): Likewise.
3538 (fold_single_bit_test): Likewise.
3539 (fold_sign_changed_comparison): Likewise.
3540 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3541 (fold_plusminus_mult_expr): Use wide-int interfaces.
3542 (native_encode_int): Likewise.
3543 (native_interpret_int): Likewise.
3544 (fold_unary_loc): Likewise.
3545 (pointer_may_wrap_p): Likewise.
3546 (size_low_cst): Likewise.
3547 (mask_with_tz): Likewise.
3548 (fold_binary_loc): Likewise.
3549 (fold_ternary_loc): Likewise.
3550 (multiple_of_p): Likewise.
3551 (tree_call_nonnegative_warnv_p): Update calls to
3552 tree_int_cst_min_precision and real_from_integer.
3553 (fold_negate_const): Use wide-int interfaces.
3554 (fold_abs_const): Likewise.
3555 (fold_relational_const): Use tree_int_cst_lt.
3556 (round_up_loc): Use wide-int interfaces.
3557 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3558 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3559 * gengtype.c: Remove include of double-int.h.
3560 (do_typedef): Use wide-int interfaces.
3561 (open_base_files): Add wide-int.h.
3562 (main): Add offset_int and widest_int typedefs.
3563 * gengtype-lex.l: Handle "^".
3564 (CXX_KEYWORD): Add "static".
3565 * gengtype-parse.c (require3): New.
3566 (require_template_declaration): Handle constant template arguments
3567 and nested templates.
3568 * gengtype-state.c: Don't include "double-int.h".
3569 * genpreds.c (write_one_predicate_function): Update comment.
3570 (write_tm_constrs_h): Add check for hval and lval use in
3572 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3573 (add_to_sequence): Likewise.
3574 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3575 and const_double_operand.
3576 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3578 * gimple-fold.c (get_base_constructor): Likewise.
3579 (fold_array_ctor_reference): Likewise.
3580 (fold_nonarray_ctor_reference): Likewise.
3581 (fold_const_aggregate_ref_1): Likewise.
3582 (gimple_val_nonnegative_real_p): Likewise.
3583 (gimple_fold_indirect_ref): Likewise.
3584 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3585 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
3586 (struct slsr_cand_d): Change index to be widest_int.
3587 (struct incr_info_d): Change incr to be widest_int.
3588 (alloc_cand_and_find_basis): Use wide-int interfaces.
3589 (slsr_process_phi): Likewise.
3590 (backtrace_base_for_ref): Likewise. Return a widest_int.
3591 (restructure_reference): Take a widest_int instead of a double_int.
3592 (slsr_process_ref): Use wide-int interfaces.
3593 (create_mul_ssa_cand): Likewise.
3594 (create_mul_imm_cand): Likewise.
3595 (create_add_ssa_cand): Likewise.
3596 (create_add_imm_cand): Take a widest_int instead of a double_int.
3597 (slsr_process_add): Use wide-int interfaces.
3598 (slsr_process_cast): Likewise.
3599 (slsr_process_copy): Likewise.
3600 (dump_candidate): Likewise.
3601 (dump_incr_vec): Likewise.
3602 (replace_ref): Likewise.
3603 (cand_increment): Likewise. Return a widest_int.
3604 (cand_abs_increment): Likewise.
3605 (replace_mult_candidate): Take a widest_int instead of a double_int.
3606 (replace_unconditional_candidate): Use wide-int interfaces.
3607 (incr_vec_index): Take a widest_int instead of a double_int.
3608 (create_add_on_incoming_edge): Likewise.
3609 (create_phi_basis): Use wide-int interfaces.
3610 (replace_conditional_candidate): Likewise.
3611 (record_increment): Take a widest_int instead of a double_int.
3612 (record_phi_increments): Use wide-int interfaces.
3613 (phi_incr_cost): Take a widest_int instead of a double_int.
3614 (lowest_cost_path): Likewise.
3615 (total_savings): Likewise.
3616 (analyze_increments): Use wide-int interfaces.
3617 (ncd_with_phi): Take a widest_int instead of a double_int.
3618 (ncd_of_cand_and_phis): Likewise.
3619 (nearest_common_dominator_for_cands): Likewise.
3620 (insert_initializers): Use wide-int interfaces.
3621 (all_phi_incrs_profitable): Likewise.
3622 (replace_one_candidate): Likewise.
3623 (replace_profitable_candidates): Likewise.
3624 * godump.c: Include wide-int-print.h.
3625 (go_output_typedef): Use wide-int interfaces.
3626 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
3627 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
3628 (build_loop_iteration_domains): Likewise.
3629 * hooks.h: Include wide-int.h rather than double-int.h.
3630 (hook_bool_dint_dint_uint_bool_true): Delete.
3631 (hook_bool_wint_wint_uint_bool_true): Declare.
3632 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
3633 (hook_bool_wint_wint_uint_bool_true): New.
3634 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
3636 (ubsan_expand_si_overflow_mul_check): Likewise.
3637 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
3638 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
3639 (get_ancestor_addr_info): Likewise.
3640 (ipa_modify_call_arguments): Likewise.
3641 * loop-doloop.c (doloop_modify): Likewise.
3642 (doloop_optimize): Likewise.
3643 * loop-iv.c (iv_number_of_iterations): Likewise.
3644 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
3645 (unroll_loop_constant_iterations): Likewise.
3646 (decide_unroll_runtime_iterations): Likewise.
3647 (unroll_loop_runtime_iterations): Likewise.
3648 (decide_peel_simple): Likewise.
3649 (decide_unroll_stupid): Likewise.
3650 * lto-streamer-in.c (streamer_read_wi): Add.
3651 (input_cfg): Use wide-int interfaces.
3652 (lto_input_tree_1): Likewise.
3653 * lto-streamer-out.c (streamer_write_wi): Add.
3654 (hash_tree): Use wide-int interfaces.
3655 (output_cfg): Likewise.
3656 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
3657 (GTFILES): Add wide-int.h and signop.h.
3658 (TAGS): Look for .cc files too.
3659 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
3660 * optabs.c (expand_subword_shift): Likewise.
3661 (expand_doubleword_shift): Likewise.
3662 (expand_absneg_bit): Likewise.
3663 (expand_copysign_absneg): Likewise.
3664 (expand_copysign_bit): Likewise.
3665 * postreload.c (reload_cse_simplify_set): Likewise.
3666 * predict.c (predict_iv_comparison): Likewise.
3667 * pretty-print.h: Include wide-int-print.h.
3669 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
3670 * print-tree.c: Include wide-int-print.h.
3671 (print_node_brief): Use wide-int interfaces.
3672 (print_node): Likewise.
3673 * read-rtl.c (validate_const_wide_int): New.
3674 (read_rtx_code): Add CONST_WIDE_INT case.
3675 * real.c: Include wide-int.h.
3676 (real_to_integer2): Delete.
3677 (real_to_integer): New function, returning a wide_int.
3678 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3679 (ten_to_ptwo): Update call to real_from_integer.
3680 (real_digit): Likewise.
3681 * real.h: Include signop.h, wide-int.h and insn-modes.h.
3682 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
3683 (REAL_VALUE_TO_INT): Delete.
3684 (real_to_integer): Declare a wide-int form.
3685 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3686 * recog.c (const_int_operand): Improve comment.
3687 (const_scalar_int_operand): New.
3688 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
3689 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
3690 (split_double): Likewise.
3691 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
3692 (rtx_size): Likewise.
3693 (rtx_alloc_stat_v): New.
3694 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
3695 (cwi_output_hex): New.
3696 (iterative_hash_rtx): Handle CONST_WIDE_INT.
3697 (cwi_check_failed_bounds): New.
3698 * rtl.def (CONST_WIDE_INT): New.
3699 * rtl.h: Include <utility> and wide-int.h.
3700 (struct hwivec_def): New.
3701 (CWI_GET_NUM_ELEM): New.
3702 (CWI_PUT_NUM_ELEM): New.
3703 (struct rtx_def): Add num_elem and hwiv.
3704 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
3705 (CASE_CONST_UNIQUE): Likewise.
3706 (CASE_CONST_ANY): Likewise.
3707 (CONST_SCALAR_INT_P): Likewise.
3708 (CONST_WIDE_INT_P): New.
3710 (HWIVEC_CHECK): New.
3711 (cwi_check_failed_bounds): New.
3713 (HWIVEC_CHECK): New.
3714 (CONST_WIDE_INT_VEC) New.
3715 (CONST_WIDE_INT_NUNITS) New.
3716 (CONST_WIDE_INT_ELT) New.
3717 (rtx_mode_t): New type.
3718 (wi::int_traits <rtx_mode_t>): New.
3720 (wi::min_value): New.
3721 (wi::max_value): New.
3723 (const_wide_int_alloc): New.
3724 (immed_wide_int_const): New.
3725 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
3726 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
3727 * signop.h: New file.
3728 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
3729 (simplify_const_unary_operation): Use wide-int interfaces.
3730 (simplify_binary_operation_1): Likewise.
3731 (simplify_const_binary_operation): Likewise.
3732 (simplify_const_relational_operation): Likewise.
3733 (simplify_immed_subreg): Likewise.
3734 * stmt.c (expand_case): Likewise.
3735 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
3736 signop rather than a bool.
3737 * stor-layout.c (layout_type): Use wide-int interfaces.
3738 (initialize_sizetypes): Update calls to
3739 set_min_and_max_values_for_integral_type.
3740 (set_min_and_max_values_for_integral_type): Take a signop rather
3741 than a bool. Use wide-int interfaces.
3742 (fixup_signed_type): Update accordingly. Remove
3743 HOST_BITS_PER_DOUBLE_INT limit.
3744 (fixup_unsigned_type): Likewise.
3745 * system.h (STATIC_CONSTANT_P): New.
3746 (STATIC_ASSERT): New.
3747 * target.def (can_use_doloop_p): Take widest_ints rather than
3749 * target.h: Include wide-int.h rather than double-int.h.
3750 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
3752 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
3753 rather than INT_CST_LT_UNSIGNED.
3754 (can_use_doloop_if_innermost): Take widest_ints rather than
3756 * tree-affine.c: Include wide-int-print.h.
3757 (double_int_ext_for_comb): Delete.
3758 (wide_int_ext_for_comb): New.
3759 (aff_combination_zero): Use wide-int interfaces.
3760 (aff_combination_const): Take a widest_int instead of a double_int.
3761 (aff_combination_elt): Use wide-int interfaces.
3762 (aff_combination_scale): Take a widest_int instead of a double_int.
3763 (aff_combination_add_elt): Likewise.
3764 (aff_combination_add_cst): Likewise.
3765 (aff_combination_add): Use wide-int interfaces.
3766 (aff_combination_convert): Likewise.
3767 (tree_to_aff_combination): Likewise.
3768 (add_elt_to_tree): Take a widest_int instead of a double_int.
3769 (aff_combination_to_tree): Use wide-int interfaces.
3770 (aff_combination_remove_elt): Likewise.
3771 (aff_combination_add_product): Take a widest_int instead of
3773 (aff_combination_mult): Use wide-int interfaces.
3774 (aff_combination_expand): Likewise.
3775 (double_int_constant_multiple_p): Delete.
3776 (wide_int_constant_multiple_p): New.
3777 (aff_combination_constant_multiple_p): Take a widest_int pointer
3778 instead of a double_int pointer.
3779 (print_aff): Use wide-int interfaces.
3780 (get_inner_reference_aff): Take a widest_int pointer
3781 instead of a double_int pointer.
3782 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
3783 * tree-affine.h: Include wide-int.h.
3784 (struct aff_comb_elt): Change type of coef to widest_int.
3785 (struct affine_tree_combination): Change type of offset to widest_int.
3786 (double_int_ext_for_comb): Delete.
3787 (wide_int_ext_for_comb): New.
3788 (aff_combination_const): Use widest_int instead of double_int.
3789 (aff_combination_scale): Likewise.
3790 (aff_combination_add_elt): Likewise.
3791 (aff_combination_constant_multiple_p): Likewise.
3792 (get_inner_reference_aff): Likewise.
3793 (aff_comb_cannot_overlap_p): Likewise.
3794 (aff_combination_zero_p): Use wide-int interfaces.
3795 * tree.c: Include tree.h.
3796 (init_ttree): Use make_int_cst.
3797 (tree_code_size): Removed code for INTEGER_CST case.
3798 (tree_size): Add INTEGER_CST case.
3799 (make_node_stat): Update comment.
3800 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
3801 (build_int_cst_type): Use wide-int interfaces.
3802 (double_int_to_tree): Likewise.
3803 (double_int_fits_to_tree_p): Delete.
3804 (force_fit_type_double): Delete.
3805 (force_fit_type): New.
3806 (int_cst_hash_hash): Use wide-int interfaces.
3807 (int_cst_hash_eq): Likewise.
3808 (build_int_cst_wide): Delete.
3809 (wide_int_to_tree): New.
3810 (cache_integer_cst): Use wide-int interfaces.
3811 (build_low_bits_mask): Likewise.
3812 (cst_and_fits_in_hwi): Likewise.
3813 (real_value_from_int_cst): Likewise.
3814 (make_int_cst_stat): New.
3815 (integer_zerop): Use wide_int interfaces.
3816 (integer_onep): Likewise.
3817 (integer_all_onesp): Likewise.
3818 (integer_pow2p): Likewise.
3819 (integer_nonzerop): Likewise.
3820 (tree_log2): Likewise.
3821 (tree_floor_log2): Likewise.
3822 (tree_ctz): Likewise.
3823 (int_size_in_bytes): Likewise.
3824 (mem_ref_offset): Return an offset_int rather than a double_int.
3825 (build_type_attribute_qual_variant): Use wide_int interfaces.
3826 (type_hash_eq): Likewise
3827 (tree_int_cst_equal): Likewise.
3828 (tree_int_cst_lt): Delete.
3829 (tree_int_cst_compare): Likewise.
3830 (tree_fits_shwi_p): Use wide_int interfaces.
3831 (tree_fits_uhwi_p): Likewise.
3832 (tree_int_cst_sign_bit): Likewise.
3833 (tree_int_cst_sgn): Likewise.
3834 (tree_int_cst_min_precision): Take a signop rather than a bool.
3835 (simple_cst_equal): Use wide_int interfaces.
3836 (compare_tree_int): Likewise.
3837 (iterative_hash_expr): Likewise.
3838 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
3840 (get_type_static_bounds): Use wide_int interfaces.
3841 (tree_int_cst_elt_check_failed): New.
3842 (build_common_tree_nodes): Reordered to set prec before filling in
3844 (int_cst_value): Check cst_and_fits_in_hwi.
3845 (widest_int_cst_value): Use wide_int interfaces.
3846 (upper_bound_in_type): Likewise.
3847 (lower_bound_in_type): Likewise.
3848 (num_ending_zeros): Likewise.
3849 (drop_tree_overflow): Likewise.
3850 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
3851 (gen_conditions_for_pow_cst_base): Likewise.
3852 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
3853 (group_case_labels_stmt): Use wide-int interfaces.
3854 (verify_gimple_assign_binary): Likewise.
3855 (print_loop): Likewise.
3856 * tree-chrec.c (tree_fold_binomial): Likewise.
3857 * tree-core.h (struct tree_base): Add int_length.
3858 (struct tree_int_cst): Change rep of value.
3859 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
3860 (dr_may_alias_p): Likewise.
3861 (max_stmt_executions_tree): Likewise.
3862 * tree.def (INTEGER_CST): Update comment.
3863 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
3864 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
3865 * tree-dump.c: Include wide-int.h and wide-int-print.h.
3866 (dequeue_and_dump): Use wide-int interfaces.
3867 * tree.h: Include wide-int.h.
3868 (NULL_TREE): Moved to earlier loc in file.
3869 (TREE_INT_CST_ELT_CHECK): New.
3870 (tree_int_cst_elt_check_failed): New.
3872 (TREE_INT_CST): Delete.
3873 (TREE_INT_CST_LOW): Use wide-int interfaces.
3874 (TREE_INT_CST_HIGH): Delete.
3875 (TREE_INT_CST_NUNITS): New.
3876 (TREE_INT_CST_EXT_NUNITS): Likewise.
3877 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
3878 (TREE_INT_CST_ELT): Likewise.
3879 (INT_CST_LT): Delete.
3880 (tree_int_cst_elt_check): New (two forms).
3881 (type_code_size): Update comment.
3882 (make_int_cst_stat, make_int_cst): New.
3883 (tree_to_double_int): Delete.
3884 (double_int_fits_to_tree_p): Delete.
3885 (force_fit_type_double): Delete.
3886 (build_int_cstu): Replace with out-of-line function.
3887 (build_int_cst_wide): Delete.
3888 (tree_int_cst_lt): Define inline.
3889 (tree_int_cst_le): New.
3890 (tree_int_cst_compare): Define inline.
3891 (tree_int_cst_min_precision): Take a signop rather than a bool.
3892 (wi::int_traits <const_tree>): New.
3893 (wi::int_traits <tree>): New.
3894 (wi::extended_tree): New.
3895 (wi::int_traits <wi::extended_tree>): New.
3896 (wi::to_widest): New.
3897 (wi::to_offset): New.
3898 (wi::fits_to_tree_p): New.
3899 (wi::min_value): New.
3900 (wi::max_value): New.
3901 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
3902 (copy_tree_body_r): Likewise.
3903 * tree-object-size.c (compute_object_offset): Likewise.
3904 (addr_object_size): Likewise.
3905 * tree-predcom.c: Include wide-int-print.h.
3906 (struct dref_d): Change type of offset to widest_int.
3907 (dump_dref): Call wide-int printer.
3908 (aff_combination_dr_offset): Use wide-int interfaces.
3909 (determine_offset): Take a widest_int pointer rather than a
3911 (split_data_refs_to_components): Use wide-int interfaces.
3912 (suitable_component_p): Likewise.
3913 (order_drefs): Likewise.
3914 (add_ref_to_chain): Likewise.
3915 (valid_initializer_p): Likewise.
3916 (determine_roots_comp): Likewise.
3917 * tree-pretty-print.c: Include wide-int-print.h.
3918 (dump_generic_node): Use wide-int interfaces.
3919 * tree-sra.c (sra_ipa_modify_expr): Likewise.
3920 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
3921 (move_fixed_address_to_symbol): Likewise.
3922 (move_hint_to_base): Likewise.
3923 (move_pointer_to_base): Likewise.
3924 (move_variant_to_index): Likewise.
3925 (most_expensive_mult_to_index): Likewise.
3926 (addr_to_parts): Likewise.
3927 (copy_ref_info): Likewise.
3928 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
3929 (indirect_refs_may_alias_p): Likewise.
3930 (stmt_kills_ref_p_1): Likewise.
3931 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3932 * tree-ssa-ccp.c: Update comment at top of file. Include
3934 (struct prop_value_d): Change type of mask to widest_int.
3935 (extend_mask): New function.
3936 (dump_lattice_value): Use wide-int interfaces.
3937 (get_default_value): Likewise.
3938 (set_constant_value): Likewise.
3939 (set_value_varying): Likewise.
3940 (valid_lattice_transition): Likewise.
3941 (set_lattice_value): Likewise.
3942 (value_to_double_int): Delete.
3943 (value_to_wide_int): New.
3944 (get_value_from_alignment): Use wide-int interfaces.
3945 (get_value_for_expr): Likewise.
3946 (do_dbg_cnt): Likewise.
3947 (ccp_finalize): Likewise.
3948 (ccp_lattice_meet): Likewise.
3949 (bit_value_unop_1): Use widest_ints rather than double_ints.
3950 (bit_value_binop_1): Likewise.
3951 (bit_value_unop): Use wide-int interfaces.
3952 (bit_value_binop): Likewise.
3953 (bit_value_assume_aligned): Likewise.
3954 (evaluate_stmt): Likewise.
3955 (ccp_fold_stmt): Likewise.
3956 (visit_cond_stmt): Likewise.
3957 (ccp_visit_stmt): Likewise.
3958 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
3959 (constant_pointer_difference): Likewise.
3960 (associate_pointerplus): Likewise.
3961 (combine_conversions): Likewise.
3962 * tree-ssa-loop.h: Include wide-int.h.
3963 (struct tree_niter_desc): Change type of max to widest_int.
3964 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
3965 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
3966 (remove_redundant_iv_tests): Likewise.
3967 (canonicalize_loop_induction_variables): Likewise.
3968 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
3969 (constant_multiple_of): Take a widest_int pointer instead of
3970 a double_int pointer.
3971 (get_computation_aff): Use wide-int interfaces.
3972 (ptr_difference_cost): Likewise.
3973 (difference_cost): Likewise.
3974 (get_loop_invariant_expr_id): Likewise.
3975 (get_computation_cost_at): Likewise.
3976 (iv_elimination_compare_lt): Likewise.
3977 (may_eliminate_iv): Likewise.
3978 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
3979 instead of double_int.
3980 (max_loop_iterations): Likewise.
3981 (max_stmt_executions): Likewise.
3982 (estimated_stmt_executions): Likewise.
3983 * tree-ssa-loop-niter.c: Include wide-int-print.h.
3984 (split_to_var_and_offset): Use wide-int interfaces.
3985 (determine_value_range): Likewise.
3986 (bound_difference_of_offsetted_base): Likewise.
3987 (bounds_add): Take a widest_int instead of a double_int.
3988 (number_of_iterations_ne_max): Use wide-int interfaces.
3989 (number_of_iterations_ne): Likewise.
3990 (number_of_iterations_lt_to_ne): Likewise.
3991 (assert_loop_rolls_lt): Likewise.
3992 (number_of_iterations_lt): Likewise.
3993 (number_of_iterations_le): Likewise.
3994 (number_of_iterations_cond): Likewise.
3995 (number_of_iterations_exit): Likewise.
3996 (finite_loop_p): Likewise.
3997 (derive_constant_upper_bound_assign): Likewise.
3998 (derive_constant_upper_bound): Return a widest_int.
3999 (derive_constant_upper_bound_ops): Likewise.
4000 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
4001 (record_estimate): Take a widest_int rather than a double_int.
4002 (record_nonwrapping_iv): Use wide-int interfaces.
4003 (double_int_cmp): Delete.
4004 (wide_int_cmp): New.
4005 (bound_index): Take a widest_int rather than a double_int.
4006 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
4007 (maybe_lower_iteration_bound): Likewise.
4008 (estimate_numbers_of_iterations_loop): Likewise.
4009 (estimated_loop_iterations): Take a widest_int pointer than than
4010 a double_int pointer.
4011 (estimated_loop_iterations_int): Use wide-int interfaces.
4012 (max_loop_iterations): Take a widest_int pointer than than
4013 a double_int pointer.
4014 (max_loop_iterations_int): Use wide-int interfaces.
4015 (max_stmt_executions): Take a widest_int pointer than than
4016 a double_int pointer.
4017 (estimated_stmt_executions): Likewise.
4018 (n_of_executions_at_most): Use wide-int interfaces.
4019 (scev_probably_wraps_p): Likewise.
4020 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
4022 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
4024 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
4025 double_ints. Adjust for trailing_wide_ints <3> representation.
4026 (set_nonzero_bits): Likewise.
4027 (get_range_info): Return wide_ints rather than double_ints.
4028 Adjust for trailing_wide_ints <3> representation.
4029 (get_nonzero_bits): Likewise.
4030 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
4032 * tree-ssanames.h (struct range_info_def): Replace min, max and
4033 nonzero_bits with a trailing_wide_ints <3>.
4034 (set_range_info): Use wide_int_refs rather than double_ints.
4035 (set_nonzero_bits): Likewise.
4036 (get_range_info): Return wide_ints rather than double_ints.
4037 (get_nonzero_bits): Likewise.
4038 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
4039 * tree-ssa-pre.c (phi_translate_1): Likewise.
4040 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
4041 (acceptable_pow_call): Likewise.
4042 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
4044 (vn_reference_fold_indirect): Likewise.
4045 (vn_reference_maybe_forwprop_address): Likewise.
4046 (valueize_refs_1): Likewise.
4047 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
4048 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
4049 tree_int_cst_lt and tree_int_cst_le.
4050 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
4052 (streamer_alloc_tree): Likewise.
4053 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
4054 (streamer_write_tree_header): Likewise.
4055 (streamer_write_integer_cst): Likewise.
4056 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
4057 (build_constructors): Likewise.
4058 (array_value_type): Likewise.
4059 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
4060 (vect_check_gather): Likewise.
4061 * tree-vect-generic.c (build_replicated_const): Likewise.
4062 (expand_vector_divmod): Likewise.
4063 * tree-vect-loop.c (vect_transform_loop): Likewise.
4064 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
4065 (vect_do_peeling_for_alignment): Likewise.
4066 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
4067 * tree-vrp.c: Include wide-int.h.
4068 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
4069 (extract_range_from_assert): Use wide-int interfaces.
4070 (vrp_int_const_binop): Likewise.
4071 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
4072 double_int pointers.
4073 (ranges_from_anti_range): Use wide-int interfaces.
4074 (quad_int_cmp): Delete.
4075 (quad_int_pair_sort): Likewise.
4076 (extract_range_from_binary_expr_1): Use wide-int interfaces.
4077 (extract_range_from_unary_expr_1): Likewise.
4078 (adjust_range_with_scev): Likewise.
4079 (masked_increment): Take and return wide_ints rather than double_ints.
4080 (register_edge_assert_for_2): Use wide-int interfaces.
4081 (check_array_ref): Likewise.
4082 (search_for_addr_array): Likewise.
4083 (maybe_set_nonzero_bits): Likewise.
4084 (union_ranges): Pass an integer of the correct type instead of
4085 using integer_one_node.
4086 (intersect_ranges): Likewise.
4087 (simplify_truth_ops_using_ranges): Likewise.
4088 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
4089 (range_fits_type_p): Likewise.
4090 (simplify_cond_using_ranges): Likewise. Take a signop rather than
4092 (simplify_conversion_using_ranges): Use wide-int interfaces.
4093 (simplify_float_conversion_using_ranges): Likewise.
4094 (vrp_finalize): Likewise.
4095 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
4096 (gimple_stringops_transform): Likewise.
4097 * varasm.c (decode_addr_const): Likewise.
4098 (const_hash_1): Likewise.
4099 (const_rtx_hash_1): Likewise
4100 (output_constant): Likewise.
4101 (array_size_for_constructor): Likewise.
4102 (output_constructor_regular_field): Likewise.
4103 (output_constructor_bitfield): Likewise.
4104 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
4105 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
4107 * gencheck.c: Define BITS_PER_UNIT.
4110 * wide-int-print.cc: New.
4111 * wide-int-print.h: New.
4113 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4115 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
4117 2014-05-06 Richard Biener <rguenther@suse.de>
4119 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
4120 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
4121 (TODO_verify_all): Adjust.
4122 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
4123 TODO_verify_stmts and TODO_verify_rtl_sharing.
4124 * bb-reorder.c: Likewise.
4125 * cfgexpand.c: Likewise.
4126 * cprop.c: Likewise.
4128 * function.c: Likewise.
4129 * fwprop.c: Likewise.
4131 * gimple-ssa-isolate-paths.c: Likewise.
4132 * gimple-ssa-strength-reduction.c: Likewise.
4133 * ipa-split.c: Likewise.
4134 * loop-init.c: Likewise.
4135 * loop-unroll.c: Likewise.
4136 * lower-subreg.c: Likewise.
4137 * modulo-sched.c: Likewise.
4138 * postreload-gcse.c: Likewise.
4139 * predict.c: Likewise.
4140 * recog.c: Likewise.
4141 * sched-rgn.c: Likewise.
4142 * store-motion.c: Likewise.
4143 * tracer.c: Likewise.
4144 * trans-mem.c: Likewise.
4145 * tree-call-cdce.c: Likewise.
4146 * tree-cfg.c: Likewise.
4147 * tree-cfgcleanup.c: Likewise.
4148 * tree-complex.c: Likewise.
4149 * tree-eh.c: Likewise.
4150 * tree-emutls.c: Likewise.
4151 * tree-if-conv.c: Likewise.
4152 * tree-into-ssa.c: Likewise.
4153 * tree-loop-distribution.c: Likewise.
4154 * tree-object-size.c: Likewise.
4155 * tree-parloops.c: Likewise.
4156 * tree-pass.h: Likewise.
4157 * tree-sra.c: Likewise.
4158 * tree-ssa-ccp.c: Likewise.
4159 * tree-ssa-copy.c: Likewise.
4160 * tree-ssa-copyrename.c: Likewise.
4161 * tree-ssa-dce.c: Likewise.
4162 * tree-ssa-dom.c: Likewise.
4163 * tree-ssa-dse.c: Likewise.
4164 * tree-ssa-forwprop.c: Likewise.
4165 * tree-ssa-ifcombine.c: Likewise.
4166 * tree-ssa-loop-ch.c: Likewise.
4167 * tree-ssa-loop-ivcanon.c: Likewise.
4168 * tree-ssa-loop.c: Likewise.
4169 * tree-ssa-math-opts.c: Likewise.
4170 * tree-ssa-phiopt.c: Likewise.
4171 * tree-ssa-phiprop.c: Likewise.
4172 * tree-ssa-pre.c: Likewise.
4173 * tree-ssa-reassoc.c: Likewise.
4174 * tree-ssa-sink.c: Likewise.
4175 * tree-ssa-strlen.c: Likewise.
4176 * tree-ssa-tail-merge.c: Likewise.
4177 * tree-ssa-uncprop.c: Likewise.
4178 * tree-switch-conversion.c: Likewise.
4179 * tree-tailcall.c: Likewise.
4180 * tree-vect-generic.c: Likewise.
4181 * tree-vectorizer.c: Likewise.
4182 * tree-vrp.c: Likewise.
4184 * var-tracking.c: Likewise.
4185 * bt-load.c: Likewise.
4186 * cfgcleanup.c: Likewise.
4187 * combine-stack-adj.c: Likewise.
4188 * combine.c: Likewise.
4189 * compare-elim.c: Likewise.
4190 * config/epiphany/resolve-sw-modes.c: Likewise.
4191 * config/i386/i386.c: Likewise.
4192 * config/mips/mips.c: Likewise.
4193 * config/s390/s390.c: Likewise.
4194 * config/sh/sh_treg_combine.cc: Likewise.
4195 * config/sparc/sparc.c: Likewise.
4198 * final.c: Likewise.
4199 * ifcvt.c: Likewise.
4200 * mode-switching.c: Likewise.
4201 * passes.c: Likewise.
4202 * postreload.c: Likewise.
4204 * reg-stack.c: Likewise.
4205 * regcprop.c: Likewise.
4206 * regrename.c: Likewise.
4209 2014-05-06 Richard Biener <rguenther@suse.de>
4212 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
4213 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
4215 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
4218 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
4220 2014-05-05 Radovan Obradovic <robradovic@mips.com>
4221 Tom de Vries <tom@codesourcery.com>
4223 * target.def (call_fusage_contains_non_callee_clobbers): New
4225 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
4227 (@node Miscellaneous Register Hooks): New node.
4228 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
4229 * doc/tm.texi: Regenerate.
4231 2014-05-05 Marek Polacek <polacek@redhat.com>
4234 * opts.c (common_handle_option): Call error_at instead of warning_at.
4236 2014-05-05 Richard Biener <rguenther@suse.de>
4238 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
4239 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
4240 under the TODO_verify_il umbrella.
4242 2014-05-05 Richard Biener <rguenther@suse.de>
4244 * passes.c (execute_function_todo): Move TODO_verify_flow under
4245 the TODO_verify_ul umbrella.
4247 2014-05-05 Richard Biener <rguenther@suse.de>
4250 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
4251 X & CST away from a CST that is the mask of a mode.
4253 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4255 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
4256 int argument to enum machine_mode.
4257 (picochip_class_max_nregs): Ditto.
4258 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
4259 (picochip_class_max_nregs): Ditto.
4261 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4263 * target.def: Add new target hook.
4264 * doc/tm.texi: Regenerate.
4265 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
4266 * targhooks.c (default_keep_leaf_when_profiled): New function.
4268 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
4269 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
4271 2014-05-05 Bin Cheng <bin.cheng@arm.com>
4273 PR tree-optimization/60363
4274 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
4275 (copy_phi_args): New parameters. Call get_value_locus_in_path.
4276 (update_destination_phis): New parameter.
4277 (create_edge_and_update_destination_phis): Ditto.
4278 (ssa_fix_duplicate_block_edges): Pass new arguments.
4279 (thread_single_edge): Ditto.
4281 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
4283 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
4284 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
4285 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
4286 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
4287 Use RS6000_BTM_HARD_FLOAT.
4288 (BU_MISC_2): Likewise.
4289 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
4290 RS6000_BTM_HARD_FLOAT.
4291 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
4293 (rs6000_invalid_builtin): Add hard floating builtin support.
4294 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
4295 hard float builtins.
4296 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
4298 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4300 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
4301 Add missing function* argument.
4303 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4305 * lra-constraints.c (valid_address_p): Move earlier in file.
4306 Add a constraint argument to the address_info version.
4307 (satisfies_memory_constraint_p): New function.
4308 (satisfies_address_constraint_p): Likewise.
4309 (process_alt_operands, curr_insn_transform): Use them.
4310 (process_address): Pass the constraint to valid_address_p when
4311 checking address operands.
4313 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4315 * config/mips/mips.c (mips_isa_rev): New variable.
4316 (mips_set_architecture): Set it.
4317 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
4319 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
4320 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
4321 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
4322 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
4323 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
4324 conditions in terms of mips_isa_rev.
4325 (mips_isa_rev): Declare.
4327 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4329 * config/sh/sh-mem.cc: Use tabs instead of spaces.
4330 (prob_unlikely, prob_likely): Make variables const.
4332 2014-05-03 Denis Chertykov <chertykov@gmail.com>
4334 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
4336 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4338 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
4340 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4342 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
4343 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
4344 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
4346 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
4347 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
4349 Replace usage of ROUND_REG with sh_round_reg.
4350 Use CEIL instead of ROUND_ADVANCE.
4352 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4355 * config/sh/sh.c: Include stdlib headers before everything else.
4357 2014-05-02 Jakub Jelinek <jakub@redhat.com>
4359 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
4360 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
4361 (gimplify_adjust_omp_clauses): Simd region is never
4362 directly nested in combined parallel. Instead, for linear
4363 with copyin/copyout, if in combined for simd loop, make decl
4364 firstprivate/lastprivate on OMP_FOR.
4365 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
4366 expand_omp_for_static_chunk): When setting endvar, also set
4367 fd->loop.v to the same value.
4369 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4371 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
4373 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
4375 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
4378 2014-05-02 Marek Polacek <polacek@redhat.com>
4380 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
4382 2014-05-02 Kito Cheng <kito@0xlab.org>
4384 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
4385 to a C expression marco.
4386 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
4387 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
4388 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
4389 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
4390 HONOR_REG_ALLOC_ORDER.
4391 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
4393 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4395 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
4397 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4399 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
4401 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
4403 * tree-if-conv.c (is_cond_scalar_reduction): New function.
4404 (convert_scalar_cond_reduction): Likewise.
4405 (predicate_scalar_phi): Add recognition and transformation
4406 of simple conditioanl reduction to be vectorizable.
4408 2014-05-01 Marek Polacek <polacek@redhat.com>
4411 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
4413 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
4415 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
4416 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
4417 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
4418 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
4419 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
4420 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
4421 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
4422 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
4424 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
4426 * config/arc/arc.opt (mlra): Move comment above option name
4427 to avoid mis-parsing as language options.
4429 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4431 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
4432 * config/sol2.h: ... here.
4433 * config/sol2-10.h: Remove.
4435 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
4436 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
4437 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
4438 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
4439 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
4440 * config/sol2.h: ... here.
4441 (SECTION_NAME_FORMAT): Don't redefine.
4442 (STARTFILE_ARCH32_SPEC): Rename to ...
4443 (STARTFILE_ARCH_SPEC): ... this.
4444 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
4445 * config/sparc/sol2.h: ... here.
4446 (SECTION_NAME_FORMAT): Don't undef.
4447 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
4448 (SUBTARGET_EXTRA_SPECS): Remove.
4449 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
4451 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
4452 (MD_STARTFILE_PREFIX): Remove.
4453 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
4454 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
4455 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
4456 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
4457 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
4458 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
4459 * config/i386/sol2.h: ... here.
4460 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
4461 * config/i386/sol2-bi.h: Remove.
4462 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
4463 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
4465 * config/i386/t-sol2-64: Rename to ...
4466 * config/i386/t-sol2: ... this.
4467 * config/sparc/t-sol2-64: Rename to ...
4468 * config/sparc/t-sol2: ... this.
4470 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
4471 sol2_tm_file_head, sol2_tm_file_tail.
4472 Include ${cpu_type}/sol2.h before sol2.h.
4474 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
4475 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
4476 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
4477 Reflect i386/t-sol2-64 renaming.
4478 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
4479 Reflect sparc/t-sol2-64 renaming.
4481 2014-04-30 Richard Biener <rguenther@suse.de>
4483 * passes.c (execute_function_todo): Move TODO_verify_stmts
4484 and TODO_verify_ssa under the TODO_verify_il umbrella.
4485 * tree-ssa.h (verify_ssa): Adjust prototype.
4486 * tree-ssa.c (verify_ssa): Add parameter to tell whether
4487 we should verify SSA operands.
4488 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
4489 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
4490 whether we should verify whether not throwing stmts have EH info.
4491 * graphite-scop-detection.c (create_sese_edges): Adjust.
4492 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
4493 * tree-eh.c (lower_try_finally_switch): Do not add the
4494 default case label twice.
4496 2014-04-30 Marek Polacek <polacek@redhat.com>
4498 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
4499 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
4500 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
4501 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
4503 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
4505 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
4506 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
4507 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
4508 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
4509 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
4510 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
4511 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
4512 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
4514 2014-04-29 David Malcolm <dmalcolm@redhat.com>
4516 * tree-cfg.c (dump_function_to_file): Dump the return type of
4517 functions, in a line to itself before the function body, mimicking
4518 the layout of a C function.
4520 2014-04-29 Jakub Jelinek <jakub@redhat.com>
4522 PR tree-optimization/60971
4523 * tree-tailcall.c (process_assignment): Reject conversions which
4526 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4528 * calls.c (initialize_argument_information): Always treat
4529 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4530 (expand_call): Likewise.
4531 (emit_library_call_calue_1): Likewise.
4532 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4533 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4536 2014-04-29 Nick Clifton <nickc@redhat.com>
4538 * config/msp430/msp430.md (umulsidi): Fix typo.
4539 (mulhisi3): Enable even inside interrupt handlers.
4540 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4541 bigger return address pushed in large mode.
4543 2014-04-29 Nick Clifton <nickc@redhat.com>
4545 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4546 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4548 * config/m32r/m32r.c (init_reg_tables): Likewise.
4549 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4550 enum to hold the modes.
4552 2014-04-29 Richard Biener <rguenther@suse.de>
4554 * dominance.c (free_dominance_info): Add overload with
4556 (dom_info_state): Likewise.
4557 (dom_info_available_p): Likewise.
4558 * basic-block.h (free_dominance_info, dom_info_state,
4559 dom_info_available_p): Declare overloads.
4560 * passes.c (execute_function_todo): Verify that verifiers
4561 don't change dominator info state. Drop dominator info
4562 for IPA pass invocations.
4563 * cgraph.c (release_function_body): Restore asserts that
4564 dominator information is released.
4566 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4568 * doc/invoke.texi: Fix typo.
4569 * tree-vrp.c: Fix typos.
4570 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4572 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4574 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4576 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4578 * config/aarch64/aarch64-builtins.c
4579 (aarch64_types_storestruct_lane_qualifiers): New.
4580 (TYPES_STORESTRUCT_LANE): Likewise.
4581 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4582 (st3_lane): Likewise.
4583 (st4_lane): Likewise.
4584 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4585 (vec_store_lanesci_lane<mode>): Likewise.
4586 (vec_store_lanesxi_lane<mode>): Likewise.
4587 (aarch64_st2_lane<VQ:mode>): Likewise.
4588 (aarch64_st3_lane<VQ:mode>): Likewise.
4589 (aarch64_st4_lane<VQ:mode>): Likewise.
4590 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
4591 * config/aarch64/arm_neon.h
4592 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
4593 use new macro arguments.
4594 (__ST3_LANE_FUNC): Likewise.
4595 (__ST4_LANE_FUNC): Likewise.
4596 * config/aarch64/iterators.md (V_TWO_ELEM): New.
4597 (V_THREE_ELEM): Likewise.
4598 (V_FOUR_ELEM): Likewise.
4600 2014-04-28 David Malcolm <dmalcolm@redhat.com>
4602 * doc/gimple.texi: Replace the description of the now-defunct
4603 union gimple_statement_d with a diagram showing the
4604 gimple_statement_base class hierarchy and its relationships to
4605 the GSS_ and GIMPLE_ enums.
4607 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4609 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4610 * config/aarch64/aarch64.c
4611 (aarch64_cannot_change_mode_class): Weaken conditions.
4612 (aarch64_modes_tieable_p): New.
4613 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
4615 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
4617 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4618 (loadsync_<mode>): Change mode.
4619 (load_quadpti, store_quadpti): New.
4620 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4621 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4623 2014-04-28 Martin Jambor <mjambor@suse.cz>
4625 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
4626 same alias type as the original statement.
4627 (subreplacement_assignment_data): New type.
4628 (handle_unscalarized_data_in_subtree): New type of parameter,
4629 generate new memory accesses with same alias type as the original
4631 (load_assign_lhs_subreplacements): Likewise.
4632 (sra_modify_constructor_assign): Generate new memory accesses with
4633 same alias type as the original statement.
4635 2014-04-28 Richard Biener <rguenther@suse.de>
4637 * tree-pass.h (TODO_verify_il): Define.
4638 (TODO_verify_all): Complete properly.
4639 * passes.c (execute_function_todo): Move existing loop-closed
4640 SSA verification under TODO_verify_il.
4641 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
4642 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
4643 Fix tree sharing issue.
4645 2014-04-28 Richard Biener <rguenther@suse.de>
4648 * builtins.def (DEF_C11_BUILTIN): Add.
4649 (BUILT_IN_ALIGNED_ALLOC): Likewise.
4650 * coretypes.h (enum function_class): Add function_c11_misc.
4651 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
4652 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
4653 (call_may_clobber_ref_p_1): Likewise.
4654 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
4655 (mark_all_reaching_defs_necessary_1): Likewise.
4656 (propagate_necessity): Likewise.
4657 (eliminate_unnecessary_stmts): Likewise.
4658 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
4660 2014-04-28 Richard Biener <rguenther@suse.de>
4662 * tree-vrp.c (vrp_var_may_overflow): Remove.
4663 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
4664 with overflow immediately bump to one before that value and
4665 let iteration figure out overflow status.
4667 2014-04-28 Richard Biener <rguenther@suse.de>
4669 * configure.ac: Do valgrind header checks unconditionally.
4670 Add --enable-valgrind-annotations.
4671 * system.h: Guard valgrind header inclusion with
4672 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
4673 * alloc-pool.c (pool_alloc, pool_free): Use
4674 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
4675 to guard possibly dead code.
4676 * config.in: Regenerated.
4677 * configure: Likewise.
4679 2014-04-28 Jeff Law <law@redhat.com>
4681 PR tree-optimization/60902
4682 * tree-ssa-threadedge.c
4683 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
4684 over real defs when invalidating outputs from statements that do not
4685 produce useful outputs for threading.
4687 2014-04-28 Richard Biener <rguenther@suse.de>
4689 PR tree-optimization/60979
4690 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
4691 SCOPs that end in a block with a successor with abnormal
4694 2014-04-28 Richard Biener <rguenther@suse.de>
4696 * tree-pass.h (execute_pass_list): Adjust prototype.
4697 * passes.c (pass_manager::execute_early_local_passes): Adjust.
4698 (do_per_function): Change callback signature, push all actual
4699 work to the callbals.
4700 (do_per_function_toporder): Likewise.
4701 (execute_function_dump): Adjust.
4702 (execute_function_todo): Likewise.
4703 (clear_last_verified): Likewise.
4704 (verify_curr_properties): Likewise.
4705 (update_properties_after_pass): Likewise.
4706 (execute_pass_list_1): Split out from ...
4707 (execute_pass_list): ... here. Adjust.
4708 (execute_ipa_pass_list): Likewise.
4709 * cgraphunit.c (cgraph_add_new_function): Adjust.
4710 (analyze_function): Likewise.
4711 (expand_function): Likewise.
4712 * cgraph.c (release_function_body): Free dominance info
4713 here instead of asserting it was magically freed elsewhere.
4715 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
4717 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
4718 * configure: Regenerate.
4719 * config/sparc/sparc.opt (muser-mode): New option.
4720 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
4722 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
4723 * doc/invoke.texi (SPARC options): Document -muser-mode.
4725 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
4727 * cselib.c (find_slot_memmode): Delete.
4728 (cselib_hasher): Change compare_type to a struct.
4729 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
4731 (preserve_constants_and_equivs): Adjust for new compare_type.
4732 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
4733 (wrap_constant): Delete.
4734 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
4736 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
4738 * doc/install.texi (Building with profile feedback): Remove
4741 2014-04-26 Tom de Vries <tom@codesourcery.com>
4743 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
4746 2014-04-25 Cary Coutant <ccoutant@google.com>
4749 * dwarf2out.c (should_move_die_to_comdat): A type definition
4750 can contain a subprogram definition, but don't move it to a
4752 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
4753 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
4755 (clone_tree_hash): Rename to...
4756 (clone_tree_partial): ...this; change callers. Copy
4757 DW_TAG_subprogram DIEs as declarations.
4758 (copy_decls_walk): Don't copy children of a declaration into a
4761 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
4764 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
4767 2014-04-25 Jiong Wang <jiong.wang@arm.com>
4769 * config/arm/predicates.md (call_insn_operand): Add long_call check.
4770 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
4772 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
4775 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4777 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
4779 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4781 PR tree-optimization/60930
4782 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
4783 creating a multiply candidate by folding two constant
4784 multiplicands when the result overflows.
4786 2014-04-25 Jakub Jelinek <jakub@redhat.com>
4788 PR tree-optimization/60960
4789 * tree-vect-generic.c (expand_vector_operation): Only call
4790 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
4792 2014-04-25 Tom de Vries <tom@codesourcery.com>
4794 * expr.c (clobber_reg_mode): New function.
4795 * expr.h (clobber_reg): New function.
4797 2014-04-25 Tom de Vries <tom@codesourcery.com>
4799 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
4802 2014-04-25 Radovan Obradovic <robradovic@mips.com>
4803 Tom de Vries <tom@codesourcery.com>
4805 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
4807 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
4808 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
4809 new argument to find_all_hard_reg_sets call.
4811 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4813 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
4814 Use HOST_WIDE_INT_C for mask literal.
4815 (aarch_rev16_shleft_mask_imm_p): Likewise.
4817 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
4820 * config/sparc/sparc.md (ashlsi3_extend): Delete.
4822 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
4824 PR preprocessor/56540
4825 * config/i386/i386-c.c (ix86_target_macros): Define
4826 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
4828 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4830 * configure.ac (tga_func): Remove.
4831 (LIB_TLS_SPEC): Remove.
4832 * configure: Regenerate.
4833 * config.in: Regenerate.
4834 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
4836 2014-04-25 Richard Biener <rguenther@suse.de>
4839 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
4840 call stmt use/clobber sets during stmt walk instead of
4841 walking the possibly incomplete set of caller edges.
4843 2014-04-25 Richard Biener <rguenther@suse.de>
4846 * passes.c (apply_ipa_transforms): Inline into only caller ...
4847 (execute_one_pass): ... here. Properly bring in function
4848 bodies for nodes we want to apply IPA transforms to.
4850 2014-04-24 Cong Hou <congh@google.com>
4852 PR tree-optimization/60896
4853 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
4854 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
4855 (vect_mark_pattern_stmts): Set the def type of all statements in
4856 PATTERN_DEF_SEQ as vect_internal_def.
4858 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
4860 * doc/extend.texi (PowerPC Built-in Functions): Document new
4861 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
4862 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
4864 * config/rs6000/predicates.md (const_0_to_3_operand): New
4865 predicate to match 0..3 integer constants.
4867 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
4868 to support adding miscellaneous builtin functions.
4869 (BU_DFP_MISC_2): Likewise.
4870 (BU_P7_MISC_1): Likewise.
4871 (BU_P7_MISC_2): Likewise.
4872 (BU_P8V_MISC_3): Likewise.
4873 (BU_MISC_1): Likewise.
4874 (BU_MISC_2): Likewise.
4875 (DIVWE): Add extended divide builtin functions.
4878 (DIVWEUO): Likewise.
4882 (DIVDEUO): Likewise.
4883 (DXEX): Add decimal floating-point builtin functions.
4886 (DDEDPDQ): Likewise.
4888 (DENBCDQ): Likewise.
4895 (CDTBCD): Add new BCD builtin functions.
4899 (BCDADD_LT): Likewise.
4900 (BCDADD_EQ): Likewise.
4901 (BCDADD_GT): Likewise.
4902 (BCDADD_OV): Likewise.
4904 (BCDSUB_LT): Likewise.
4905 (BCDSUB_EQ): Likewise.
4906 (BCDSUB_GT): Likewise.
4907 (BCDSUB_OV): Likewise.
4908 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
4909 (UNPACK_TD): Likewise.
4910 (PACK_TF): Likewise.
4911 (UNPACK_TF): Likewise.
4912 (UNPACK_TF_0): Likewise.
4913 (UNPACK_TF_1): Likewise.
4914 (PACK_V1TI): Likewise.
4915 (UNPACK_V1TI): Likewise.
4917 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
4918 support for decimal floating point builtin functions.
4919 (rs6000_expand_ternop_builtin): Add checks for the new builtin
4920 functions that take constant arguments.
4921 (rs6000_invalid_builtin): Add decimal floating point builtin support.
4922 (rs6000_init_builtins): Setup long double, _Decimal64, and
4923 _Decimal128 types for new builtin functions.
4924 (builtin_function_type): Set the unsigned flags appropriately for
4925 the new builtin functions.
4926 (rs6000_opt_masks): Add support for decimal floating point builtin
4929 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
4930 floating point builtin functions.
4931 (RS6000_BTM_COMMON): Likewise.
4932 (RS6000_BTI_long_double): Likewise.
4933 (RS6000_BTI_dfloat64): Likewise.
4934 (RS6000_BTI_dfloat128): Likewise.
4935 (long_double_type_internal_node): Likewise.
4936 (dfloat64_type_internal_node): Likewise.
4937 (dfloat128_type_internal_node): Likewise.
4939 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
4940 2.07 bcd arithmetic instructions.
4941 (UNSPEC_BCDSUB): Likewise.
4942 (UNSPEC_BCD_OVERFLOW): Likewise.
4943 (UNSPEC_BCD_ADD_SUB): Likewise.
4944 (bcd_add_sub): Likewise.
4945 (BCD_TEST): Likewise.
4946 (bcd<bcd_add_sub>): Likewise.
4947 (bcd<bcd_add_sub>_test): Likewise.
4948 (bcd<bcd_add_sub>_test2): Likewise.
4949 (bcd<bcd_add_sub>_<code>): Likewise.
4950 (peephole2 for combined bcd ops): Likewise.
4952 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
4953 decimal floating point builtin functions.
4954 (UNSPEC_DENBCD): Likewise.
4955 (UNSPEC_DXEX): Likewise.
4956 (UNSPEC_DIEX): Likewise.
4957 (UNSPEC_DSCLI): Likewise.
4958 (UNSPEC_DSCRI): Likewise.
4959 (D64_D128): Likewise.
4960 (dfp_suffix): Likewise.
4961 (dfp_ddedpd_<mode>): Likewise.
4962 (dfp_denbcd_<mode>): Likewise.
4963 (dfp_dxex_<mode>): Likewise.
4964 (dfp_diex_<mode>): Likewise.
4965 (dfp_dscli_<mode>): Likewise.
4966 (dfp_dscri_<mode>): Likewise.
4968 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
4970 (UNSPEC_CDTBCD): Likewise.
4971 (UNSPEC_CBCDTD): Likewise.
4972 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
4973 (UNSPEC_DIVEO): Likewise.
4974 (UNSPEC_DIVEU): Likewise.
4975 (UNSPEC_DIVEUO): Likewise.
4976 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
4977 pack/unpack 128-bit types.
4978 (UNSPEC_PACK_128BIT): Likewise.
4979 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
4980 (udiv<mode>3): Use idiv_ldiv mode attribute.
4981 (div<mode>3): Likewise.
4982 (addg6s): Add new BCD builtin functions.
4985 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
4986 (div_extend): Likewise.
4987 (div<div_extend>_<mode>"): Likewise.
4988 (FP128_64): Add support for new builtin functions to pack/unpack
4990 (unpack<mode>): Likewise.
4991 (unpacktf_0): Likewise.
4992 (unpacktf_1): Likewise.
4993 (unpack<mode>_dm): Likewise.
4994 (unpack<mode>_nodm): Likewise.
4995 (pack<mode>): Likewise.
4996 (unpackv1ti): Likewise.
4997 (packv1ti): Likewise.
4999 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
5001 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
5004 2014-04-24 Jakub Jelinek <jakub@redhat.com>
5006 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
5007 * gimplify.c (omp_is_private): Change last argument's type to int.
5008 Only diagnose lastprivate if the simd argument is 1, only diagnose
5009 linear if the simd argument is 2.
5010 (gimplify_omp_for): Adjust omp_is_private callers. When adding
5011 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
5012 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
5013 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
5014 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
5015 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5016 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
5017 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5018 * tree-nested.c (convert_nonlocal_omp_clauses,
5019 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
5021 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
5024 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
5027 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
5029 * flag-types.h (enum ivar_visibility): Add.
5031 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
5033 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
5034 function * argument.
5036 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
5038 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
5040 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5041 Tom de Vries <tom@codesourcery.com>
5043 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
5044 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
5046 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
5047 * emit-rtl.c (try_split): Same.
5049 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5050 Tom de Vries <tom@codesourcery.com>
5052 * common.opt (fuse-caller-save): New option.
5054 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
5056 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
5057 elements for big-endian.
5059 2014-04-24 Richard Biener <rguenther@suse.de>
5061 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
5062 during TER and instead use the sepops interface for expanding
5063 non-GIMPLE_SINGLE_RHS.
5065 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5067 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
5068 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
5070 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5072 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
5073 assembler 64-bit option.
5074 * configure: Regenerate.
5076 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5078 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
5079 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
5080 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
5081 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
5082 (TARGET_CRYPTO): Take TARGET_SIMD into account.
5084 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5086 * config/aarch64/aarch64-builtins.c
5087 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
5088 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
5089 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
5090 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
5092 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
5093 (Vrevsuff): New mode attribute.
5095 2014-04-24 Terry Guo <terry.guo@arm.com>
5097 * config/arm/arm.h (machine_function): Define variable
5098 after_arm_reorg here.
5099 * config/arm/arm.c (after_arm_reorg): Remove the definition.
5100 (arm_split_constant): Update the way to access variable
5103 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
5105 2014-04-23 Tom de Vries <tom@codesourcery.com>
5107 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
5109 2014-04-23 David Malcolm <dmalcolm@redhat.com>
5111 * is-a.h: Update comments to reflect the following changes to the
5112 "pointerness" of the API, making the template parameter match the
5113 return type, allowing use of is-a.h with typedefs of pointers.
5114 (is_a_helper::cast): Return a T rather then a pointer to a T, so
5115 that the return type matches the parameter to the is_a_helper.
5117 (dyn_cast): Likewise.
5119 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
5120 pointer from the is-a.h API.
5122 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
5123 (is_a_helper <cgraph_node *>::test): ...this, matching change to
5125 (is_a_helper <varpool_node>::test): Likewise, convert to...
5126 (is_a_helper <varpool_node *>::test): ...this.
5128 (varpool_first_variable): Update for removal of implicit pointer
5129 from the is-a.h API.
5130 (varpool_next_variable): Likewise.
5131 (varpool_first_static_initializer): Likewise.
5132 (varpool_next_static_initializer): Likewise.
5133 (varpool_first_defined_variable): Likewise.
5134 (varpool_next_defined_variable): Likewise.
5135 (cgraph_first_defined_function): Likewise.
5136 (cgraph_next_defined_function): Likewise.
5137 (cgraph_first_function): Likewise.
5138 (cgraph_next_function): Likewise.
5139 (cgraph_first_function_with_gimple_body): Likewise.
5140 (cgraph_next_function_with_gimple_body): Likewise.
5141 (cgraph_alias_target): Likewise.
5142 (varpool_alias_target): Likewise.
5143 (cgraph_function_or_thunk_node): Likewise.
5144 (varpool_variable_node): Likewise.
5145 (symtab_real_symbol_p): Likewise.
5146 * cgraphunit.c (referred_to_p): Likewise.
5147 (analyze_functions): Likewise.
5148 (handle_alias_pairs): Likewise.
5149 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
5150 * gimple-ssa.h (gimple_vuse_op): Likewise.
5151 (gimple_vdef_op): Likewise.
5152 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
5153 * gimple.c (gimple_build_asm_1): Likewise.
5154 (gimple_build_try): Likewise.
5155 (gimple_build_resx): Likewise.
5156 (gimple_build_eh_dispatch): Likewise.
5157 (gimple_build_omp_for): Likewise.
5158 (gimple_omp_for_set_clauses): Likewise.
5160 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
5161 (is_a_helper <gimple_statement_asm *>::test): ...this.
5162 (is_a_helper <gimple_statement_bind>::test): Convert to...
5163 (is_a_helper <gimple_statement_bind *>::test): ...this.
5164 (is_a_helper <gimple_statement_call>::test): Convert to...
5165 (is_a_helper <gimple_statement_call *>::test): ...this.
5166 (is_a_helper <gimple_statement_catch>::test): Convert to...
5167 (is_a_helper <gimple_statement_catch *>::test): ...this.
5168 (is_a_helper <gimple_statement_resx>::test): Convert to...
5169 (is_a_helper <gimple_statement_resx *>::test): ...this.
5170 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
5171 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
5172 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
5173 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
5174 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
5175 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
5176 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
5177 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
5178 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
5179 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
5180 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
5181 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
5182 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
5183 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
5184 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
5185 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
5186 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
5187 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
5188 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
5189 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
5190 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
5191 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
5192 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
5193 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
5194 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
5195 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
5196 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
5197 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
5198 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
5199 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
5200 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
5201 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
5202 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
5203 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
5204 (is_a_helper <gimple_statement_phi>::test): Convert to...
5205 (is_a_helper <gimple_statement_phi *>::test): ...this.
5206 (is_a_helper <gimple_statement_transaction>::test): Convert to...
5207 (is_a_helper <gimple_statement_transaction *>::test): ...this.
5208 (is_a_helper <gimple_statement_try>::test): Convert to...
5209 (is_a_helper <gimple_statement_try *>::test): ...this.
5210 (is_a_helper <gimple_statement_wce>::test): Convert to...
5211 (is_a_helper <gimple_statement_wce *>::test): ...this.
5212 (is_a_helper <const gimple_statement_asm>::test): Convert to...
5213 (is_a_helper <const gimple_statement_asm *>::test): ...this.
5214 (is_a_helper <const gimple_statement_bind>::test): Convert to...
5215 (is_a_helper <const gimple_statement_bind *>::test): ...this.
5216 (is_a_helper <const gimple_statement_call>::test): Convert to...
5217 (is_a_helper <const gimple_statement_call *>::test): ...this.
5218 (is_a_helper <const gimple_statement_catch>::test): Convert to...
5219 (is_a_helper <const gimple_statement_catch *>::test): ...this.
5220 (is_a_helper <const gimple_statement_resx>::test): Convert to...
5221 (is_a_helper <const gimple_statement_resx *>::test): ...this.
5222 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
5223 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
5224 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
5225 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
5226 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
5228 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
5230 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
5232 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
5234 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
5235 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
5236 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
5238 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
5239 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
5241 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
5242 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
5243 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
5244 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
5245 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
5246 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
5248 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
5249 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
5250 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
5251 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
5253 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
5254 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
5255 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
5256 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
5257 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
5258 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
5259 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
5260 (is_a_helper <const gimple_statement_phi>::test): Convert to...
5261 (is_a_helper <const gimple_statement_phi *>::test): ...this.
5262 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
5263 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
5264 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
5265 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
5266 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
5267 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
5268 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
5270 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
5272 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
5273 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
5275 (gimple_use_ops): Update for removal of implicit pointer from the
5277 (gimple_set_use_ops): Likewise.
5278 (gimple_vuse): Likewise.
5279 (gimple_vdef): Likewise.
5280 (gimple_vuse_ptr): Likewise.
5281 (gimple_vdef_ptr): Likewise.
5282 (gimple_set_vuse): Likewise.
5283 (gimple_set_vdef): Likewise.
5284 (gimple_omp_return_set_lhs): Likewise.
5285 (gimple_omp_return_lhs): Likewise.
5286 (gimple_omp_return_lhs_ptr): Likewise.
5287 (gimple_call_fntype): Likewise.
5288 (gimple_call_set_fntype): Likewise.
5289 (gimple_call_set_internal_fn): Likewise.
5290 (gimple_call_use_set): Likewise.
5291 (gimple_call_clobber_set): Likewise.
5292 (gimple_bind_vars): Likewise.
5293 (gimple_bind_set_vars): Likewise.
5294 (gimple_bind_body_ptr): Likewise.
5295 (gimple_bind_set_body): Likewise.
5296 (gimple_bind_add_stmt): Likewise.
5297 (gimple_bind_block): Likewise.
5298 (gimple_bind_set_block): Likewise.
5299 (gimple_asm_ninputs): Likewise.
5300 (gimple_asm_noutputs): Likewise.
5301 (gimple_asm_nclobbers): Likewise.
5302 (gimple_asm_nlabels): Likewise.
5303 (gimple_asm_input_op): Likewise.
5304 (gimple_asm_input_op_ptr): Likewise.
5305 (gimple_asm_output_op): Likewise.
5306 (gimple_asm_output_op_ptr): Likewise.
5307 (gimple_asm_set_output_op): Likewise.
5308 (gimple_asm_clobber_op): Likewise.
5309 (gimple_asm_set_clobber_op): Likewise.
5310 (gimple_asm_label_op): Likewise.
5311 (gimple_asm_set_label_op): Likewise.
5312 (gimple_asm_string): Likewise.
5313 (gimple_catch_types): Likewise.
5314 (gimple_catch_types_ptr): Likewise.
5315 (gimple_catch_handler_ptr): Likewise.
5316 (gimple_catch_set_types): Likewise.
5317 (gimple_catch_set_handler): Likewise.
5318 (gimple_eh_filter_types): Likewise.
5319 (gimple_eh_filter_types_ptr): Likewise.
5320 (gimple_eh_filter_failure_ptr): Likewise.
5321 (gimple_eh_filter_set_types): Likewise.
5322 (gimple_eh_filter_set_failure): Likewise.
5323 (gimple_eh_must_not_throw_fndecl): Likewise.
5324 (gimple_eh_must_not_throw_set_fndecl): Likewise.
5325 (gimple_eh_else_n_body_ptr): Likewise.
5326 (gimple_eh_else_e_body_ptr): Likewise.
5327 (gimple_eh_else_set_n_body): Likewise.
5328 (gimple_eh_else_set_e_body): Likewise.
5329 (gimple_try_eval_ptr): Likewise.
5330 (gimple_try_cleanup_ptr): Likewise.
5331 (gimple_try_set_eval): Likewise.
5332 (gimple_try_set_cleanup): Likewise.
5333 (gimple_wce_cleanup_ptr): Likewise.
5334 (gimple_wce_set_cleanup): Likewise.
5335 (gimple_phi_capacity): Likewise.
5336 (gimple_phi_num_args): Likewise.
5337 (gimple_phi_result): Likewise.
5338 (gimple_phi_result_ptr): Likewise.
5339 (gimple_phi_set_result): Likewise.
5340 (gimple_phi_arg): Likewise.
5341 (gimple_phi_set_arg): Likewise.
5342 (gimple_resx_region): Likewise.
5343 (gimple_resx_set_region): Likewise.
5344 (gimple_eh_dispatch_region): Likewise.
5345 (gimple_eh_dispatch_set_region): Likewise.
5346 (gimple_omp_critical_name): Likewise.
5347 (gimple_omp_critical_name_ptr): Likewise.
5348 (gimple_omp_critical_set_name): Likewise.
5349 (gimple_omp_for_clauses): Likewise.
5350 (gimple_omp_for_clauses_ptr): Likewise.
5351 (gimple_omp_for_set_clauses): Likewise.
5352 (gimple_omp_for_collapse): Likewise.
5353 (gimple_omp_for_index): Likewise.
5354 (gimple_omp_for_index_ptr): Likewise.
5355 (gimple_omp_for_set_index): Likewise.
5356 (gimple_omp_for_initial): Likewise.
5357 (gimple_omp_for_initial_ptr): Likewise.
5358 (gimple_omp_for_set_initial): Likewise.
5359 (gimple_omp_for_final): Likewise.
5360 (gimple_omp_for_final_ptr): Likewise.
5361 (gimple_omp_for_set_final): Likewise.
5362 (gimple_omp_for_incr): Likewise.
5363 (gimple_omp_for_incr_ptr): Likewise.
5364 (gimple_omp_for_set_incr): Likewise.
5365 (gimple_omp_for_pre_body_ptr): Likewise.
5366 (gimple_omp_for_set_pre_body): Likewise.
5367 (gimple_omp_parallel_clauses): Likewise.
5368 (gimple_omp_parallel_clauses_ptr): Likewise.
5369 (gimple_omp_parallel_set_clauses): Likewise.
5370 (gimple_omp_parallel_child_fn): Likewise.
5371 (gimple_omp_parallel_child_fn_ptr): Likewise.
5372 (gimple_omp_parallel_set_child_fn): Likewise.
5373 (gimple_omp_parallel_data_arg): Likewise.
5374 (gimple_omp_parallel_data_arg_ptr): Likewise.
5375 (gimple_omp_parallel_set_data_arg): Likewise.
5376 (gimple_omp_task_clauses): Likewise.
5377 (gimple_omp_task_clauses_ptr): Likewise.
5378 (gimple_omp_task_set_clauses): Likewise.
5379 (gimple_omp_task_child_fn): Likewise.
5380 (gimple_omp_task_child_fn_ptr): Likewise.
5381 (gimple_omp_task_set_child_fn): Likewise.
5382 (gimple_omp_task_data_arg): Likewise.
5383 (gimple_omp_task_data_arg_ptr): Likewise.
5384 (gimple_omp_task_set_data_arg): Likewise.
5385 (gimple_omp_taskreg_clauses): Likewise.
5386 (gimple_omp_taskreg_clauses_ptr): Likewise.
5387 (gimple_omp_taskreg_set_clauses): Likewise.
5388 (gimple_omp_taskreg_child_fn): Likewise.
5389 (gimple_omp_taskreg_child_fn_ptr): Likewise.
5390 (gimple_omp_taskreg_set_child_fn): Likewise.
5391 (gimple_omp_taskreg_data_arg): Likewise.
5392 (gimple_omp_taskreg_data_arg_ptr): Likewise.
5393 (gimple_omp_taskreg_set_data_arg): Likewise.
5394 (gimple_omp_task_copy_fn): Likewise.
5395 (gimple_omp_task_copy_fn_ptr): Likewise.
5396 (gimple_omp_task_set_copy_fn): Likewise.
5397 (gimple_omp_task_arg_size): Likewise.
5398 (gimple_omp_task_arg_size_ptr): Likewise.
5399 (gimple_omp_task_set_arg_size): Likewise.
5400 (gimple_omp_task_arg_align): Likewise.
5401 (gimple_omp_task_arg_align_ptr): Likewise.
5402 (gimple_omp_task_set_arg_align): Likewise.
5403 (gimple_omp_single_clauses): Likewise.
5404 (gimple_omp_single_clauses_ptr): Likewise.
5405 (gimple_omp_single_set_clauses): Likewise.
5406 (gimple_omp_target_clauses): Likewise.
5407 (gimple_omp_target_clauses_ptr): Likewise.
5408 (gimple_omp_target_set_clauses): Likewise.
5409 (gimple_omp_target_child_fn): Likewise.
5410 (gimple_omp_target_child_fn_ptr): Likewise.
5411 (gimple_omp_target_set_child_fn): Likewise.
5412 (gimple_omp_target_data_arg): Likewise.
5413 (gimple_omp_target_data_arg_ptr): Likewise.
5414 (gimple_omp_target_set_data_arg): Likewise.
5415 (gimple_omp_teams_clauses): Likewise.
5416 (gimple_omp_teams_clauses_ptr): Likewise.
5417 (gimple_omp_teams_set_clauses): Likewise.
5418 (gimple_omp_sections_clauses): Likewise.
5419 (gimple_omp_sections_clauses_ptr): Likewise.
5420 (gimple_omp_sections_set_clauses): Likewise.
5421 (gimple_omp_sections_control): Likewise.
5422 (gimple_omp_sections_control_ptr): Likewise.
5423 (gimple_omp_sections_set_control): Likewise.
5424 (gimple_omp_for_set_cond): Likewise.
5425 (gimple_omp_for_cond): Likewise.
5426 (gimple_omp_atomic_store_set_val): Likewise.
5427 (gimple_omp_atomic_store_val): Likewise.
5428 (gimple_omp_atomic_store_val_ptr): Likewise.
5429 (gimple_omp_atomic_load_set_lhs): Likewise.
5430 (gimple_omp_atomic_load_lhs): Likewise.
5431 (gimple_omp_atomic_load_lhs_ptr): Likewise.
5432 (gimple_omp_atomic_load_set_rhs): Likewise.
5433 (gimple_omp_atomic_load_rhs): Likewise.
5434 (gimple_omp_atomic_load_rhs_ptr): Likewise.
5435 (gimple_omp_continue_control_def): Likewise.
5436 (gimple_omp_continue_control_def_ptr): Likewise.
5437 (gimple_omp_continue_set_control_def): Likewise.
5438 (gimple_omp_continue_control_use): Likewise.
5439 (gimple_omp_continue_control_use_ptr): Likewise.
5440 (gimple_omp_continue_set_control_use): Likewise.
5441 (gimple_transaction_body_ptr): Likewise.
5442 (gimple_transaction_label): Likewise.
5443 (gimple_transaction_label_ptr): Likewise.
5444 (gimple_transaction_set_body): Likewise.
5445 (gimple_transaction_set_label): Likewise.
5447 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
5448 * ipa-inline-analysis.c (inline_write_summary): Likewise.
5449 * ipa-ref.c (ipa_record_reference): Likewise.
5450 * ipa-reference.c (analyze_function): Likewise.
5451 (ipa_reference_write_optimization_summary): Likewise.
5452 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
5453 (address_taken_from_non_vtable_p): Likewise.
5454 (comdat_can_be_unshared_p_1): Likewise.
5455 * lto-cgraph.c (lto_output_ref): Likewise.
5456 (add_references): Likewise.
5457 (compute_ltrans_boundary): Likewise.
5458 (output_symtab): Likewise.
5459 (input_ref): Likewise.
5460 (input_cgraph_1): Likewise.
5461 (output_cgraph_opt_summary): Likewise.
5462 * lto-streamer-out.c (lto_output): Likewise.
5463 (output_symbol_p): Likewise.
5464 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
5465 (lsei_start_function_in_partition): Likewise.
5466 (lsei_next_variable_in_partition): Likewise.
5467 (lsei_start_variable_in_partition): Likewise.
5468 * symtab.c (insert_to_assembler_name_hash): Likewise.
5469 (unlink_from_assembler_name_hash): Likewise.
5470 (symtab_unregister_node): Likewise.
5471 (symtab_remove_node): Likewise.
5472 (dump_symtab_node): Likewise.
5473 (verify_symtab_base): Likewise.
5474 (verify_symtab_node): Likewise.
5475 (symtab_make_decl_local): Likewise.
5476 (symtab_alias_ultimate_target): Likewise.
5477 (symtab_resolve_alias): Likewise.
5478 (symtab_get_symbol_partitioning_class): Likewise.
5479 * tree-phinodes.c (allocate_phi_node): Likewise.
5480 (reserve_phi_args_for_new_edge): Likewise.
5481 (remove_phi_args): Likewise.
5482 * varpool.c (varpool_node_for_asm): Likewise.
5483 (varpool_remove_unreferenced_decls): Likewise.
5485 2014-04-23 Jeff Law <law@redhat.com>
5487 PR tree-optimization/60902
5488 * tree-ssa-threadedge.c
5489 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
5490 invalidate outputs from statements that do not produce useful
5491 outputs for threading.
5493 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5495 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
5496 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
5497 machine descriptions for Stack Smashing Protector.
5499 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
5501 * aarch64.md (<optab>_rol<mode>3): New pattern.
5502 (<optab>_rolsi3_uxtw): Likewise.
5503 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
5505 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
5507 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
5508 (arm_cortex_a12_tune): Likewise.
5510 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5512 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
5514 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5516 * config/arm/arm.md (arm_rev16si2): New pattern.
5517 (arm_rev16si2_alt): Likewise.
5518 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
5520 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5522 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
5523 (rev16<mode>2_alt): Likewise.
5524 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
5525 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
5526 (aarch_rev16_shleft_mask_imm_p): Likewise.
5527 (aarch_rev16_p_1): Likewise.
5528 (aarch_rev16_p): Likewise.
5529 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5530 (aarch_rev16_shright_mask_imm_p): Likewise.
5531 (aarch_rev16_shleft_mask_imm_p): Likewise.
5533 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5535 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5536 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5538 (cortex_a53_extra_costs): Likewise.
5539 (cortex_a57_extra_costs): Likewise.
5540 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5541 (cortexa7_extra_costs): Likewise.
5542 (cortexa8_extra_costs): Likewise.
5543 (cortexa12_extra_costs): Likewise.
5544 (cortexa15_extra_costs): Likewise.
5545 (v7m_extra_costs): Likewise.
5546 (arm_new_rtx_costs): Handle BSWAP.
5548 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5550 * config/arm/arm.c (cortexa8_extra_costs): New table.
5551 (arm_cortex_a8_tune): New tuning struct.
5552 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5554 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5556 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5558 2014-04-23 Richard Biener <rguenther@suse.de>
5560 * Makefile.in (OBJS): Remove loop-unswitch.o.
5561 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5562 * passes.def (pass_rtl_unswitch): Likewise.
5563 * loop-init.c (gate_rtl_unswitch): Likewise.
5564 (rtl_unswitch): Likewise.
5565 (pass_data_rtl_unswitch): Likewise.
5566 (pass_rtl_unswitch): Likewise.
5567 (make_pass_rtl_unswitch): Likewise.
5568 * rtl.h (reversed_condition): Likewise.
5569 (compare_and_jump_seq): Likewise.
5570 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5572 * loop-unroll.c (compare_and_jump_seq): Likewise.
5574 2014-04-23 Richard Biener <rguenther@suse.de>
5576 PR tree-optimization/60903
5577 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5578 commented code block.
5579 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5580 loop flags to newly created BBs and edges.
5582 2014-04-23 Nick Clifton <nickc@redhat.com>
5584 * config/msp430/msp430.c (msp430_handle_option): Move function
5586 (msp430_option_override): Simplify mcu and mcpu option handling.
5587 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
5588 support for -mhwmult command line option.
5589 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
5590 -mhwmult command line option.
5591 (msp430_hwmult_enabled): Delete.
5592 (msp43o_output_labelref): Add support for -mhwmult command line option.
5593 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
5594 (umulsidi3): Likewise.
5595 * config/msp430/msp430.opt (mmcu): Add Report attribute.
5596 (mcpu, mlarge, msmall): Likewise.
5597 (mhwmult): New option.
5598 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
5600 (msp430_is_f5_mcu): Remove prototype.
5601 (msp430_use_f5_series_hwmult): Add prototype.
5602 * config/msp430/msp430-opts.h: New file.
5603 * common/config/msp430: New directory.
5604 * common/config/msp430/msp430-common.c: New file.
5605 * config.gcc (msp430): Remove target_has_targetm_common.
5606 * doc/invoke.texi: Document -mhwmult command line option.
5608 2014-04-23 Nick Clifton <nickc@redhat.com>
5610 * config/i386/cygwin.h (ENDFILE_SPEC): Include
5611 default-manifest.o if it can be found in the search path.
5612 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
5614 2014-04-23 Terry Guo <terry.guo@arm.com>
5616 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
5618 2014-04-23 Richard Biener <rguenther@suse.de>
5621 * tree-inline.c (declare_return_variable): Use mark_addressable.
5623 2014-04-23 Richard Biener <rguenther@suse.de>
5626 * loop-init.c (loop_optimizer_init): Make sure to apply
5627 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
5629 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5632 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
5634 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
5635 if flag_sanitize_undefined_trap_on_error.
5636 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
5637 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
5638 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
5639 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
5640 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
5641 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
5642 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
5643 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
5644 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
5645 * ubsan.c (ubsan_instrument_unreachable): Return
5646 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
5647 (ubsan_expand_null_ifn): Emit __builtin_trap ()
5648 if flag_sanitize_undefined_trap_on_error and
5649 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
5650 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
5651 instrument_bool_enum_load): Emit __builtin_trap () if
5652 flag_sanitize_undefined_trap_on_error and
5653 __builtin_handle_*_abort () if !flag_sanitize_recover.
5654 * doc/invoke.texi (-fsanitize-recover,
5655 -fsanitize-undefined-trap-on-error): Document.
5657 2014-04-22 Christian Bruel <christian.bruel@st.com>
5659 * config/sh/sh.md (mov<mode>): Replace movQIHI.
5660 Force immediates to SImode.
5662 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
5664 * config/nios2/nios2.md (UNSPEC_ROUND): New.
5666 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
5667 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
5668 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
5669 (nios2_fpu_insn): Add entry for round.
5670 (N2FPU_NO_ERRNO_P): Define.
5671 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
5673 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
5675 2014-04-22 Richard Henderson <rth@redhat.com>
5677 * config/aarch64/aarch64 (addti3, subti3): New expanders.
5678 (add<GPI>3_compare0): Remove leading * from name.
5679 (add<GPI>3_carryin): Likewise.
5680 (sub<GPI>3_compare0): Likewise.
5681 (sub<GPI>3_carryin): Likewise.
5682 (<su_optab>mulditi3): New expander.
5683 (multi3): New expander.
5684 (madd<GPI>): Remove leading * from name.
5686 2014-04-22 Martin Jambor <mjambor@suse.cz>
5688 * cgraphclones.c (cgraph_function_versioning): Copy
5689 ipa_transforms_to_apply instead of asserting it is empty.
5691 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
5694 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
5695 on count_exp to get mode.
5697 2014-04-22 Andrew Pinski <apinski@cavium.com>
5699 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5700 Handle TLS for ILP32.
5701 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
5702 (tlsie_small_<mode>): this and handle PTR.
5703 (tlsie_small_sidi): New pattern.
5704 (tlsle_small): Change to an expand to handle ILP32.
5705 (tlsle_small_<mode>): New pattern.
5706 (tlsdesc_small): Rename to ...
5707 (tlsdesc_small_<mode>): this and handle PTR.
5709 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5711 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
5713 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5715 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5716 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
5717 (aarch64_types_signed_poly_qualifiers): Likewise.
5718 (aarch64_types_unsigned_signed_qualifiers): Likewise.
5719 (aarch64_types_poly_signed_qualifiers): Likewise.
5720 (TYPES_REINTERP_SS): Type macro added.
5721 (TYPES_REINTERP_SU): Likewise.
5722 (TYPES_REINTERP_SP): Likewise.
5723 (TYPES_REINTERP_US): Likewise.
5724 (TYPES_REINTERP_PS): Likewise.
5725 (aarch64_fold_builtin): New expression folding added.
5726 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
5727 Declarations removed.
5728 (REINTERP_SS): Declarations added.
5729 (REINTERP_US): Likewise.
5730 (REINTERP_PS): Likewise.
5731 (REINTERP_SU): Likewise.
5732 (REINTERP_SP): Likewise.
5733 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
5734 (vreinterpretq_p8_f64): Likewise.
5735 (vreinterpret_p16_f64): Likewise.
5736 (vreinterpretq_p16_f64): Likewise.
5737 (vreinterpret_f32_f64): Likewise.
5738 (vreinterpretq_f32_f64): Likewise.
5739 (vreinterpret_f64_f32): Likewise.
5740 (vreinterpret_f64_p8): Likewise.
5741 (vreinterpret_f64_p16): Likewise.
5742 (vreinterpret_f64_s8): Likewise.
5743 (vreinterpret_f64_s16): Likewise.
5744 (vreinterpret_f64_s32): Likewise.
5745 (vreinterpret_f64_s64): Likewise.
5746 (vreinterpret_f64_u8): Likewise.
5747 (vreinterpret_f64_u16): Likewise.
5748 (vreinterpret_f64_u32): Likewise.
5749 (vreinterpret_f64_u64): Likewise.
5750 (vreinterpretq_f64_f32): Likewise.
5751 (vreinterpretq_f64_p8): Likewise.
5752 (vreinterpretq_f64_p16): Likewise.
5753 (vreinterpretq_f64_s8): Likewise.
5754 (vreinterpretq_f64_s16): Likewise.
5755 (vreinterpretq_f64_s32): Likewise.
5756 (vreinterpretq_f64_s64): Likewise.
5757 (vreinterpretq_f64_u8): Likewise.
5758 (vreinterpretq_f64_u16): Likewise.
5759 (vreinterpretq_f64_u32): Likewise.
5760 (vreinterpretq_f64_u64): Likewise.
5761 (vreinterpret_s64_f64): Likewise.
5762 (vreinterpretq_s64_f64): Likewise.
5763 (vreinterpret_u64_f64): Likewise.
5764 (vreinterpretq_u64_f64): Likewise.
5765 (vreinterpret_s8_f64): Likewise.
5766 (vreinterpretq_s8_f64): Likewise.
5767 (vreinterpret_s16_f64): Likewise.
5768 (vreinterpretq_s16_f64): Likewise.
5769 (vreinterpret_s32_f64): Likewise.
5770 (vreinterpretq_s32_f64): Likewise.
5771 (vreinterpret_u8_f64): Likewise.
5772 (vreinterpretq_u8_f64): Likewise.
5773 (vreinterpret_u16_f64): Likewise.
5774 (vreinterpretq_u16_f64): Likewise.
5775 (vreinterpret_u32_f64): Likewise.
5776 (vreinterpretq_u32_f64): Likewise.
5778 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5780 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5781 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
5782 (vreinterpret_p8_s8): Likewise.
5783 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
5784 (vreinterpret_p8_s16): Likewise.
5785 (vreinterpret_p8_s32): Likewise.
5786 (vreinterpret_p8_s64): Likewise.
5787 (vreinterpret_p8_f32): Likewise.
5788 (vreinterpret_p8_u8): Likewise.
5789 (vreinterpret_p8_u16): Likewise.
5790 (vreinterpret_p8_u32): Likewise.
5791 (vreinterpret_p8_u64): Likewise.
5792 (vreinterpret_p8_p16): Likewise.
5793 (vreinterpretq_p8_s8): Likewise.
5794 (vreinterpretq_p8_s16): Likewise.
5795 (vreinterpretq_p8_s32): Likewise.
5796 (vreinterpretq_p8_s64): Likewise.
5797 (vreinterpretq_p8_f32): Likewise.
5798 (vreinterpretq_p8_u8): Likewise.
5799 (vreinterpretq_p8_u16): Likewise.
5800 (vreinterpretq_p8_u32): Likewise.
5801 (vreinterpretq_p8_u64): Likewise.
5802 (vreinterpretq_p8_p16): Likewise.
5803 (vreinterpret_p16_s8): Likewise.
5804 (vreinterpret_p16_s16): Likewise.
5805 (vreinterpret_p16_s32): Likewise.
5806 (vreinterpret_p16_s64): Likewise.
5807 (vreinterpret_p16_f32): Likewise.
5808 (vreinterpret_p16_u8): Likewise.
5809 (vreinterpret_p16_u16): Likewise.
5810 (vreinterpret_p16_u32): Likewise.
5811 (vreinterpret_p16_u64): Likewise.
5812 (vreinterpret_p16_p8): Likewise.
5813 (vreinterpretq_p16_s8): Likewise.
5814 (vreinterpretq_p16_s16): Likewise.
5815 (vreinterpretq_p16_s32): Likewise.
5816 (vreinterpretq_p16_s64): Likewise.
5817 (vreinterpretq_p16_f32): Likewise.
5818 (vreinterpretq_p16_u8): Likewise.
5819 (vreinterpretq_p16_u16): Likewise.
5820 (vreinterpretq_p16_u32): Likewise.
5821 (vreinterpretq_p16_u64): Likewise.
5822 (vreinterpretq_p16_p8): Likewise.
5823 (vreinterpret_f32_s8): Likewise.
5824 (vreinterpret_f32_s16): Likewise.
5825 (vreinterpret_f32_s32): Likewise.
5826 (vreinterpret_f32_s64): Likewise.
5827 (vreinterpret_f32_u8): Likewise.
5828 (vreinterpret_f32_u16): Likewise.
5829 (vreinterpret_f32_u32): Likewise.
5830 (vreinterpret_f32_u64): Likewise.
5831 (vreinterpret_f32_p8): Likewise.
5832 (vreinterpret_f32_p16): Likewise.
5833 (vreinterpretq_f32_s8): Likewise.
5834 (vreinterpretq_f32_s16): Likewise.
5835 (vreinterpretq_f32_s32): Likewise.
5836 (vreinterpretq_f32_s64): Likewise.
5837 (vreinterpretq_f32_u8): Likewise.
5838 (vreinterpretq_f32_u16): Likewise.
5839 (vreinterpretq_f32_u32): Likewise.
5840 (vreinterpretq_f32_u64): Likewise.
5841 (vreinterpretq_f32_p8): Likewise.
5842 (vreinterpretq_f32_p16): Likewise.
5843 (vreinterpret_s64_s8): Likewise.
5844 (vreinterpret_s64_s16): Likewise.
5845 (vreinterpret_s64_s32): Likewise.
5846 (vreinterpret_s64_f32): Likewise.
5847 (vreinterpret_s64_u8): Likewise.
5848 (vreinterpret_s64_u16): Likewise.
5849 (vreinterpret_s64_u32): Likewise.
5850 (vreinterpret_s64_u64): Likewise.
5851 (vreinterpret_s64_p8): Likewise.
5852 (vreinterpret_s64_p16): Likewise.
5853 (vreinterpretq_s64_s8): Likewise.
5854 (vreinterpretq_s64_s16): Likewise.
5855 (vreinterpretq_s64_s32): Likewise.
5856 (vreinterpretq_s64_f32): Likewise.
5857 (vreinterpretq_s64_u8): Likewise.
5858 (vreinterpretq_s64_u16): Likewise.
5859 (vreinterpretq_s64_u32): Likewise.
5860 (vreinterpretq_s64_u64): Likewise.
5861 (vreinterpretq_s64_p8): Likewise.
5862 (vreinterpretq_s64_p16): Likewise.
5863 (vreinterpret_u64_s8): Likewise.
5864 (vreinterpret_u64_s16): Likewise.
5865 (vreinterpret_u64_s32): Likewise.
5866 (vreinterpret_u64_s64): Likewise.
5867 (vreinterpret_u64_f32): Likewise.
5868 (vreinterpret_u64_u8): Likewise.
5869 (vreinterpret_u64_u16): Likewise.
5870 (vreinterpret_u64_u32): Likewise.
5871 (vreinterpret_u64_p8): Likewise.
5872 (vreinterpret_u64_p16): Likewise.
5873 (vreinterpretq_u64_s8): Likewise.
5874 (vreinterpretq_u64_s16): Likewise.
5875 (vreinterpretq_u64_s32): Likewise.
5876 (vreinterpretq_u64_s64): Likewise.
5877 (vreinterpretq_u64_f32): Likewise.
5878 (vreinterpretq_u64_u8): Likewise.
5879 (vreinterpretq_u64_u16): Likewise.
5880 (vreinterpretq_u64_u32): Likewise.
5881 (vreinterpretq_u64_p8): Likewise.
5882 (vreinterpretq_u64_p16): Likewise.
5883 (vreinterpret_s8_s16): Likewise.
5884 (vreinterpret_s8_s32): Likewise.
5885 (vreinterpret_s8_s64): Likewise.
5886 (vreinterpret_s8_f32): Likewise.
5887 (vreinterpret_s8_u8): Likewise.
5888 (vreinterpret_s8_u16): Likewise.
5889 (vreinterpret_s8_u32): Likewise.
5890 (vreinterpret_s8_u64): Likewise.
5891 (vreinterpret_s8_p8): Likewise.
5892 (vreinterpret_s8_p16): Likewise.
5893 (vreinterpretq_s8_s16): Likewise.
5894 (vreinterpretq_s8_s32): Likewise.
5895 (vreinterpretq_s8_s64): Likewise.
5896 (vreinterpretq_s8_f32): Likewise.
5897 (vreinterpretq_s8_u8): Likewise.
5898 (vreinterpretq_s8_u16): Likewise.
5899 (vreinterpretq_s8_u32): Likewise.
5900 (vreinterpretq_s8_u64): Likewise.
5901 (vreinterpretq_s8_p8): Likewise.
5902 (vreinterpretq_s8_p16): Likewise.
5903 (vreinterpret_s16_s8): Likewise.
5904 (vreinterpret_s16_s32): Likewise.
5905 (vreinterpret_s16_s64): Likewise.
5906 (vreinterpret_s16_f32): Likewise.
5907 (vreinterpret_s16_u8): Likewise.
5908 (vreinterpret_s16_u16): Likewise.
5909 (vreinterpret_s16_u32): Likewise.
5910 (vreinterpret_s16_u64): Likewise.
5911 (vreinterpret_s16_p8): Likewise.
5912 (vreinterpret_s16_p16): Likewise.
5913 (vreinterpretq_s16_s8): Likewise.
5914 (vreinterpretq_s16_s32): Likewise.
5915 (vreinterpretq_s16_s64): Likewise.
5916 (vreinterpretq_s16_f32): Likewise.
5917 (vreinterpretq_s16_u8): Likewise.
5918 (vreinterpretq_s16_u16): Likewise.
5919 (vreinterpretq_s16_u32): Likewise.
5920 (vreinterpretq_s16_u64): Likewise.
5921 (vreinterpretq_s16_p8): Likewise.
5922 (vreinterpretq_s16_p16): Likewise.
5923 (vreinterpret_s32_s8): Likewise.
5924 (vreinterpret_s32_s16): Likewise.
5925 (vreinterpret_s32_s64): Likewise.
5926 (vreinterpret_s32_f32): Likewise.
5927 (vreinterpret_s32_u8): Likewise.
5928 (vreinterpret_s32_u16): Likewise.
5929 (vreinterpret_s32_u32): Likewise.
5930 (vreinterpret_s32_u64): Likewise.
5931 (vreinterpret_s32_p8): Likewise.
5932 (vreinterpret_s32_p16): Likewise.
5933 (vreinterpretq_s32_s8): Likewise.
5934 (vreinterpretq_s32_s16): Likewise.
5935 (vreinterpretq_s32_s64): Likewise.
5936 (vreinterpretq_s32_f32): Likewise.
5937 (vreinterpretq_s32_u8): Likewise.
5938 (vreinterpretq_s32_u16): Likewise.
5939 (vreinterpretq_s32_u32): Likewise.
5940 (vreinterpretq_s32_u64): Likewise.
5941 (vreinterpretq_s32_p8): Likewise.
5942 (vreinterpretq_s32_p16): Likewise.
5943 (vreinterpret_u8_s8): Likewise.
5944 (vreinterpret_u8_s16): Likewise.
5945 (vreinterpret_u8_s32): Likewise.
5946 (vreinterpret_u8_s64): Likewise.
5947 (vreinterpret_u8_f32): Likewise.
5948 (vreinterpret_u8_u16): Likewise.
5949 (vreinterpret_u8_u32): Likewise.
5950 (vreinterpret_u8_u64): Likewise.
5951 (vreinterpret_u8_p8): Likewise.
5952 (vreinterpret_u8_p16): Likewise.
5953 (vreinterpretq_u8_s8): Likewise.
5954 (vreinterpretq_u8_s16): Likewise.
5955 (vreinterpretq_u8_s32): Likewise.
5956 (vreinterpretq_u8_s64): Likewise.
5957 (vreinterpretq_u8_f32): Likewise.
5958 (vreinterpretq_u8_u16): Likewise.
5959 (vreinterpretq_u8_u32): Likewise.
5960 (vreinterpretq_u8_u64): Likewise.
5961 (vreinterpretq_u8_p8): Likewise.
5962 (vreinterpretq_u8_p16): Likewise.
5963 (vreinterpret_u16_s8): Likewise.
5964 (vreinterpret_u16_s16): Likewise.
5965 (vreinterpret_u16_s32): Likewise.
5966 (vreinterpret_u16_s64): Likewise.
5967 (vreinterpret_u16_f32): Likewise.
5968 (vreinterpret_u16_u8): Likewise.
5969 (vreinterpret_u16_u32): Likewise.
5970 (vreinterpret_u16_u64): Likewise.
5971 (vreinterpret_u16_p8): Likewise.
5972 (vreinterpret_u16_p16): Likewise.
5973 (vreinterpretq_u16_s8): Likewise.
5974 (vreinterpretq_u16_s16): Likewise.
5975 (vreinterpretq_u16_s32): Likewise.
5976 (vreinterpretq_u16_s64): Likewise.
5977 (vreinterpretq_u16_f32): Likewise.
5978 (vreinterpretq_u16_u8): Likewise.
5979 (vreinterpretq_u16_u32): Likewise.
5980 (vreinterpretq_u16_u64): Likewise.
5981 (vreinterpretq_u16_p8): Likewise.
5982 (vreinterpretq_u16_p16): Likewise.
5983 (vreinterpret_u32_s8): Likewise.
5984 (vreinterpret_u32_s16): Likewise.
5985 (vreinterpret_u32_s32): Likewise.
5986 (vreinterpret_u32_s64): Likewise.
5987 (vreinterpret_u32_f32): Likewise.
5988 (vreinterpret_u32_u8): Likewise.
5989 (vreinterpret_u32_u16): Likewise.
5990 (vreinterpret_u32_u64): Likewise.
5991 (vreinterpret_u32_p8): Likewise.
5992 (vreinterpret_u32_p16): Likewise.
5993 (vreinterpretq_u32_s8): Likewise.
5994 (vreinterpretq_u32_s16): Likewise.
5995 (vreinterpretq_u32_s32): Likewise.
5996 (vreinterpretq_u32_s64): Likewise.
5997 (vreinterpretq_u32_f32): Likewise.
5998 (vreinterpretq_u32_u8): Likewise.
5999 (vreinterpretq_u32_u16): Likewise.
6000 (vreinterpretq_u32_u64): Likewise.
6001 (vreinterpretq_u32_p8): Likewise.
6002 (vreinterpretq_u32_p16): Likewise.
6004 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6006 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
6008 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
6010 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
6011 (vqnegd_s64): Likewise.
6012 (vqabs_s64): Likewise.
6013 (vqabsd_s64): Likewise.
6015 2014-04-22 Richard Henderson <rth@redhat.com>
6017 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
6018 computation to the top of the loop.
6020 2014-04-22 Renlin <renlin.li@arm.com>
6021 Jiong Wang <jiong.wang@arm.com>
6023 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
6024 * config/aarch64/aarch64.c (aarch64_layout_frame)
6025 (aarch64_initial_elimination_offset): Likewise.
6027 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
6029 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
6032 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
6034 * machmode.h (bitwise_mode_for_mode): Declare.
6035 * stor-layout.h (bitwise_type_for_mode): Likewise.
6036 * stor-layout.c (bitwise_mode_for_mode): New function.
6037 (bitwise_type_for_mode): Likewise.
6038 * builtins.c (fold_builtin_memory_op): Use it instead of
6039 int_mode_for_mode and build_nonstandard_integer_type.
6041 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6043 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
6044 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
6045 (*-*-solaris2*): Simplify.
6046 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
6047 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
6048 *-*-solaris2.9* handling.
6050 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
6052 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
6053 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
6055 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
6056 * configure: Regenerate.
6058 * config/i386/sol2-9.h: Remove.
6060 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
6061 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
6062 Remove Solaris 9 references.
6064 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
6066 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
6067 (floatuns<GPI:mode><GPF:mode>2): Remove.
6068 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
6069 and floatuns conversions.
6070 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
6071 and floatuns conversions.
6072 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
6073 (w1,w2): New mode attributes for inequal width conversions.
6075 2014-04-22 Renlin Li <Renlin.Li@arm.com>
6077 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
6078 the output asm format.
6080 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
6082 * config/aarch64/aarch64-simd.md
6083 (aarch64_cm<optab>di): Always split.
6084 (*aarch64_cm<optab>di): New.
6085 (aarch64_cmtstdi): Always split.
6086 (*aarch64_cmtstdi): New.
6088 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6090 PR tree-optimization/60823
6091 * omp-low.c (ipa_simd_modify_function_body): Go through
6092 all SSA_NAMEs and for those refering to vector arguments
6093 which are going to be replaced adjust SSA_NAME_VAR and,
6094 if it is a default definition, change it into a non-default
6095 definition assigned at the beginning of function from new_decl.
6096 (ipa_simd_modify_stmt_ops): Rewritten.
6097 * tree-dfa.c (set_ssa_default_def): When removing default def,
6098 check for NULL loc instead of NULL *loc.
6100 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6102 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
6103 restrictions on core registers for DImode values in Thumb2.
6105 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6107 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
6108 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
6110 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6112 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
6113 (*iordi_notzesidi_di): Likewise.
6114 (*iordi_notsesidi_di): Likewise.
6116 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6118 * config/arm/arm-protos.h (tune_params): New struct members.
6119 * config/arm/arm.c: Initialise tune_params per processor.
6120 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
6121 for speed, based on new tune_params.
6123 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6125 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
6126 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
6127 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
6128 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
6129 * config/aarch64/arm_neon.h (vrnd_f64): Added.
6130 (vrnda_f64): Likewise.
6131 (vrndi_f64): Likewise.
6132 (vrndm_f64): Likewise.
6133 (vrndn_f64): Likewise.
6134 (vrndp_f64): Likewise.
6135 (vrndx_f64): Likewise.
6137 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6139 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
6140 GET_MODE_SIZE argument is enum machine_mode.
6142 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6145 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
6146 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
6148 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
6151 * asan.c (asan_emit_stack_protection): Force the base to align to
6152 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
6153 appropriate bits if STRICT_ALIGNMENT.
6154 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
6156 (expand_used_vars): Leave a space in the stack frame for alignment
6157 if STRICT_ALIGNMENT.
6159 2014-04-21 David Malcolm <dmalcolm@redhat.com>
6161 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
6163 (gimple_store_p): Likewise.
6164 (gimple_assign_load_p): Likewise.
6165 (gimple_assign_cast_p): Likewise.
6166 (gimple_clobber_p): Likewise.
6168 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
6169 rather than a gimple.
6170 (gimple_assign_cast_p): Likewise.
6172 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
6175 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
6176 If mode is DDmode and TARGET_E500_DOUBLE allow move.
6178 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
6179 more debug information for E500 if -mdebug=reg.
6181 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
6184 * config/i386/i386.c (ix86_expand_builtin)
6185 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
6186 register for target RTX.
6187 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
6189 2014-04-18 Cong Hou <congh@google.com>
6191 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
6192 the widen-mult pattern by handling two operands with different sizes,
6193 and operands whose size is smaller than half of the result type.
6195 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6197 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
6198 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
6199 (do_estimate_edge_time): Compute it.
6200 * ipa-inline.c (want_inline_small_function_p): Bypass
6201 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
6203 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6205 * ipa-inline.c (spec_rem): New static variable.
6206 (dump_overall_stats): New function.
6207 (dump_inline_stats): New function.
6209 2014-04-18 Richard Henderson <rth@redhat.com>
6211 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
6212 to GET_MODE_SIZE, not a reg_class_t.
6214 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6216 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
6217 (vsx_xxmrglw_<mode>): Likewise.
6219 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
6222 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
6223 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
6224 (rs6000_init_hard_regno_mode_ok): Likewise.
6226 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
6228 * ipa-inline.c (inline_small_functions): Account only non-cold
6230 * doc/invoke.texi (inline-unit-growth): Update documentation.
6232 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
6234 * config/rs6000/rs6000.md (addti3, subti3): New.
6236 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
6239 * config/i386/i386.c (ix86_expand_clear): Remove outdated
6240 comment. Check optimize_insn_for_size_p instead of
6241 optimize_insn_for_speed_p.
6243 2014-04-17 Martin Jambor <mjambor@suse.cz>
6245 * gimple-iterator.c (gsi_start_edge): New function.
6246 * gimple-iterator.h (gsi_start_edge): Declare.
6247 * tree-sra.c (single_non_eh_succ): New function.
6248 (disqualify_ops_if_throwing_stmt): Renamed to
6249 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
6250 having one non-EH successor BB.
6251 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
6252 generate loads into replacements.
6253 (sra_modify_assign): Likewise and and also use the simple path for
6255 (sra_modify_function_body): Commit statements on edges.
6257 2014-04-17 Richard Biener <rguenther@suse.de>
6260 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
6261 comparison results and add clarifying comment.
6263 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6265 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
6266 (blank_mode): Initialize it.
6267 (emit_mode_size_inline, emit_mode_nunits_inline,
6268 emit_mode_inner_inline): New functions.
6269 (emit_insn_modes_h): Call them and surround their output with
6270 #if GCC_VERSION >= 4001 ... #endif.
6271 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
6272 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
6273 mode_* arrays if the argument is __builtin_constant_p.
6274 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
6275 is enum machine_mode.
6277 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6279 * passes.c (opt_pass::execute): Adjust.
6280 (pass_manager::execute_pass_mode_switching): Likewise.
6281 (early_local_passes::execute): Likewise.
6282 (execute_one_pass): Pass cfun to the pass's execute method.
6283 * tree-pass.h (opt_pass::execute): Add function * argument.
6284 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6285 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6286 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6287 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6288 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6289 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
6290 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
6291 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6292 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6293 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
6294 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
6295 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
6296 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
6297 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
6298 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6299 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6300 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6301 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
6302 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
6303 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6304 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6305 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6306 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6307 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6308 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6309 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6310 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6311 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6312 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6315 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6317 * passes.c (opt_pass::gate): Take function * argument.
6318 (gate_all_early_local_passes): Merge into
6319 (early_local_passes::gate): this.
6320 (gate_all_early_optimizations): Merge into
6321 (all_early_optimizations::gate): this.
6322 (gate_all_optimizations): Mege into
6323 (all_optimizations::gate): this.
6324 (gate_all_optimizations_g): Merge into
6325 (all_optimizations_g::gate): this.
6326 (gate_rest_of_compilation): Mege into
6327 (rest_of_compilation::gate): this.
6328 (gate_postreload): Merge into
6329 (postreload::gate): this.
6330 (dump_one_pass): Pass cfun to the pass's gate method.
6331 (execute_ipa_summary_passes): Likewise.
6332 (execute_one_pass): Likewise.
6333 (ipa_write_summaries_2): Likewise.
6334 (ipa_write_optimization_summaries_1): Likewise.
6335 (ipa_read_summaries_1): Likewise.
6336 (ipa_read_optimization_summaries_1): Likewise.
6337 (execute_ipa_stmt_fixups): Likewise.
6338 * tree-pass.h (opt_pass::gate): Add function * argument.
6339 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
6340 combine-stack-adj.c, combine.c, compare-elim.c,
6341 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6342 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
6343 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
6344 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
6345 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6346 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6347 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6348 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
6349 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
6350 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
6351 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6352 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6353 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
6354 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6355 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6356 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6357 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6358 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6359 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6360 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6361 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6362 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6363 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6364 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
6365 var-tracking.c, vtable-verify.c, web.c: Adjust.
6367 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6369 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
6370 * configure: Regenerate.
6372 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6374 * passes.c (dump_one_pass): don't check pass->has_gate.
6375 (execute_ipa_summary_passes): Likewise.
6376 (execute_one_pass): Likewise.
6377 (ipa_write_summaries_2): Likewise.
6378 (ipa_write_optimization_summaries_1): Likewise.
6379 (ipa_read_optimization_summaries_1): Likewise.
6380 (execute_ipa_stmt_fixups): Likewise.
6381 * tree-pass.h (pass_data::has_gate): Remove.
6382 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6383 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6384 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6385 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6386 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6387 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
6388 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
6389 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
6390 gimple-low.c, gimple-ssa-isolate-paths.c,
6391 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
6392 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6393 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6394 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
6395 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
6396 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
6397 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
6398 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6399 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
6400 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6401 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6402 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6403 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6404 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6405 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6406 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6407 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6408 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6409 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6410 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6411 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6412 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6415 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6417 * pass_manager.h (pass_manager::register_dump_files_1): Remove
6419 * passes.c (pass_manager::register_dump_files_1): Merge into
6420 (pass_manager::register_dump_files): this, and remove its handling of
6421 properties since the pass always has the properties anyway.
6422 (pass_manager::pass_manager): Adjust.
6424 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6426 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
6427 * passes.c (pass_manager::register_dump_files_1): Remove dead code
6428 dealing with properties.
6429 (pass_manager::register_dump_files): Adjust.
6431 2014-03-20 Mark Wielaard <mjw@redhat.com>
6433 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
6434 then represent the bound as normal constant value.
6436 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6439 Forward port from 4.8 branch
6440 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
6442 * config/i386/bmiintrin.h (_blsi_u32): New.
6446 (_blsmsk_u32): Ditto.
6447 (_blsmsk_u64): Ditto.
6448 (_tzcnt_u32): Ditto.
6449 (_tzcnt_u64): Ditto.
6451 2014-04-17 Kito Cheng <kito@0xlab.org>
6453 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
6455 2014-04-17 Richard Biener <rguenther@suse.de>
6458 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
6459 boolean results for comparisons.
6461 2014-04-17 Richard Biener <rguenther@suse.de>
6463 PR tree-optimization/60836
6464 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
6465 initial PHI args to be gimple values.
6467 2014-04-17 Richard Biener <rguenther@suse.de>
6469 PR tree-optimization/60841
6470 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
6471 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
6472 of stmts to SLP build.
6473 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
6474 (vect_analyze_slp): Likewise.
6475 (vect_analyze_slp_instance): Likewise.
6476 (vect_build_slp_tree): Limit overall SLP tree growth.
6477 * tree-vectorizer.h (vect_analyze_data_refs,
6478 vect_analyze_slp): Adjust prototypes.
6480 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6482 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
6485 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6487 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
6488 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
6489 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
6490 for TARGET_SLOW_PSHUFB
6492 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6494 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
6495 * config/i386/i386.c (intel_cost): Ditto.
6497 2014-04-17 Joey Ye <joey.ye@arm.com>
6499 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
6501 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6503 * opts.c (common_handle_option): Disable -fipa-reference coorectly
6506 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6508 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
6509 (type_all_derivations_known_p): New predicate.
6510 (type_all_ctors_visible_p): New predicate.
6511 (type_possibly_instantiated_p): New predicate.
6512 (get_odr_type): Compute all_derivations_known.
6513 (dump_odr_type): Dump the flag.
6514 (maybe_record_type): Cleanup.
6515 (record_target_from_binfo): Add bases_to_consider array;
6516 record bases for types w/o instances and skip CXX destructor.
6517 (possible_polymorphic_call_targets_1): Add bases_to_consider
6518 and consider_construction parameters; check if type may have instance.
6519 (get_polymorphic_call_info): Set maybe_in_construction to true
6520 when we know nothing.
6521 (record_targets_from_bases): Skip CXX destructors; they are
6522 never called for types in construction.
6523 (possible_polymorphic_call_targets): Do not record target when
6524 type may not have instance.
6526 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6529 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6530 external aliases alive, too.
6532 2014-04-16 Andrew Pinski <apinski@cavium.com>
6534 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6537 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6539 * final.c (compute_alignments): Do not apply loop alignment to a block
6540 falling through to the exit.
6542 2014-04-16 Catherine Moore <clm@codesourcery.com>
6544 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6545 Adjust constraints for microMIPS store patterns.
6547 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6549 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6551 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6553 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6554 (append_use): Run at -O0.
6555 (append_vdef): Likewise.
6556 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6557 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6559 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6561 PR tree-optimization/60844
6562 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6563 (propagate_op_to_single_use, remove_visited_stmt_chain,
6564 linearize_expr, repropagate_negates, reassociate_bb): Use it
6565 instead of gsi_remove.
6567 2014-04-16 Martin Jambor <mjambor@suse.cz>
6569 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6570 ipa_transforms_to_apply.
6571 (cgraph_function_versioning): Assert that old_node has empty
6572 ipa_transforms_to_apply.
6573 * trans-mem.c (ipa_tm_create_version): Likewise.
6574 * tree-inline.c (tree_function_versioning): Do not duplicate
6575 ipa_transforms_to_apply.
6577 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6580 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6582 Pass necessary as flags on 64-bit Solaris/x86.
6583 Use lowercase relocs for x86_64-*-*.
6584 * configure: Regenerate.
6586 2014-04-15 Jan Hubicka <jh@suse.cz>
6588 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
6589 (maybe_record_node, likely_target_p): Use it.
6591 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6594 Revert following patch
6596 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6599 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6600 software floating point or no floating point registers, do not
6601 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6602 in GPRs that occurs after we tested for GPRs that would never be
6605 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6606 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6607 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6608 specifically allow DDmode, since that does not use the SPE SIMD
6611 2014-03-21 Mark Wielaard <mjw@redhat.com>
6613 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
6614 as unsigned or int depending on type and value used.
6616 2014-04-15 Richard Biener <rguenther@suse.de>
6618 PR rtl-optimization/56965
6619 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
6620 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
6622 * alias.c (true_dependence_1): Do not call
6623 nonoverlapping_component_refs_p.
6624 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
6625 nonoverlapping_component_refs_p.
6626 (indirect_refs_may_alias_p): Likewise.
6628 2014-04-15 Teresa Johnson <tejohnson@google.com>
6630 * cfg.c (dump_bb_info): Fix flags check.
6631 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
6633 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6635 PR rtl-optimization/60663
6636 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
6639 2014-04-15 Richard Biener <rguenther@suse.de>
6641 * lto-streamer.h (LTO_major_version): Bump to 4.
6643 2014-04-15 Richard Biener <rguenther@suse.de>
6645 * common.opt (lto_partition_model): New enum.
6646 (flto-partition=): Merge separate options with a single with argument,
6647 add -flto-partition=one support.
6648 * flag-types.h (enum lto_partition_model): Declare.
6649 * opts.c (finish_options): Remove duplicate -flto-partition=
6651 * lto-wrapper.c (run_gcc): Adjust.
6653 2014-04-15 Richard Biener <rguenther@suse.de>
6655 * alias.c (ncr_compar): New function.
6656 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
6658 2014-04-15 Richard Biener <rguenther@suse.de>
6660 * alias.c (record_component_aliases): Do not walk BINFOs.
6662 2014-04-15 Richard Biener <rguenther@suse.de>
6664 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6665 Add struct function argument and adjust.
6666 (find_func_aliases_for_call): Likewise.
6667 (find_func_aliases): Likewise.
6668 (find_func_clobbers): Likewise.
6669 (intra_create_variable_infos): Likewise.
6670 (compute_points_to_sets): Likewise.
6671 (ipa_pta_execute): Adjust. Do not push/pop cfun.
6673 2014-04-15 Richard Biener <rguenther@suse.de>
6675 * tree.c (iterative_hash_expr): Use enum tree_code_class
6676 to store TREE_CODE_CLASS.
6677 (tree_block): Likewise.
6678 (tree_set_block): Likewise.
6679 * tree.h (fold_build_pointer_plus_loc): Use
6680 convert_to_ptrofftype_loc.
6682 2014-04-15 Jakub Jelinek <jakub@redhat.com>
6685 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
6688 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
6690 * cfgloop.h (struct loop): Move force_vectorize down.
6691 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
6692 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
6693 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
6694 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
6695 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
6696 * tree-core.h (enum annot_expr_kind): Add new kind values.
6697 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
6698 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
6700 * tree.def (ANNOTATE_EXPR): Tweak comment.
6702 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6704 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
6707 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
6709 * tree.h (TYPE_IDENTIFIER): Declare.
6710 * tree.c (subrange_type_for_debug_p): Use it.
6711 * godump.c (go_format_type): Likewise.
6712 * dwarf2out.c (is_cxx_auto, modified_type_die,
6713 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
6714 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
6716 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6719 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
6721 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
6723 * config/i386/i386.c (examine_argument): Return bool. Return true if
6724 parameter should be passed in memory.
6725 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
6726 (construct_container): Update calls to examine_argument.
6727 (function_arg_advance_64): Ditto.
6728 (return_in_memory_32): Merge with ix86_return_in_memory.
6729 (return_in_memory_64): Ditto.
6730 (return_in_memory_ms_64): Ditto.
6732 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6734 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
6735 * coverage.c (coverage_compute_profile_id): Handle externally visible
6738 2014-04-14 Martin Jambor <mjambor@suse.cz>
6740 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
6741 DECL_DISREGARD_INLINE_LIMITS functions.
6743 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6746 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
6748 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6751 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
6752 optimize_insn_for_speed_p instead of
6753 optimize_function_for_speed_p.
6755 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
6757 * doc/invoke.texi (free): Document AArch64.
6759 2014-04-14 Richard Biener <rguenther@suse.de>
6761 PR tree-optimization/60042
6762 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
6763 (insert_into_preds_of_block): Do not prevent PHI insertion
6764 for REFERENCE exprs here ...
6765 (eliminate_dom_walker::before_dom_children): ... but prevent
6766 their use here under similar conditions when applied to the
6767 IL after PRE optimizations.
6769 2014-04-14 Richard Biener <rguenther@suse.de>
6771 * passes.def: Move early points-to after early SRA.
6773 2014-04-14 Richard Biener <rguenther@suse.de>
6775 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
6776 check for which sign-changes we allow when forwarding
6777 a converted value into a switch.
6779 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6781 * stor-layout.c (place_field): Finalize non-constant offset for the
6784 2014-04-14 Richard Biener <rguenther@suse.de>
6786 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
6788 (expand_switch_using_bit_tests_p): Likewise.
6789 (process_switch): Compute and pass on speed_p based on the
6791 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
6792 optimize_bb_for_speed_p.
6794 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6796 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
6797 * function.h (struct function): Rename has_force_vect_loops into
6798 has_force_vectorize_loops.
6799 * lto-streamer-in.c (input_cfg): Adjust for renaming.
6800 (input_struct_function_base): Likewise.
6801 * lto-streamer-out.c (output_cfg): Likewise.
6802 (output_struct_function_base): Likewise.
6803 * omp-low.c (expand_omp_simd): Likewise.
6804 * tree-cfg.c (move_sese_region_to_fn): Likewise.
6805 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
6806 (version_loop_for_if_conversion): Likewise.
6807 (tree_if_conversion): Likewise.
6808 (main_tree_if_conversion): Likewise.
6809 (gate_tree_if_conversion): Likewise.
6810 * tree-inline.c (copy_loops): Likewise.
6811 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
6812 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
6813 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
6814 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
6815 * tree-vectorizer.c (vectorize_loops): Likewise.
6816 * tree-vectorizer.h (unlimited_cost_model): Likewise.
6818 2014-04-14 Richard Biener <rguenther@suse.de>
6821 * lto-streamer-out.c (wrap_refs): New function.
6822 (lto_output): Wrap symbol references in global initializes in
6823 type-preserving MEM_REFs.
6825 2014-04-14 Christian Bruel <christian.bruel@st.com>
6827 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
6829 2014-04-14 Christian Bruel <christian.bruel@st.com>
6831 * config/sh/sh.md (setmemqi): New expand pattern.
6832 * config/sh/sh.h (CLEAR_RATIO): Define.
6833 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
6834 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
6836 2014-04-14 Richard Biener <rguenther@suse.de>
6839 * fold-const.c (negate_expr_p): Don't negate directional rounding
6841 (fold_negate_expr): Likewise.
6843 2014-04-14 Richard Biener <rguenther@suse.de>
6845 PR tree-optimization/59817
6846 PR tree-optimization/60453
6847 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
6848 recursion to catch all CHRECs in the scalar evolution and restrict
6849 the predicate for the remains appropriately.
6851 2014-04-12 Catherine Moore <clm@codesourcery.com>
6853 * config/mips/constraints.md: Add new register constraint "kb".
6854 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
6855 (*movhi_internal): Likewise.
6856 (*movqi_internal): Likewise.
6857 * config/mips/mips.h (M16_STORE_REGS): New register class.
6858 (REG_CLASS_NAMES): Add M16_STORE_REGS.
6859 (REG_CLASS_CONTENTS): Likewise.
6860 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
6862 2014-04-11 Tobias Burnus <burnus@net-b.de>
6865 * doc/invoke.texi (-Wformat-signedness): Document it.
6866 (Wformat=2): Mention that this enables -Wformat-signedness.
6868 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6870 * common/config/epiphany/epiphany-common.c
6871 (epiphany_option_optimization_table): Enable section anchors by
6872 default at -O1 or higher.
6873 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
6874 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
6875 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
6876 carries no extra cost.
6877 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
6878 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
6879 * config/epiphany/predicates.md (memclob_operand): New predicate.
6880 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
6881 Use memclob_operand predicate and X constraint for operand 3.
6883 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6885 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
6886 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
6889 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6891 PR rtl-optimization/60651
6892 * mode-switching.c (optimize_mode_switching): Make sure to emit
6893 sets of a lower numbered entity before sets of a higher numbered
6894 entity to a mode of the same or lower priority.
6895 When creating a seginfo for a basic block that starts with a code
6896 label, move the insertion point past the code label.
6897 (new_seginfo): Document and enforce requirement that
6898 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
6899 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
6900 * doc/tm.texi: Regenerate.
6902 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
6905 * config/arc/arc.c (arc_save_restore): Fix assert typo.
6907 2013-04-11 Jakub Jelinek <jakub@redhat.com>
6909 * BASE-VER: Set to 4.10.0.
6911 2014-04-11 Tobias Burnus <burnus@net-b.de>
6914 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
6915 * doc/gcc.texi (Service): Update description in the @menu
6916 * doc/invoke.texi (Option Summary): Remove misplaced and
6919 2014-04-11 Steve Ellcey <sellcey@mips.com>
6920 Jakub Jelinek <jakub@redhat.com>
6923 * expr.c (convert_move): Use emit_store_flag_force instead of
6924 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
6927 2014-04-11 Richard Biener <rguenther@suse.de>
6930 * varasm.c (assemble_alias): Avoid endless error reporting
6931 recursion by setting TREE_ASM_WRITTEN.
6933 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6935 * config/s390/s390.md: Add a splitter for NOT rtx.
6937 2014-04-11 Jakub Jelinek <jakub@redhat.com>
6939 PR rtl-optimization/60663
6940 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
6942 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
6943 Jakub Jelinek <jakub@redhat.com>
6946 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
6947 flag from decl_node to node.
6949 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6952 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
6953 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
6954 ameliorating the cases where it can be.
6956 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
6959 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6961 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6962 (loadsync_<mode>): Change mode.
6963 (load_quadpti, store_quadpti): New.
6964 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6965 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6966 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
6968 2014-04-09 Cong Hou <congh@google.com>
6971 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
6974 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6976 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
6977 instead of vnor to exploit possible fusion opportunity in the
6979 (altivec_expand_vec_perm_const_le): Likewise.
6981 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6983 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6984 (loadsync_<mode>): Change mode.
6985 (load_quadpti, store_quadpti): New.
6986 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6987 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6989 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
6992 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
6993 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
6994 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
6996 2014-04-08 Richard Biener <rguenther@suse.de>
6999 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
7000 a 64bit widest int print double-int similar to on HWI64 hosts.
7002 2014-04-08 Richard Biener <rguenther@suse.de>
7004 PR tree-optimization/60785
7005 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
7006 default defs properly.
7008 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
7010 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
7011 (Weffc++): Likewise.
7013 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
7015 * ipa-devirt.c (maybe_record_node): When node is not recorded,
7016 set completep to false rather than true.
7018 2014-04-07 Douglas B Rupp <rupp@adacore.com>
7021 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
7022 ARM_TARGET2_DWARF_FORMAT.
7024 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
7027 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
7028 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
7031 2014-04-07 Richard Biener <rguenther@suse.de>
7033 PR tree-optimization/60766
7034 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
7035 (may_eliminate_iv): Convert cand_value_at result to desired type.
7037 2014-04-07 Jason Merrill <jason@redhat.com>
7040 * common.opt (-fno-gnu-unique): Add.
7041 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
7043 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7045 * haifa-sched.c: Fix outdated function reference and minor
7046 grammar errors in introductory comment.
7048 2014-04-07 Richard Biener <rguenther@suse.de>
7051 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
7053 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
7055 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
7058 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
7059 size accounting for thunks.
7060 (pa_asm_output_mi_thunk): Use final_start_function() and
7061 final_end_function() to output function start and end directives.
7063 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
7065 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
7066 device specific ISA/ feature information. Remove short_sp and
7067 errata_skip ds. Add avr_device_specific_features enum to have device
7069 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
7070 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
7071 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
7072 updated device specific info.
7073 * config/avr/avr-mcus.def: Merge device specific details to
7074 dev_attribute field.
7075 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
7077 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
7078 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
7079 assembler if RMW isa supported by current device.
7080 * config/avr/genmultilib.awk: Update as device info structure changed.
7081 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
7083 2014-04-04 Cong Hou <congh@google.com>
7085 PR tree-optimization/60656
7086 * tree-vect-stmts.c (supportable_widening_operation):
7087 Fix a bug that elements in a vector with vect_used_by_reduction
7088 property are incorrectly reordered when the operation on it is not
7089 consistant with the one in reduction operation.
7091 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
7093 PR rtl-optimization/60155
7094 * gcse.c (record_set_data): New function.
7095 (single_set_gcse): New function.
7096 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
7097 (hoist_code): Likewise.
7098 (get_pressure_class_and_nregs): Likewise.
7100 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
7102 * explow.c (probe_stack_range): Emit a final optimization blockage.
7104 2014-04-04 Anthony Green <green@moxielogic.com>
7106 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
7109 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
7112 * lto-cgraph.c (input_overwrite_node): Check that partitioning
7113 flags are set only during streaming.
7114 * ipa.c (process_references, walk_polymorphic_call_targets,
7115 symtab_remove_unreachable_nodes): Drop bodies of always inline
7116 after early inlining.
7117 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
7119 2014-04-04 Jakub Jelinek <jakub@redhat.com>
7120 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7123 * dwarf2out.c (const_ok_for_output_1): Reject expressions
7126 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7129 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
7131 (cortex_a53_fdivd): Likewise.
7133 2014-04-04 Martin Jambor <mjambor@suse.cz>
7136 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
7138 * cgraph.c (clone_of_p): Also return true if thunks match.
7139 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
7140 cgraph_function_or_thunk_node and an obsolete comment.
7141 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
7143 (build_function_decl_skip_args): Likewise.
7144 (set_new_clone_decl_and_node_flags): New function.
7145 (duplicate_thunk_for_node): Likewise.
7146 (redirect_edge_duplicating_thunks): Likewise.
7147 (cgraph_clone_node): New parameter args_to_skip, pass it to
7148 redirect_edge_duplicating_thunks which is called instead of
7149 cgraph_redirect_edge_callee.
7150 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
7151 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
7153 2014-04-04 Jeff Law <law@redhat.com>
7156 * config/arm/predicates.md (const_int_I_operand): New predicate.
7157 (const_int_M_operand): Similarly.
7158 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
7160 (insv_t2, extv_reg, extzv_t2): Likewise.
7161 (load_multiple_with_writeback): Similarly for const_int_I_operand.
7162 (pop_multiple_with_writeback_and_return): Likewise.
7163 (vfp_pop_multiple_with_writeback): Likewise
7165 2014-04-04 Richard Biener <rguenther@suse.de>
7168 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
7169 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
7171 * gimplify.h (gimple_add_tmp_var_fn): Declare.
7172 * gimplify.c (gimple_add_tmp_var_fn): New function.
7173 * gimple-expr.h (create_tmp_reg_fn): Declare.
7174 * gimple-expr.c (create_tmp_reg_fn): New function.
7175 * gimple-low.c (record_vars_into): Don't change cfun.
7176 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
7177 code generation without cfun.
7179 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
7182 * Makefile.in (install-driver): Fix shell scripting.
7184 2014-04-03 Cong Hou <congh@google.com>
7186 PR tree-optimization/60505
7187 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
7188 threshold of number of iterations below which no vectorization
7190 * tree-vect-loop.c (new_loop_vec_info):
7191 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
7192 * tree-vect-loop.c (vect_analyze_loop_operations):
7193 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
7194 * tree-vect-loop.c (vect_transform_loop):
7195 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
7196 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
7197 of iterations of the loop and see if we should build the epilogue.
7199 2014-04-03 Richard Biener <rguenther@suse.de>
7201 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
7202 (streamer_tree_cache_create): Adjust.
7203 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
7204 to allow optional nodes array.
7205 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
7206 (streamer_tree_cache_append): Likewise.
7207 (streamer_tree_cache_create): Create nodes array optionally
7208 as specified by parameter.
7209 * lto-streamer-out.c (create_output_block): Avoid maintaining
7210 the node array in the writer cache.
7211 (DFS_write_tree): Remove assertion.
7212 (produce_asm_for_decls): Free the out decl state hash table early.
7213 * lto-streamer-in.c (lto_data_in_create): Adjust for
7214 streamer_tree_cache_create prototype change.
7216 2014-04-03 Richard Biener <rguenther@suse.de>
7218 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
7219 set TREE_CHAIN to NULL_TREE.
7221 2014-04-03 Richard Biener <rguenther@suse.de>
7223 PR tree-optimization/60740
7224 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
7225 over all GIMPLE_COND operands.
7227 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
7229 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
7230 (Weffc++): Remove Scott's numbering, merge lists and reference
7233 2014-04-03 Nick Clifton <nickc@redhat.com>
7235 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
7238 2014-04-03 Martin Jambor <mjambor@suse.cz>
7240 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
7241 mention gcc_unreachable before failing.
7242 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
7245 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
7248 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
7249 inconsistent code and instead mark the context inconsistent.
7250 (possible_polymorphic_call_targets): For inconsistent contexts
7251 return empty complete list.
7253 2014-04-02 Anthony Green <green@moxielogic.com>
7255 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
7256 (extendqisi2, extendhisi2): Define.
7257 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
7258 (WCHAR_TYPE): Change to unsigned int.
7260 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7262 PR tree-optimization/60733
7263 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
7264 insertion point for PHI candidates to be the end of the feeding
7265 block for the PHI argument.
7267 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
7269 PR rtl-optimization/60650
7270 * lra-constraints.c (process_alt_operands): Decrease reject for
7271 earlyclobber matching.
7273 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7275 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
7277 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7279 * config/spu/spu.c (pad_bb): Do not crash when the last
7280 insn is CODE_FOR_blockage.
7282 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7284 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
7285 lies outside the target mode.
7287 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7290 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7291 software floating point or no floating point registers, do not
7292 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7293 in GPRs that occurs after we tested for GPRs that would never be
7296 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7297 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7298 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7299 specifically allow DDmode, since that does not use the SPE SIMD
7302 2014-04-02 Richard Biener <rguenther@suse.de>
7305 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
7306 MODE_INTs. Properly use negv_optab.
7307 (expand_abs): Likewise.
7309 2014-04-02 Richard Biener <rguenther@suse.de>
7312 * Makefile.in (install-driver): Guard extra installs with special
7315 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
7317 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7320 2014-04-01 Richard Henderson <rth@redhat.com>
7323 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
7324 alternative enabled before register allocation.
7326 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
7328 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
7329 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
7331 (nios2_large_got_address): Remove unneeded 'sym' parameter.
7332 (nios2_got_address): Update nios2_large_got_address call site.
7333 (nios2_delegitimize_address): New function.
7334 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
7335 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
7336 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
7338 2014-04-01 Martin Husemann <martin@duskware.de>
7340 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
7343 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
7345 PR rtl-optimization/60604
7346 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
7347 check from register_operand.
7348 (register_operand): Redefine in terms of general_operand.
7349 (nonmemory_operand): Use register_operand for the non-constant cases.
7351 2014-04-01 Richard Biener <rguenther@suse.de>
7353 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
7355 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
7357 * doc/invoke.texi (mapp-regs): Clarify.
7359 2014-03-31 Ulrich Drepper <drepper@gmail.com>
7361 * config/i386/avx512fintrin.h (__v32hi): Define type.
7362 (__v64qi): Likewise.
7363 (_mm512_set1_epi8): Define.
7364 (_mm512_set1_epi16): Define.
7365 (_mm512_set4_epi32): Define.
7366 (_mm512_set4_epi64): Define.
7367 (_mm512_set4_pd): Define.
7368 (_mm512_set4_ps): Define.
7369 (_mm512_setr4_epi64): Define.
7370 (_mm512_setr4_epi32): Define.
7371 (_mm512_setr4_pd): Define.
7372 (_mm512_setr4_ps): Define.
7373 (_mm512_setzero_epi32): Define.
7375 2014-03-31 Martin Jambor <mjambor@suse.cz>
7378 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
7379 callsite_arguments_match_p. Updated all callers. Also check types of
7380 corresponding formal parameters and actual arguments.
7381 (not_all_callers_have_enough_arguments_p) Renamed to
7382 some_callers_have_mismatched_arguments_p.
7384 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
7386 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
7388 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
7391 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
7394 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
7396 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
7397 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
7399 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
7400 Use FMAMODE_NOVF512 mode iterator.
7401 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
7402 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
7403 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
7404 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
7406 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
7407 Use VF_128_256 mode iterator.
7408 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
7411 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7413 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
7414 static chain if needed.
7416 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7419 * lra-constraints.c (index_part_to_reg): New.
7420 (process_address): Use it.
7422 2014-03-27 Jeff Law <law@redhat.com>
7423 Jakub Jelinek <jakub@redhat.com>
7426 * expr.c (do_tablejump): Use simplify_gen_binary rather than
7427 gen_rtx_{PLUS,MULT} to build up the address expression.
7429 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
7430 creating non-canonical RTL.
7432 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7435 * ipa-inline.c (want_inline_small_function_p): Short circuit large
7436 functions; reorganize to make cheap checks first.
7437 (inline_small_functions): Do not estimate growth when dumping;
7439 * ipa-inline.h (inline_summary): Add min_size.
7440 (growth_likely_positive): New function.
7441 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
7442 (set_cond_stmt_execution_predicate): Cleanup.
7443 (estimate_edge_size_and_time): Compute min_size.
7444 (estimate_calls_size_and_time): Likewise.
7445 (estimate_node_size_and_time): Likewise.
7446 (inline_update_overall_summary): Update min_size.
7447 (do_estimate_edge_time): Likewise.
7448 (do_estimate_edge_size): Update.
7449 (do_estimate_edge_hints): Update.
7450 (growth_likely_positive): New function.
7452 2014-03-28 Jakub Jelinek <jakub@redhat.com>
7455 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
7456 also if addr has VOIDmode.
7458 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7460 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
7461 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
7463 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
7464 instructions as well as AdvancedSIMD loads.
7466 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7468 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
7469 Use crypto_aese type.
7470 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
7471 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
7472 crypto_aese, crypto_aesmc. Move to types.md.
7473 * config/arm/types.md (crypto_aes): Split into crypto_aese,
7475 * config/arm/iterators.md (crypto_type): Likewise.
7477 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7479 * cgraph.c: Include expr.h and tree-dfa.h.
7480 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
7483 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7486 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
7487 regs from checking multi-reg pseudos.
7489 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7491 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
7493 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7495 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
7496 if it would clobber the stack pointer, even temporarily.
7498 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
7500 * mode-switching.c: Make small adjustments to the top comment.
7502 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7504 * config/rs6000/constraints.md (wD constraint): New constraint to
7505 match the constant integer to get the top DImode/DFmode out of a
7506 vector in a VSX register.
7508 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
7509 match the constant integer to get the top DImode/DFmode out of a
7510 vector in a VSX register.
7512 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
7515 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7518 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
7519 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
7521 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
7522 Optimize vec_extract of 64-bit values, where the value being
7523 extracted is in the top word, where we can use scalar
7524 instructions. Add direct move and store support. Combine the big
7525 endian/little endian vector select load support into a single insn.
7526 (vsx_extract_<mode>_internal1): Likewise.
7527 (vsx_extract_<mode>_internal2): Likewise.
7528 (vsx_extract_<mode>_load): Likewise.
7529 (vsx_extract_<mode>_store): Likewise.
7530 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7531 combined into vsx_extract_<mode>_load.
7532 (vsx_extract_<mode>_one_le): Likewise.
7534 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7535 define the top 64-bit vector element.
7537 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7540 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7541 Document vec_vbpermq builtin.
7544 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7545 enable use of xxsldwi and xxpermdi builtin functions.
7546 (vec_xxpermdi): Likewise.
7548 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7549 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7551 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7553 PR rtl-optimization/60650
7554 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7556 (find_spills_for): New.
7557 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7558 Spill all pseudos on the second iteration.
7560 2014-03-27 Marek Polacek <polacek@redhat.com>
7563 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7566 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7568 * config/s390/s390.c (s390_can_use_return_insn): Check for
7569 call-saved FPRs on 31 bit.
7571 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7574 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7575 if they need regimplification, just drop them instead of
7576 calling gimple_regimplify_operands on them.
7578 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7581 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7582 (aarch64_frame_pointer_required): Adjust logic.
7583 (aarch64_can_eliminate): Adjust logic.
7584 (aarch64_override_options_after_change): Adjust logic.
7586 2014-03-27 Dehao Chen <dehao@google.com>
7588 * ipa-inline.c (early_inliner): Update node's inline info.
7590 2014-03-26 Dehao Chen <dehao@google.com>
7592 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
7593 compiler inserted conditional jumps for NAN float check.
7595 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7597 * ubsan.h (ubsan_create_data): Change second argument's type
7598 to const location_t *.
7599 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
7601 (ubsan_create_data): Change second argument to const location_t *PLOC.
7602 Create Loc field whenever PLOC is non-NULL.
7603 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
7604 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
7608 * real.c (real_to_integer2): Change type of low to UHWI.
7610 2014-03-26 Tobias Burnus <burnus@net-b.de>
7612 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
7613 (CILK_SELF_SPECS): New define.
7614 (driver_self_specs): Use it.
7616 2014-03-26 Richard Biener <rguenther@suse.de>
7618 * tree-pretty-print.c (percent_K_format): Implement special
7619 case for LTO and its stripped down BLOCK tree.
7621 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7624 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
7626 * tree-vrp.c (simplify_internal_call_using_ranges): If only
7627 one range is range_int_cst_p, but not both, at least optimize
7628 addition/subtraction of 0 and multiplication by 0 or 1.
7629 * gimple-fold.c (gimple_fold_call): Fold
7630 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
7631 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
7632 INTEGER_CSTs, try to fold at least x * 0 and y - y.
7634 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
7636 PR rtl-optimization/60452
7637 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
7638 <case REG>: Return 1 for invalid offsets from the frame pointer.
7640 2014-03-26 Marek Polacek <polacek@redhat.com>
7643 * doc/extend.texi (C Extensions): Mention variable-length arrays in
7646 2014-03-26 Marek Polacek <polacek@redhat.com>
7649 * doc/extend.texi (Designated Inits): Describe what happens to omitted
7652 2014-03-26 Marek Polacek <polacek@redhat.com>
7655 * ira-color.c (update_conflict_hard_regno_costs): Perform the
7656 multiplication in unsigned type.
7658 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7660 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
7662 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7664 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
7666 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7669 * cif-code.def (UNREACHABLE) New code.
7670 * ipa-inline.c (inline_small_functions): Skip edges to
7671 __builtlin_unreachable.
7672 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
7673 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
7674 predicate to __bulitin_unreachable.
7675 (set_cond_stmt_execution_predicate): Fix issue when
7676 invert_tree_comparison returns ERROR_MARK.
7677 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
7678 propagate to inline clones.
7679 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
7681 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
7682 * cgraphclones.c (cgraph_clone_node): If call destination is already
7683 ureachable, do not redirect it back.
7684 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
7687 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7689 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
7690 Do not modify inline clones.
7692 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7694 * config/i386/i386.md (general_sext_operand): New mode attr.
7695 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
7696 don't generate (sign_extend (const_int)).
7697 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
7698 operands[2]. Use We constraint instead of <i> and
7699 <general_sext_operand> predicate instead of <general_operand>.
7700 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
7701 * config/i386/constraints.md (We): New constraint.
7702 * config/i386/predicates.md (x86_64_sext_operand,
7703 sext_operand): New predicates.
7705 2014-03-25 Martin Jambor <mjambor@suse.cz>
7708 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
7709 inconsistent devirtualizations to __builtin_unreachable.
7711 2014-03-25 Marek Polacek <polacek@redhat.com>
7714 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
7716 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
7718 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
7719 order of elements for big-endian.
7721 2014-03-25 Richard Biener <rguenther@suse.de>
7724 * gimplify-me.c (gimple_regimplify_operands): Update the
7727 2014-03-25 Martin Jambor <mjambor@suse.cz>
7730 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
7731 (lto_output_varpool_node): Likewise.
7732 (input_overwrite_node): Likewise.
7733 (input_varpool_node): Likewise.
7735 2014-03-25 Richard Biener <rguenther@suse.de>
7737 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
7738 (run_gcc): Likewise.
7740 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7742 * combine.c (simplify_compare_const): Add MODE argument.
7743 Handle mode_width 0 as very large mode_width.
7744 (try_combine, simplify_comparison): Adjust callers.
7746 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
7747 type to avoid signed integer overflow.
7748 * explow.c (plus_constant): Likewise.
7750 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
7752 * doc/generic.texi: Correct typos.
7754 2014-03-24 Tobias Burnus <burnus@net-b.de>
7756 * doc/invoke.texi (-flto): Expand section about
7757 using static libraries with LTO.
7759 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7761 PR rtl-optimization/60501
7762 * optabs.def (addptr3_optab): New optab.
7763 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
7764 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
7765 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
7767 * lra.c (emit_add3_insn): Use the addptr pattern if available.
7769 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
7771 2014-03-24 Ulrich Drepper <drepper@gmail.com>
7773 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
7776 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
7777 (_mm256_undefined_ps): Define.
7778 (_mm256_undefined_pd): Define.
7779 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
7780 (_mm_undefined_pd): Define.
7781 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
7782 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
7783 (_mm512_undefined_ps): Define.
7784 (_mm512_undefined_pd): Define.
7785 Use _mm*_undefined_*.
7786 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
7788 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
7790 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
7791 (lshr_simd): DI mode added.
7792 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
7793 (aarch64_ushr_simddi): Likewise.
7794 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
7795 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
7796 (vshrd_n_u64): Likewise.
7798 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7800 * Makefile.in (s-macro_list): Depend on cc1.
7802 2014-03-23 Teresa Johnson <tejohnson@google.com>
7804 * ipa-utils.c (ipa_print_order): Use specified dump file.
7806 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
7808 PR rtl-optimization/60601
7809 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
7811 * gcc.c (eval_spec_function): Initialize save_growing_value.
7813 2014-03-22 Jakub Jelinek <jakub@redhat.com>
7816 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
7817 code == MINUS_EXPR, never swap op0 with op1.
7819 * toplev.c (init_local_tick): Avoid signed integer multiplication
7821 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
7822 shift by first operand's bitsize.
7824 2014-03-21 Jakub Jelinek <jakub@redhat.com>
7827 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
7829 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
7830 TARGET_ISA_64BIT_P(x).
7832 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7834 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
7835 pattern for vector nor instead of subtract from splat(-1).
7836 (altivec_expand_vec_perm_const_le): Likewise.
7838 2014-03-21 Richard Henderson <rth@twiddle.net>
7841 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
7842 related insns after epilogue_completed.
7844 2014-03-21 Martin Jambor <mjambor@suse.cz>
7847 * cgraph.h (symtab_node): New flag body_removed.
7848 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
7849 when removing bodies.
7850 * symtab.c (dump_symtab_base): Dump body_removed flag.
7851 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
7852 had their bodies removed.
7854 2014-03-21 Martin Jambor <mjambor@suse.cz>
7857 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
7860 2014-03-21 Richard Biener <rguenther@suse.de>
7862 PR tree-optimization/60577
7863 * tree-core.h (struct tree_base): Document nothrow_flag use
7864 in VAR_DECL_NONALIASED.
7865 * tree.h (VAR_DECL_NONALIASED): New.
7866 (may_be_aliased): Adjust.
7867 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
7869 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7871 * expr.c (expand_expr_real_1): Remove outdated comment.
7873 2014-03-20 Jakub Jelinek <jakub@redhat.com>
7876 * ira.c (adjust_cleared_regs): Call copy_rtx on
7877 *reg_equiv[REGNO (loc)].src_p before passing it to
7878 simplify_replace_fn_rtx.
7881 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
7882 into CONST, put pic register as first operand of PLUS. Use
7883 gen_const_mem for both 32-bit and 64-bit PIC got loads.
7885 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7887 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
7889 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7891 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
7892 around for store forwarding issue in the FPU on the UT699.
7893 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
7894 loads and operations if -mfix-ut699 is specified.
7895 (divtf3_hq): Tweak attribute.
7896 (sqrttf2_hq): Likewise.
7898 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7900 * calls.c (store_one_arg): Remove incorrect const qualification on the
7901 type of the temporary.
7902 * cfgexpand.c (expand_return): Likewise.
7903 * expr.c (expand_constructor): Likewise.
7904 (expand_expr_real_1): Likewise.
7906 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7908 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
7911 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
7914 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
7916 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
7918 * config/arm/aarch-common-protos.h
7919 (alu_cost_table): Fix spelling of "extend".
7920 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
7922 2014-03-19 Richard Biener <rguenther@suse.de>
7925 * tree-core.h (tree_type_common): Re-order pointer members
7926 to reduce recursion depth during GC walks.
7928 2014-03-19 Marek Polacek <polacek@redhat.com>
7931 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
7932 before accessing it.
7934 2014-03-19 Richard Biener <rguenther@suse.de>
7937 * lto-streamer-in.c (input_function): In WPA stage do not drop
7940 2014-03-19 Jakub Jelinek <jakub@redhat.com>
7942 PR tree-optimization/60559
7943 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
7944 with build_zero_cst assignment.
7946 2014-03-18 Kai Tietz <ktietz@redhat.com>
7948 PR rtl-optimization/56356
7949 * sdbout.c (sdbout_parms): Verify that parms'
7950 incoming argument is valid.
7951 (sdbout_reg_parms): Likewise.
7953 2014-03-18 Richard Henderson <rth@redhat.com>
7956 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
7957 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
7958 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
7960 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
7962 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
7963 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
7964 Italicize plugin event names in description. Explain that
7965 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
7966 Remind that no GCC functions should be called after PLUGIN_FINISH.
7967 Explain what pragmas with expansion are.
7969 2014-03-18 Martin Liska <mliska@suse.cz>
7971 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
7972 gimple call statement is update.
7973 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
7974 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
7976 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7979 * ubsan.c (ubsan_instrument_unreachable): Call
7980 initialize_sanitizer_builtins.
7981 (ubsan_pass): Likewise.
7984 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
7985 varpool_finalize_decl instead of rest_of_decl_compilation.
7987 2014-03-18 Richard Biener <rguenther@suse.de>
7989 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
7990 by using bitmap_and_compl instead of bitmap_and_compl_into.
7991 (df_rd_transfer_function): Likewise.
7993 2014-03-18 Richard Biener <rguenther@suse.de>
7995 * doc/lto.texi (fresolution): Fix typo.
7997 2014-03-18 Richard Biener <rguenther@suse.de>
7999 * doc/invoke.texi (flto): Update for changes in 4.9.
8001 2014-03-18 Richard Biener <rguenther@suse.de>
8003 * doc/loop.texi: Remove section on the removed lambda framework.
8004 Update loop docs with recent changes in preserving loop structure.
8006 2014-03-18 Richard Biener <rguenther@suse.de>
8008 * doc/lto.texi (-fresolution): Document.
8010 2014-03-18 Richard Biener <rguenther@suse.de>
8012 * doc/contrib.texi: Adjust my name.
8014 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8017 * internal-fn.c: Include diagnostic-core.h.
8018 (expand_BUILTIN_EXPECT): New function.
8019 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
8020 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
8021 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
8022 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
8024 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
8025 Revert 3 argument __builtin_expect code.
8026 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
8027 * gimple-fold.c (gimple_fold_call): Likewise.
8028 * tree.h (fold_builtin_expect): New prototype.
8029 * builtins.c (build_builtin_expect_predicate): Add predictor
8030 argument, if non-NULL, create 3 argument __builtin_expect.
8031 (fold_builtin_expect): No longer static. Add ARG2 argument,
8032 pass it through to build_builtin_expect_predicate.
8033 (fold_builtin_2): Adjust caller.
8034 (fold_builtin_3): Handle BUILT_IN_EXPECT.
8035 * internal-fn.def (BUILTIN_EXPECT): New.
8037 2014-03-18 Tobias Burnus <burnus@net-b.de>
8040 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
8041 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
8042 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
8044 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
8047 * predict.c (combine_predictions_for_bb): Fix up formatting.
8048 (expr_expected_value_1, expr_expected_value): Add predictor argument,
8049 fill what it points to if non-NULL.
8050 (tree_predict_by_opcode): Adjust caller, use the predictor.
8051 * predict.def (PRED_COMPARE_AND_SWAP): Add.
8053 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
8055 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
8056 proper constant for the store mode.
8058 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
8060 * symtab.c (change_decl_assembler_name): Fix transparent alias
8063 2014-03-16 Renlin Li <Renlin.Li@arm.com>
8065 * config/aarch64/aarch64.c: Correct the comments about the
8066 aarch64 stack layout.
8068 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
8070 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
8071 check for GF_OMP_FOR_KIND_FOR.
8073 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
8075 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
8076 ymm and zmm register names.
8078 2014-03-17 Jakub Jelinek <jakub@redhat.com>
8081 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
8082 note creation for the 2010-08-31 changes.
8084 2014-03-17 Marek Polacek <polacek@redhat.com>
8087 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
8088 as -fno-tree-loop-vectorize.
8089 (expand_omp_simd): Likewise.
8091 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
8093 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
8094 (eligible_for_call_delay): New prototype.
8095 * config/sparc/sparc.c (tls_call_delay): Rename into...
8096 (eligible_for_call_delay): ...this. Return false if the instruction
8097 cannot be put in the delay slot of a branch.
8098 (eligible_for_restore_insn): Simplify.
8099 (eligible_for_return_delay): Return false if the instruction cannot be
8100 put in the delay slot of a branch and simplify.
8101 (eligible_for_sibcall_delay): Return false if the instruction cannot be
8102 put in the delay slot of a branch.
8103 * config/sparc/sparc.md (fix_ut699): New attribute.
8104 (tls_call_delay): Delete.
8105 (in_call_delay): Reimplement.
8106 (eligible_for_sibcall_delay): Rename into...
8107 (in_sibcall_delay): ...this.
8108 (eligible_for_return_delay): Rename into...
8109 (in_return_delay): ...this.
8110 (in_branch_delay): Reimplement.
8111 (in_uncond_branch_delay): Delete.
8112 (in_annul_branch_delay): Delete.
8114 2014-03-14 Richard Henderson <rth@redhat.com>
8117 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
8118 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
8119 (*floathi<X87MODEF>2_i387_with_temp): Remove.
8120 (floathi splitters): Remove.
8121 (float<SWI48x>xf2): New pattern.
8122 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
8123 code that tried to handle DImode for 32-bit, but which was excluded
8124 by the pattern's condition. Drop allocation of stack temporary.
8125 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
8126 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
8127 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
8128 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
8129 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
8130 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
8131 (*float<SWI48><MODEF>2_sse_interunit): Remove.
8132 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
8133 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
8134 (*float<SWI48x><X87MODEF>2_i387): Remove.
8135 (all float _with_temp splitters): Remove.
8136 (*float<SWI48x><MODEF>2_i387): New pattern.
8137 (*float<SWI48><MODEF>2_sse): New pattern.
8138 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
8139 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
8141 2014-03-14 Jakub Jelinek <jakub@redhat.com>
8142 Marek Polacek <polacek@redhat.com>
8145 * common.opt (dump_base_name_prefixed): New Variable.
8146 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
8147 if x_dump_base_name_prefixed is already set, set it at the end.
8149 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
8151 PR rtl-optimization/60508
8152 * lra-constraints.c (get_reload_reg): Add new parameter
8154 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
8155 Pass the new parameter values.
8157 2014-03-14 Richard Biener <rguenther@suse.de>
8159 * common.opt: Revert unintented changes from r205065.
8162 2014-03-14 Richard Biener <rguenther@suse.de>
8165 * cfghooks.c (split_block): Properly adjust all loops the
8166 block was a latch of.
8168 2014-03-14 Martin Jambor <mjambor@suse.cz>
8171 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
8174 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
8177 * config/avr/avr.c (avr_set_current_function): Pass function name
8178 through default_strip_name_encoding before sanity checking instead
8179 of skipping the first char of the assembler name.
8181 2014-03-13 Richard Henderson <rth@redhat.com>
8184 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
8185 (ix86_force_to_memory, ix86_free_from_memory): Remove.
8186 * config/i386/i386-protos.h: Likewise.
8187 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
8188 in the expander instead of a splitter.
8189 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
8190 any possibility of requiring a memory.
8191 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
8192 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
8193 (fp branch splitters): Update for ix86_split_fp_branch.
8194 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
8195 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
8196 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
8197 (*fop_<MODEF>_2_i387): Remove f/r alternative.
8198 (*fop_<MODEF>_3_i387): Likewise.
8199 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
8200 (splitters for the fop_* register patterns): Remove.
8201 (fscalexf4_i387): Rename from *fscalexf4_i387.
8202 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
8204 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8206 PR tree-optimization/59779
8207 * tree-dfa.c (get_ref_base_and_extent): Use double_int
8208 type for bitsize and maxsize instead of HOST_WIDE_INT.
8210 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
8212 PR rtl-optimization/57320
8213 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
8214 the CFG after thread_prologue_and_epilogue_insns.
8216 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
8218 PR rtl-optimization/57189
8219 * lra-constraints.c (process_alt_operands): Disfavor spilling
8222 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
8224 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
8226 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8228 PR tree-optimization/59025
8230 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
8231 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
8233 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
8236 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
8237 calls of avr_out_plus_1.
8239 2014-03-13 Bin Cheng <bin.cheng@arm.com>
8241 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
8242 BB's single pred and update the father loop's latch info later.
8244 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8246 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
8250 (VEC_base): Likewise.
8251 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
8252 registers, we need to swap double words in little endian mode.
8254 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
8255 to be a container mode for 128-bit integer operations added in ISA
8256 2.07. Unlike TImode and PTImode, the preferred register set is
8257 the Altivec/VMX registers for the 128-bit operations.
8259 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
8261 (rs6000_split_128bit_ok_p): Likewise.
8263 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
8264 macros for creating ISA 2.07 normal and overloaded builtin
8265 functions with 3 arguments.
8266 (BU_P8V_OVERLOAD_3): Likewise.
8267 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
8268 for use as overloaded functions.
8269 (VPERM_1TI_UNS): Likewise.
8270 (VSEL_1TI): Likewise.
8271 (VSEL_1TI_UNS): Likewise.
8272 (ST_INTERNAL_1ti): Likewise.
8273 (LD_INTERNAL_1ti): Likewise.
8274 (XXSEL_1TI): Likewise.
8275 (XXSEL_1TI_UNS): Likewise.
8276 (VPERM_1TI): Likewise.
8277 (VPERM_1TI_UNS): Likewise.
8278 (XXPERMDI_1TI): Likewise.
8279 (SET_1TI): Likewise.
8280 (LXVD2X_V1TI): Likewise.
8281 (STXVD2X_V1TI): Likewise.
8282 (VEC_INIT_V1TI): Likewise.
8283 (VEC_SET_V1TI): Likewise.
8284 (VEC_EXT_V1TI): Likewise.
8285 (EQV_V1TI): Likewise.
8286 (NAND_V1TI): Likewise.
8287 (ORC_V1TI): Likewise.
8288 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
8289 added in ISA 2.07. Add both normal 'altivec' builtins, and the
8291 (VADDUQM): Likewise.
8292 (VSUBCUQ): Likewise.
8293 (VADDEUQM): Likewise.
8294 (VADDECUQ): Likewise.
8295 (VSUBEUQM): Likewise.
8296 (VSUBECUQ): Likewise.
8298 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
8299 __int128_t and __uint128_t types.
8300 (__uint128_type): Likewise.
8301 (altivec_categorize_keyword): Add support for vector __int128_t,
8302 vector __uint128_t, vector __int128, and vector unsigned __int128
8303 as a container type for TImode operations that need to be done in
8304 VSX/Altivec registers.
8305 (rs6000_macro_to_expand): Likewise.
8306 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
8307 to support 128-bit integer instructions vaddcuq, vadduqm,
8308 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
8309 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
8311 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
8312 for V1TImode, and set up preferences to use VSX/Altivec registers.
8313 Setup VSX reload handlers.
8314 (rs6000_debug_reg_global): Likewise.
8315 (rs6000_init_hard_regno_mode_ok): Likewise.
8316 (rs6000_preferred_simd_mode): Likewise.
8317 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
8318 (easy_altivec_constant): Likewise.
8319 (output_vec_const_move): Likewise.
8320 (rs6000_expand_vector_set): Convert V1TImode set and extract to
8322 (rs6000_expand_vector_extract): Likewise.
8323 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
8325 (rs6000_const_vec): Add support for V1TImode.
8326 (rs6000_emit_le_vsx_load): Swap double words when loading or
8327 storing TImode/V1TImode.
8328 (rs6000_emit_le_vsx_store): Likewise.
8329 (rs6000_emit_le_vsx_move): Likewise.
8330 (rs6000_emit_move): Add support for V1TImode.
8331 (altivec_expand_ld_builtin): Likewise.
8332 (altivec_expand_st_builtin): Likewise.
8333 (altivec_expand_vec_init_builtin): Likewise.
8334 (altivec_expand_builtin): Likewise.
8335 (rs6000_init_builtins): Add support for V1TImode type. Add
8336 support for ISA 2.07 128-bit integer builtins. Define type names
8337 for the VSX/Altivec vector types.
8338 (altivec_init_builtins): Add support for overloaded vector
8339 functions with V1TImode type.
8340 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
8341 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
8343 (rs6000_split_128bit_ok_p): Likewise.
8344 (rs6000_handle_altivec_attribute): Create V1TImode from vector
8345 __int128_t and vector __uint128_t.
8347 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
8348 and mode attributes.
8355 (VS_scalar): Likewise.
8356 (VS_double): Likewise.
8357 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
8359 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
8360 we support the ISA 2.07 128-bit integer arithmetic instructions.
8361 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
8362 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
8363 and TImode types for use with the builtin functions.
8364 (V1TI_type_node): Likewise.
8365 (unsigned_V1TI_type_node): Likewise.
8366 (intTI_type_internal_node): Likewise.
8367 (uintTI_type_internal_node): Likewise.
8369 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
8370 128-bit builtin functions.
8371 (UNSPEC_VADDEUQM): Likewise.
8372 (UNSPEC_VADDECUQ): Likewise.
8373 (UNSPEC_VSUBCUQ): Likewise.
8374 (UNSPEC_VSUBEUQM): Likewise.
8375 (UNSPEC_VSUBECUQ): Likewise.
8376 (VM): Add V1TImode to vector mode iterators.
8378 (VI_unit): Likewise.
8379 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
8380 (altivec_vaddcuq): Likewise.
8381 (altivec_vsubuqm): Likewise.
8382 (altivec_vsubcuq): Likewise.
8383 (altivec_vaddeuqm): Likewise.
8384 (altivec_vaddecuq): Likewise.
8385 (altivec_vsubeuqm): Likewise.
8386 (altivec_vsubecuq): Likewise.
8388 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
8390 (BOOL_128): Likewise.
8391 (BOOL_REGS_OUTPUT): Likewise.
8392 (BOOL_REGS_OP1): Likewise.
8393 (BOOL_REGS_OP2): Likewise.
8394 (BOOL_REGS_UNARY): Likewise.
8395 (BOOL_REGS_AND_CR0): Likewise.
8397 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
8398 128-bit integer builtin support.
8399 (vec_vadduqm): Likewise.
8400 (vec_vaddecuq): Likewise.
8401 (vec_vaddeuqm): Likewise.
8402 (vec_vsubecuq): Likewise.
8403 (vec_vsubeuqm): Likewise.
8404 (vec_vsubcuq): Likewise.
8405 (vec_vsubuqm): Likewise.
8407 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8408 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
8409 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
8410 128-bit integer add/subtract to ISA 2.07.
8412 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
8414 * config/arc/arc.c (arc_predicate_delay_insns):
8415 Fix third argument passed to conditionalize_nonjump.
8417 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
8419 * config/aarch64/aarch64-builtins.c
8420 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
8421 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
8422 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
8423 instead of __builtin_lfloor.
8424 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
8426 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8428 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
8429 (tree_ssa_ifcombine_bb_1): New function.
8430 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
8431 is an empty forwarder block to then_bb or vice versa and then_bb
8432 and else_bb are effectively swapped.
8434 2014-03-12 Christian Bruel <christian.bruel@st.com>
8437 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
8438 REG_CFA_DEF_CFA note.
8439 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
8440 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
8442 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
8444 PR tree-optimization/60454
8445 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
8447 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8449 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
8450 Do not define target_cpu_default2 to generic.
8451 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
8452 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
8453 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
8455 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8456 Marc Glisse <marc.glisse@inria.fr>
8458 PR tree-optimization/60502
8459 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
8460 instead of build_low_bits_mask.
8462 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8465 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
8466 if there are multiple uses, but op doesn't live on E edge.
8467 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
8468 clobber stmts before __builtin_unreachable.
8470 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
8472 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
8473 hard_frame_pointer_rtx.
8474 * cse.c (cse_insn): Remove volatile check.
8475 * cselib.c (cselib_process_insn): Likewise.
8476 * dse.c (scan_insn): Likewise.
8478 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8480 * config/arc/arc.c (conditionalize_nonjump): New function,
8482 (arc_ifcvt): ... this.
8483 (arc_predicate_delay_insns): Use it.
8485 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8487 * config/arc/predicates.md (extend_operand): During/after reload,
8488 allow const_int_operand.
8489 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
8490 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
8491 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
8493 (umulsi3_highpart_i): Likewise.
8495 2014-03-11 Richard Biener <rguenther@suse.de>
8497 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
8498 Add asserts to guard possible wrong-code bugs.
8500 2014-03-11 Richard Biener <rguenther@suse.de>
8502 PR tree-optimization/60429
8503 PR tree-optimization/60485
8504 * tree-ssa-structalias.c (set_union_with_increment): Properly
8505 take into account all fields that overlap the shifted vars.
8506 (do_sd_constraint): Likewise.
8507 (do_ds_constraint): Likewise.
8508 (get_constraint_for_ptr_offset): Likewise.
8510 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
8512 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
8513 (nios2_compute_frame_layout):
8514 Add calculation of cfun->machine->fp_save_offset.
8515 (nios2_expand_prologue): Correct setting of frame pointer register
8517 (nios2_expand_epilogue): Update recovery of stack pointer from
8518 frame pointer accordingly.
8519 (nios2_initial_elimination_offset): Update calculation of offset
8520 for eliminating to HARD_FRAME_POINTER_REGNUM.
8522 2014-03-10 Jakub Jelinek <jakub@redhat.com>
8525 * ipa.c (symtab_remove_unreachable_nodes): Don't call
8526 cgraph_get_create_node on VAR_DECLs.
8528 2014-03-10 Richard Biener <rguenther@suse.de>
8531 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8533 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8535 * config/vms/vms.opt (vms_float_format): New variable.
8537 2014-03-08 Tobias Burnus <burnus@net-b.de>
8539 * doc/invoke.texi (-fcilkplus): Update implementation status.
8541 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8542 Richard Biener <rguenther@suse.de>
8544 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8545 consistently accross all TUs.
8546 (run_gcc): Enable -fshort-double automatically at link at link-time
8547 and disallow override.
8549 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8552 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8553 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8554 if they can't be used.
8556 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8558 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8559 for Solaris 11/x86 ld.
8560 * configure: Regenerate.
8562 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8564 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8565 (LIB_TLS_SPEC): Save as ld_tls_libs.
8566 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8567 (HAVE_AS_IX86_TLSLDM): New test.
8568 * configure, config.in: Regenerate.
8569 * config/i386/i386.c (legitimize_tls_address): Fall back to
8570 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8571 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8572 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8573 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8575 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8577 * common.opt (fira-loop-pressure): Mark as optimization.
8579 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8581 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8582 an OpenMP mappable type.
8584 2014-03-06 Matthias Klose <doko@ubuntu.com>
8586 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
8587 MULTILIB_OSDIRNAMES is not defined.
8589 2014-03-06 Jakub Jelinek <jakub@redhat.com>
8590 Meador Inge <meadori@codesourcery.com>
8593 * config/arm/arm.c (arm_tls_symbol_p): Remove.
8594 (arm_legitimize_address): Call legitimize_tls_address for any
8595 arm_tls_referenced_p expression, handle constant addend. Call it
8596 before testing for !TARGET_ARM.
8597 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
8599 2014-03-06 Richard Biener <rguenther@suse.de>
8605 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8607 * tree-streamer.c (record_common_node): Assert we don't record
8608 nodes with type double.
8609 (preload_common_node): Skip type double, complex double and double
8610 pointer since it is now frontend dependent due to fshort-double option.
8612 2014-03-06 Richard Biener <rguenther@suse.de>
8614 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
8615 or -fno-lto is specified and the linker has full plugin support.
8616 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
8617 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
8618 * lto-wrapper.c (merge_and_complain): Merge compile-time
8619 optimization levels.
8620 (run_gcc): And pass it through to the link options.
8622 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
8626 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8628 * cselib.c (remove_useless_values): Skip to avoid quadratic
8629 behavior if the condition moved from...
8630 (cselib_process_insn): ... here holds.
8632 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8635 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
8636 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
8639 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
8640 (TM_H): Add x86-tune.def.
8642 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8644 * config/aarch64/aarch64.c (generic_tunings):
8645 Use cortexa57_extra_costs.
8647 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8650 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
8651 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
8652 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
8655 2014-03-04 Heiher <r@hev.cc>
8657 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
8658 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
8660 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
8662 * config/i386/predicates.md (const2356_operand): Change to ...
8663 (const2367_operand): ... this.
8664 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
8666 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8667 (*avx512pf_scatterpf<mode>sf): Ditto.
8668 (avx512pf_scatterpf<mode>df): Ditto.
8669 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8670 (*avx512pf_scatterpf<mode>df): Ditto.
8671 * config/i386/i386.c (ix86_expand_builtin): Update
8672 incorrect hint operand error message.
8674 2014-03-04 Richard Biener <rguenther@suse.de>
8676 * lto-section-in.c (lto_get_section_data): Fix const cast.
8678 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8680 * tree-streamer.c (record_common_node): Assert we don't record
8681 nodes with type double.
8682 (preload_common_node): Skip type double, complex double and double
8683 pointer since it is now frontend dependent due to fshort-double option.
8685 2014-03-04 Richard Biener <rguenther@suse.de>
8688 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
8689 (lto_input_toplevel_asms): Likewise.
8690 * lto-section-in.c (lto_get_section_data): Instead do it here
8693 2014-03-04 Richard Biener <rguenther@suse.de>
8695 PR tree-optimization/60382
8696 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
8697 dead PHIs a reduction.
8699 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
8701 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
8703 (_mm_prefetch): Move out of GCC target("sse") pragma.
8704 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
8705 GCC target("prfchw") pragma.
8706 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
8708 * config/i386/i386.c (ix86_option_override_internal): Enable
8709 -mprfchw with -mprefetchwt1.
8711 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8713 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
8716 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8718 * opts.h (CL_PCH_IGNORE): Define.
8719 * targhooks.c (option_affects_pch_p):
8720 Return false for options that have CL_PCH_IGNORE set.
8721 * opt-functions.awk: Process PchIgnore.
8722 * doc/options.texi: Document PchIgnore.
8724 * config/arc/arc.opt (misize): Add PchIgnore property.
8726 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8728 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
8729 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
8730 constraint on constants to permit them being loaded into
8731 GENERAL_REGS or BASE_REGS.
8733 2014-03-03 Nick Clifton <nickc@redhat.com>
8735 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
8736 anti-cacnonical alternatives.
8737 (negandhi3_real): New pattern.
8738 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
8740 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8742 * config/avr/avr-mcus.def: Remove atxmega16x1.
8743 * config/avr/avr-tables.opt: Regenerate.
8744 * config/avr/t-multilib: Regenerate.
8745 * doc/avr-mmcu.texi: Regenerate.
8747 2014-03-03 Tobias Grosser <tobias@grosser.es>
8748 Mircea Namolaru <mircea.namolaru@inria.fr>
8750 PR tree-optimization/58028
8751 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
8754 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8756 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
8757 not handled by recognizers.
8759 2014-03-03 Jakub Jelinek <jakub@redhat.com>
8762 * function.c (expand_function_end): Don't emit
8763 clobber_return_register sequence if clobber_after is a BARRIER.
8764 * cfgexpand.c (construct_exit_block): Append instructions before
8765 return_label to prev_bb.
8767 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8769 * config/rs6000/constraints.md: Document reserved use of "wc".
8771 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8774 * ipa.c (function_and_variable_visibility): When dissolving comdat
8775 group, also set all symbols to local.
8777 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8782 2013-12-14 Jan Hubicka <jh@suse.cz>
8784 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
8786 2014-03-02 Jon Beniston <jon@beniston.com>
8792 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
8793 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
8794 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
8795 (simple_return, *simple_return): New patterns
8796 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
8797 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
8799 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
8801 * dwarf2out.c (gen_subprogram_die): Tidy.
8803 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
8806 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
8807 (*mov_t_msb_neg_negc): ... this new insn.
8809 2014-02-28 Jason Merrill <jason@redhat.com>
8812 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
8815 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
8818 * dwarf2out.c (decltype_auto_die): New static.
8819 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
8820 (gen_type_die_with_usage): Handle 'decltype(auto)'.
8821 (is_cxx_auto): Likewise.
8823 2014-02-28 Ian Bolton <ian.bolton@arm.com>
8825 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
8826 we are not using general regs only.
8828 2014-02-28 Richard Biener <rguenther@suse.de>
8831 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
8832 previous fix and only allow to remove trivial pre-headers
8833 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
8834 (remove_forwarder_block): Properly update the latch of a loop.
8836 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8839 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
8840 (cselib_preserved_hash_table): New.
8841 (preserve_constants_and_equivs): Move preserved vals to it.
8842 (cselib_find_slot): Look it up first.
8843 (cselib_init): Initialize it.
8844 (cselib_finish): Release it.
8845 (dump_cselib_table): Dump it.
8847 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8850 * cselib.c (remove_useless_values): Skip to avoid quadratic
8851 behavior if the condition moved from...
8852 (cselib_process_insn): ... here holds.
8854 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8857 * var-tracking.c (vt_initialize): Apply the same condition to
8858 preserve the CFA base value.
8860 2014-02-28 Joey Ye <joey.ye@arm.com>
8863 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
8864 if reload in progress or completed.
8866 2014-02-28 Tobias Burnus <burnus@net-b.de>
8869 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
8872 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
8874 * doc/tm.texi.in (Condition Code Status): Update documention for
8875 relative locations of cc0-setter and cc0-user.
8877 2014-02-27 Jeff Law <law@redhat.com>
8879 PR rtl-optimization/52714
8880 * combine.c (try_combine): When splitting an unrecognized PARALLEL
8881 into two independent simple sets, if I3 is a jump, ensure the
8882 pattern we place into I3 is a (set (pc) ...).
8884 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
8885 Jeff Law <law@redhat.com>
8887 PR rtl-optimization/49847
8888 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
8889 are in different blocks.
8890 * doc/tm.texi (Condition Code Status): Update documention for
8891 relative locations of cc0-setter and cc0-user.
8893 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
8896 * lra.c (lra_emit_add): Check SUBREG too.
8898 2014-02-27 Andreas Schwab <schwab@suse.de>
8900 * config/m68k/m68k.c (m68k_option_override): Disable
8901 -flive-range-shrinkage for classic m68k.
8902 (m68k_override_options_after_change): Likewise.
8904 2014-02-27 Marek Polacek <polacek@redhat.com>
8907 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
8908 -Wmaybe-uninitialized.
8910 2014-02-27 Alan Modra <amodra@gmail.com>
8913 * reload1.c (emit_input_reload_insns): When reload_override_in,
8914 set old to rl->in_reg when rl->in_reg is a subreg.
8916 2014-02-26 Richard Biener <rguenther@suse.de>
8919 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
8921 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8923 * common/config/i386/predicates.md (const1256_operand): Remove.
8924 (const2356_operand): New.
8925 (const_1_to_2_operand): Remove.
8926 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8927 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8928 (*avx512pf_gatherpf<mode>sf): Ditto.
8929 (avx512pf_gatherpf<mode>df): Ditto.
8930 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8931 (*avx512pf_gatherpf<mode>df): Ditto.
8932 (avx512pf_scatterpf<mode>sf): Ditto.
8933 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8934 (*avx512pf_scatterpf<mode>sf): Ditto.
8935 (avx512pf_scatterpf<mode>df): Ditto.
8936 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8937 (*avx512pf_scatterpf<mode>df): Ditto.
8938 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
8940 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
8942 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
8943 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
8944 (_mm512_mask_testn_epi64_mask): Move to ...
8945 * config/i386/avx512cdintrin.h: Here.
8946 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
8947 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
8948 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
8949 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
8950 TARGET_AVX512F from TARGET_AVX512CD.
8952 2014-02-26 Richard Biener <rguenther@suse.de>
8955 * ipa.c (walk_polymorphic_call_targets): Properly guard
8956 call to inline_update_overall_summary.
8958 2014-02-26 Bin Cheng <bin.cheng@arm.com>
8961 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
8962 and latches only if requested. Fix latch if it is removed.
8963 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
8964 LOOPS_HAVE_PREHEADERS.
8966 2014-02-25 Andrew Pinski <apinski@cavium.com>
8968 * builtins.c (expand_builtin_thread_pointer): Create a new target
8969 when the target is NULL.
8971 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
8973 PR rtl-optimization/60317
8974 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8975 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8976 * lra-assigns.c: Include params.h.
8977 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
8978 other reload pseudos considerations.
8980 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8982 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
8983 to use canonical form for nor<mode>3.
8985 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8988 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
8991 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8993 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
8994 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
8995 (ix86_handle_option): Handle OPT_mprefetchwt1.
8996 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
8997 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8999 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9000 OPTION_MASK_ISA_PREFETCHWT1.
9001 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
9002 (PTA_PREFETCHWT1): New.
9003 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
9004 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
9005 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
9006 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
9007 (*prefetch_avx512pf_<mode>_: Change into ...
9008 (*prefetch_prefetchwt1_<mode>: This.
9009 * config/i386/i386.opt (mprefetchwt1): New.
9010 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
9011 (_mm_prefetch): Handle intent to write.
9012 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
9014 2014-02-25 Richard Biener <rguenther@suse.de>
9017 * emit-rtl.c (mem_attrs_htab): Remove.
9018 (mem_attrs_htab_hash): Likewise.
9019 (mem_attrs_htab_eq): Likewise.
9020 (set_mem_attrs): Always allocate new mem-attrs when something changed.
9021 (init_emit_once): Do not allocate mem_attrs_htab.
9023 2014-02-25 Richard Biener <rguenther@suse.de>
9026 * lto-opts.c (lto_write_options): Output non-explicit conservative
9027 -fwrapv, -fno-trapv and -fno-strict-overflow.
9028 * lto-wrapper.c (merge_and_complain): Handle merging those options.
9029 (run_gcc): And pass them through.
9031 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9033 * sel-sched.c (calculate_new_fences): New parameter ptime.
9034 Calculate it as a maximum over all fence cycles.
9035 (sel_sched_region_2): Adjust the call to calculate_new_fences.
9036 Print the final schedule timing when sched_verbose.
9038 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9040 PR rtl-optimization/60292
9041 * sel-sched.c (fill_vec_av_set): Do not reset target availability
9042 bit fot the fence instruction.
9044 2014-02-24 Alangi Derick <alangiderick@gmail.com>
9046 * calls.h: Fix typo in comment.
9048 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
9050 * config/pa/pa.c (pa_output_move_double): Don't valididate when
9051 adjusting offsetable addresses.
9053 2014-02-24 Guozhi Wei <carrot@google.com>
9055 * sparseset.h (sparseset_pop): Fix the wrong index.
9057 2014-02-24 Walter Lee <walt@tilera.com>
9059 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
9060 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
9062 * common/config/tilegx/tilegx-common.c
9063 (TARGET_DEFAULT_TARGET_FLAGS): Define.
9064 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
9066 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
9067 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
9068 (tilegx_gimplify_va_arg_expr): Handle big endian.
9069 (tilegx_expand_unaligned_load): Ditto.
9070 (tilegx_expand_unaligned_store): Ditto.
9071 (TARGET_RETURN_IN_MSB): New.
9072 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
9073 (TARGET_ENDIAN_DEFAULT): New.
9074 (TARGET_BIG_ENDIAN): Handle big endian.
9075 (BYTES_BIG_ENDIAN): Ditto.
9076 (WORDS_BIG_ENDIAN): Ditto.
9077 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
9080 * config/tilegx/tilegx.md (extv): Handle big endian.
9082 (insn_st<n>): Ditto.
9083 (insn_st<n>_add<bitsuffix>): Ditto.
9084 (insn_stnt<n>): Ditto.
9085 (insn_stnt<n>_add<bitsuffix>):Ditto.
9086 (vec_interleave_highv8qi): Handle big endian.
9087 (vec_interleave_highv8qi_be): New.
9088 (vec_interleave_highv8qi_le): New.
9089 (insn_v1int_h): Handle big endian.
9090 (vec_interleave_lowv8qi): Handle big endian.
9091 (vec_interleave_lowv8qi_be): New.
9092 (vec_interleave_lowv8qi_le): New.
9093 (insn_v1int_l): Handle big endian.
9094 (vec_interleave_highv4hi): Handle big endian.
9095 (vec_interleave_highv4hi_be): New.
9096 (vec_interleave_highv4hi_le): New.
9097 (insn_v2int_h): Handle big endian.
9098 (vec_interleave_lowv4hi): Handle big endian.
9099 (vec_interleave_lowv4hi_be): New.
9100 (vec_interleave_lowv4hi_le): New.
9101 (insn_v2int_l): Handle big endian.
9102 (vec_interleave_highv2si): Handle big endian.
9103 (vec_interleave_highv2si_be): New.
9104 (vec_interleave_highv2si_le): New.
9105 (insn_v4int_h): Handle big endian.
9106 (vec_interleave_lowv2si): Handle big endian.
9107 (vec_interleave_lowv2si_be): New.
9108 (vec_interleave_lowv2si_le): New.
9109 (insn_v4int_l): Handle big endian.
9110 * config/tilegx/tilegx.opt (mbig-endian): New option.
9111 (mlittle-endian): New option.
9112 * doc/install.texi: Document tilegxbe-linux.
9113 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
9115 2014-02-24 Martin Jambor <mjambor@suse.cz>
9118 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
9119 there are no parameter descriptors.
9121 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
9123 PR rtl-optimization/60268
9124 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
9125 initialization to ...
9126 (sched_rgn_init): ... here.
9127 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
9129 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9131 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
9134 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9136 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
9139 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9141 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
9142 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
9144 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9146 * config/microblaze/predicates.md: Add cmp_op predicate.
9147 * config/microblaze/microblaze.md: Add branch_compare instruction
9148 which uses cmp_op predicate and emits cmp insn before branch.
9149 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
9150 to microblaze_expand_conditional_branch and consolidate logic.
9151 (microblaze_expand_conditional_branch): emit branch_compare
9152 insn instead of handling cmp op separate from branch insn.
9154 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9156 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
9159 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9161 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
9162 define_insn with define_expand and new define_insn
9163 *altivec_lve<VI_char>x_internal.
9164 (altivec_stve<VI_char>x): Replace define_insn with define_expand
9165 and new define_insn *altivec_stve<VI_char>x_internal.
9166 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
9168 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
9170 (altivec_expand_stvex_be): New function.
9172 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
9174 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
9175 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
9176 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
9177 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
9179 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
9182 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
9183 instead of emit_move_insn.
9185 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9187 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
9189 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
9190 gen_altivec_vsumsws.
9192 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9194 * config/rs6000/altivec.md (altivec_lvxl): Rename as
9195 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
9196 (altivec_lvxl_<mode>): New define_expand incorporating
9197 -maltivec=be semantics where needed.
9198 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
9199 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
9200 semantics where needed.
9201 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
9202 (altivec_stvx_<mode>): New define_expand incorporating
9203 -maltivec=be semantics where needed.
9204 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
9205 VM2 iterator instead of V4SI.
9206 (altivec_stvxl_<mode>): New define_expand incorporating
9207 -maltivec=be semantics where needed.
9208 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
9209 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
9210 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
9211 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
9212 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
9213 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
9214 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
9215 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
9216 ALTIVEC_BUILTIN_STVXL.
9217 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
9218 (altivec_expand_stvx_be): Likewise.
9219 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
9220 (altivec_expand_lvx_be): Likewise.
9221 (altivec_expand_stvx_be): Likewise.
9222 (altivec_expand_builtin): Add cases for
9223 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
9224 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
9225 (altivec_init_builtins): Add definitions for
9226 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
9227 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
9229 2014-02-21 Catherine Moore <clm@codesourcery.com>
9231 * doc/invoke.texi (mvirt, mno-virt): Document.
9232 * config/mips/mips.opt (mvirt): New option.
9233 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
9235 2014-02-21 Richard Biener <rguenther@suse.de>
9237 PR tree-optimization/60276
9238 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
9239 (STMT_VINFO_MIN_NEG_DIST): New macro.
9240 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
9241 STMT_VINFO_MIN_NEG_DIST.
9242 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
9243 made for negative dependence distances still hold.
9245 2014-02-21 Richard Biener <rguenther@suse.de>
9248 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
9249 DECL_INITIAL for globals not in the current function context.
9251 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9253 PR tree-optimization/56490
9254 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
9255 * tree-ssa-uninit.c: Include params.h.
9256 (compute_control_dep_chain): Add num_calls argument, return false
9257 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
9258 num_calls to recursive call.
9259 (find_predicates): Change dep_chain into normal array,
9260 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
9261 variable and adjust compute_control_dep_chain caller.
9262 (find_def_preds): Likewise.
9264 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
9266 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
9267 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
9269 2014-02-21 Nick Clifton <nickc@redhat.com>
9271 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
9272 (pushhi1): Likewise.
9273 (popqi1): Add mode to pre_dec.
9276 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9278 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
9279 mode for mask of V8SFmode permutation.
9281 2014-02-20 Richard Henderson <rth@redhat.com>
9284 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
9285 a new pseudo for OLDVAL.
9287 2014-02-20 Jakub Jelinek <jakub@redhat.com>
9290 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
9291 gen_reg_rtx if d->testing_p.
9292 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
9293 if d->testing_p and we will certainly return true.
9294 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
9297 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
9299 * emit-rtl.c (gen_reg_rtx): Assert that
9300 crtl->emit.regno_pointer_align_length is non-zero.
9302 2014-02-20 Richard Henderson <rth@redhat.com>
9305 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
9306 on failure the store back into EXPECT.
9308 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
9309 Sandra Loosemore <sandra@codesourcery.com>
9311 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
9312 * config/nios2/nios2.c (nios2_function_profiler): Add
9313 -fPIC (flag_pic == 2) support.
9314 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
9315 (nios2_large_offset_p): New function.
9316 (nios2_unspec_reloc_p): Move up position, update to use
9317 nios2_large_offset_p.
9318 (nios2_unspec_address): Remove function.
9319 (nios2_unspec_offset): New function.
9320 (nios2_large_got_address): New function.
9321 (nios2_got_address): Add large offset support.
9322 (nios2_legitimize_tls_address): Update usage of removed and new
9324 (nios2_symbol_binds_local_p): New function.
9325 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
9326 (nios2_legitimize_address): Update to use nios2_large_offset_p.
9327 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
9328 (nios2_print_operand): Merge H/L processing, add hiadj/lo
9329 processing for (const (unspec ...)).
9330 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
9332 2014-02-20 Richard Biener <rguenther@suse.de>
9334 * tree-cfg.c (replace_uses_by): Mark altered BBs before
9335 doing the substitution.
9336 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
9338 2014-02-20 Martin Jambor <mjambor@suse.cz>
9341 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
9342 info when checking whether lattices are bottom.
9344 2014-02-20 Richard Biener <rguenther@suse.de>
9347 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
9350 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
9353 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
9354 parameter specifying the scaling.
9355 (inline_call): Update.
9356 (want_inline_recursively): Guard division by zero.
9357 (recursive_inlining): Update.
9358 * ipa-inline.h (clone_inlined_nodes): Update.
9360 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9363 * config/i386/i386.c (classify_argument): Pass structures of size
9364 64 bytes or less in register.
9366 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9367 Kirill Yukhin <kirill.yukhin@intel.com>
9369 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
9370 (_mm_rcp28_round_ss): Ditto.
9371 (_mm_rsqrt28_round_sd): Ditto.
9372 (_mm_rsqrt28_round_ss): Ditto.
9373 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
9374 (_mm_rcp14_round_ss): Ditto.
9375 (_mm_rsqrt14_round_sd): Ditto.
9376 (_mm_rsqrt14_round_ss): Ditto.
9377 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
9378 the first input operand, get rid of match_dup.
9379 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
9381 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9383 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
9384 operand as the first input operand, set type attribute.
9385 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9387 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
9388 operand as the first input operand, set type attribute.
9390 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9392 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
9395 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
9398 * config/i386/i386.c (construct_container): Remove TFmode check
9399 for X86_64_INTEGER_CLASS.
9401 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
9404 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
9405 only when -Wpsabi is enabled.
9407 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
9410 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
9411 passing arrays in registers are the same as for structs, so remove the
9412 special case for them.
9414 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
9416 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
9417 destination type, extract only the valid bits if the source type is not
9418 integral and has a different mode.
9420 2014-02-19 Richard Biener <rguenther@suse.de>
9423 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
9426 2014-02-19 Richard Biener <rguenther@suse.de>
9429 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
9430 (ipa_modify_call_arguments): Emit an argument load explicitely and
9431 preserve virtual SSA form there and for the replacement call.
9432 Do not update SSA form nor free dominance info.
9434 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9436 * ipa.c (function_and_variable_visibility): Also clear WEAK
9437 flag when disolving COMDAT_GROUP.
9439 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9441 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
9442 * ipa-prop.c (ipa_set_jf_known_type): Return early when
9444 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
9445 do more sanity checks.
9446 (detect_type_change): Return true when giving up early.
9447 (compute_complex_assign_jump_func): Fix type parameter of
9448 ipa_set_ancestor_jf.
9449 (compute_complex_ancestor_jump_func): Likewise.
9450 (update_jump_functions_after_inlining): Fix updating of
9452 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
9454 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9456 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
9457 inline clones when edge disappears.
9459 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9462 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
9463 Split 64-bit moves into 2 patterns. Do not allow the use of
9464 direct move for TDmode in little endian, since the decimal value
9465 has little endian bytes within a word, but the 64-bit pieces are
9466 ordered in a big endian fashion, and normal subreg's of TDmode are
9468 (mov<mode>_64bit_dm): Likewise.
9469 (movtd_64bit_nodm): Likewise.
9471 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9473 PR tree-optimization/60174
9474 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
9475 statement of an SSA_NAME that occurs in an abnormal PHI node.
9477 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9480 * final.c (SEEN_BB): Remove.
9481 (SEEN_NOTE, SEEN_EMITTED): Renumber.
9482 (final_scan_insn): Don't force_source_line on second
9483 NOTE_INSN_BASIC_BLOCK.
9485 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
9488 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
9489 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
9490 (type_natural_mode): Warn ABI change when %zmm register is not
9491 available for AVX512F vector value passing.
9493 2014-02-18 Kai Tietz <ktietz@redhat.com>
9496 * config/i386/i386.c (ix86_expand_prologue): Use value in
9497 rax register as displacement when restoring %r10 or %rax.
9498 Fix wrong offset when restoring both registers.
9500 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9502 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
9503 assertion with conditional return.
9505 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9506 Uros Bizjak <ubizjak@gmail.com>
9509 * config/i386/driver-i386.c (host_detect_local_cpu): If
9510 YMM state is not saved by the OS, also clear has_f16c. Move
9511 CPUID 0x80000001 handling before YMM state saving checking.
9513 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
9515 PR rtl-optimization/58960
9516 * haifa-sched.c (alloc_global_sched_pressure_data): New,
9517 factored out from ...
9518 (sched_init): ... here.
9519 (free_global_sched_pressure_data): New, factored out from ...
9520 (sched_finish): ... here.
9521 * sched-int.h (free_global_sched_pressure_data): Declare.
9522 * sched-rgn.c (nr_regions_initial): New static global.
9523 (haifa_find_rgns): Initialize it.
9524 (schedule_region): Disable sched-pressure for the newly
9527 2014-02-17 Richard Biener <rguenther@suse.de>
9529 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9530 release SSA defs of pattern stmts.
9532 2014-02-17 Richard Biener <rguenther@suse.de>
9534 * tree-inline.c (expand_call_inline): Release the virtual
9535 operand defined by the call we are about to inline.
9537 2014-02-17 Richard Biener <rguenther@suse.de>
9539 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9541 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9542 Ilya Tocar <ilya.tocar@intel.com>
9544 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9545 arguments order in builtin.
9546 (_mm512_permutexvar_epi64): Ditto.
9547 (_mm512_mask_permutexvar_epi64): Ditto
9548 (_mm512_maskz_permutexvar_epi32): Ditto
9549 (_mm512_permutexvar_epi32): Ditto
9550 (_mm512_mask_permutexvar_epi32): Ditto
9552 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9554 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9555 (p8_vmrgow): Likewise.
9557 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9559 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9562 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9565 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9566 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9567 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9568 using direct move instructions on ISA 2.07. Also adjust
9569 instruction length for 64-bit.
9570 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9571 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9573 2014-02-15 Alan Modra <amodra@gmail.com>
9577 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9578 find_replacement on parts of insn rtl that might be reloaded.
9580 2014-02-15 Richard Biener <rguenther@suse.de>
9582 PR tree-optimization/60183
9583 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9584 (tree_ssa_phiprop): Calculate and free post-dominators.
9586 2014-02-14 Jeff Law <law@redhat.com>
9588 PR rtl-optimization/60131
9589 * ree.c (get_extended_src_reg): New function.
9590 (combine_reaching_defs): Use it rather than assuming location of REG.
9591 (find_and_remove_re): Verify first operand of extension is
9592 a REG before adding the insns to the copy list.
9594 2014-02-14 Roland McGrath <mcgrathr@google.com>
9596 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
9597 * configure: Regenerated.
9598 * config.in: Regenerated.
9599 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
9600 instead of ASM_SHORT.
9602 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
9603 Richard Earnshaw <rearnsha@arm.com>
9605 PR rtl-optimization/59535
9606 * lra-constraints.c (process_alt_operands): Encourage alternative
9607 when unassigned pseudo class is superset of the alternative class.
9608 (inherit_reload_reg): Don't inherit when optimizing for code size.
9609 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
9610 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
9611 modes not less than 4 for Thumb1.
9613 2014-02-14 Kyle McMartin <kyle@redhat.com>
9616 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
9618 2014-02-14 Richard Biener <rguenther@suse.de>
9620 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
9621 (get_frame_arg): Drop the assert with langhook types_compatible_p.
9622 Do not strip INDIRECT_REFs.
9624 2014-02-14 Richard Biener <rguenther@suse.de>
9627 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
9628 DECL_FUNCTION_SPECIFIC_TARGET.
9629 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
9630 * tree-streamer-out.c (pack_ts_target_option): Remove.
9631 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
9632 (write_ts_function_decl_tree_pointers): Do not stream
9633 DECL_FUNCTION_SPECIFIC_TARGET.
9634 * tree-streamer-in.c (unpack_ts_target_option): Remove.
9635 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
9636 (lto_input_ts_function_decl_tree_pointers): Do not stream
9637 DECL_FUNCTION_SPECIFIC_TARGET.
9639 2014-02-14 Jakub Jelinek <jakub@redhat.com>
9641 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
9642 (get_initial_def_for_induction, vectorizable_induction): Ignore
9643 debug stmts when looking for exit_phi.
9644 (vectorizable_live_operation): Fix up condition.
9646 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9648 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
9649 nreverse() because it changes the content of original tree list.
9651 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9653 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
9654 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
9656 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9658 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
9659 GNU coding standards.
9661 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9664 * dwarf2out.c (gen_subprogram_die): Don't call
9665 add_calling_convention_attribute if subr_die is old_die.
9667 2014-02-13 Sharad Singhai <singhai@google.com>
9669 * doc/optinfo.texi: Fix order of nodes.
9671 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
9673 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
9674 operands[2], not operands[3].
9676 2014-02-13 Richard Biener <rguenther@suse.de>
9679 * doc/install.texi (ISL): Update recommended version to 0.12.2,
9680 mention the possibility of an in-tree build.
9681 (CLooG): Update recommended version to 0.18.1, mention the
9682 possibility of an in-tree build and clarify that the ISL
9683 bundled with CLooG does not work.
9685 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9688 * expr.c (compress_float_constant): If x is a hard register,
9689 extend into a pseudo and then move to x.
9691 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
9693 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
9694 caused by bad second argument to warning_at() with -mhotpatch and
9695 nested functions (e.g. with gfortran).
9697 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
9699 * opts.c (option_name): Remove "enabled by default" rider.
9701 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
9703 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
9705 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
9706 Uros Bizjak <ubizjak@gmail.com>
9709 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
9710 * configure: Regenerated.
9712 2014-02-12 Richard Biener <rguenther@suse.de>
9714 * vec.c (vec_prefix::calculate_allocation): Move as
9715 inline variant to vec.h.
9716 (vec_prefix::calculate_allocation_1): New out-of-line version.
9717 * vec.h (vec_prefix::calculate_allocation_1): Declare.
9718 (vec_prefix::m_has_auto_buf): Rename to ...
9719 (vec_prefix::m_using_auto_storage): ... this.
9720 (vec_prefix::calculate_allocation): Inline the easy cases
9721 and dispatch to calculate_allocation_1 which doesn't need the
9723 (va_heap::reserve): Use gcc_checking_assert.
9724 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
9725 m_using_auto_storage.
9726 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
9728 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
9729 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
9730 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
9732 2014-02-12 Richard Biener <rguenther@suse.de>
9734 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
9735 when we found a dependence.
9737 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
9739 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
9741 (maybe_fold_stmt): ... into this new function.
9742 * omp-low.c (lower_omp): Update comment.
9744 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
9747 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
9750 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
9752 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
9753 identifiers in comments.
9754 (cortexa53_extra_costs): Likewise.
9755 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
9756 (cortexa7_extra_costs): Likewise.
9757 (cortexa12_extra_costs): Likewise.
9758 (cortexa15_extra_costs): Likewise.
9759 (v7m_extra_costs): Likewise.
9761 2014-02-12 Richard Biener <rguenther@suse.de>
9764 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
9765 of posix_memalign being successful.
9766 (lower_stmt): Restrict lowering of posix_memalign to when
9767 -ftree-bit-ccp is enabled.
9769 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9771 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
9773 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
9775 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
9777 PR rtl-optimization/60116
9778 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
9779 other_insn once the combination has been validated.
9781 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
9784 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
9786 * ipa-devirt.c: Include demangle.h
9787 (odr_violation_reported): New static variable.
9788 (add_type_duplicate): Update odr_violations.
9789 (maybe_record_node): Add completep parameter; update it.
9790 (record_target_from_binfo): Add COMPLETEP parameter;
9791 update it as needed.
9792 (possible_polymorphic_call_targets_1): Likewise.
9793 (struct polymorphic_call_target_d): Add nonconstruction_targets;
9794 rename FINAL to COMPLETE.
9795 (record_targets_from_bases): Sanity check we found the binfo;
9796 fix COMPLETEP updating.
9797 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
9798 parameter, fix computing of COMPLETEP.
9799 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
9800 at LTO time do demangling.
9801 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
9802 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
9804 (gimple_get_virt_method_for_binfo): Likewise.
9805 * gimple-fold.h (gimple_get_virt_method_for_binfo,
9806 gimple_get_virt_method_for_vtable): Update prototypes.
9808 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
9811 * genautomata.c (add_presence_absence): Fix typo with
9812 {final_}presence_list.
9814 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9817 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
9818 for VSX/Altivec vectors that land in GPR registers.
9820 2014-02-11 Richard Henderson <rth@redhat.com>
9821 Jakub Jelinek <jakub@redhat.com>
9824 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
9825 around drhs if type conversion to lacc->type is not useless.
9827 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9829 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
9831 (cortex-a57.cortex-a53): Likewise.
9832 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
9834 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9836 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
9839 2014-02-11 Renlin Li <Renlin.Li@arm.com>
9841 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
9842 add_options_for_arm_vfp3.
9844 2014-02-11 Jeff Law <law@redhat.com>
9847 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
9848 object with an undesirable mode.
9850 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9853 * config/i386/sol2-9.h: New file.
9854 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
9855 *-*-solaris2.9*): Use it.
9857 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9859 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
9860 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
9862 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9864 * config/microblaze/microblaze.c: Extend mcpu version format
9866 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
9868 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
9870 2014-02-10 Richard Henderson <rth@redhat.com>
9873 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
9874 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
9875 ms-abi vs -mno-accumulate-outgoing-args.
9876 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
9877 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
9880 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
9883 * cfgexpand.c (expand_asm_operands): Attach source location to
9884 ASM_INPUT rtx objects.
9885 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
9887 2014-02-10 Nick Clifton <nickc@redhat.com>
9889 * config/mn10300/mn10300.c (popcount): New function.
9890 (mn10300_expand_prologue): Include saved registers in stack usage
9893 2014-02-10 Jeff Law <law@redhat.com>
9896 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
9897 when changing the SET_DEST of a prior insn to avoid an input reload.
9899 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9901 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
9902 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
9903 -mcall-openbsd, or -mcall-linux.
9904 (CC1_ENDIAN_BIG_SPEC): Remove.
9905 (CC1_ENDIAN_LITTLE_SPEC): Remove.
9906 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9907 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
9908 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
9909 and %cc1_endian_default.
9910 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9912 2014-02-10 Richard Biener <rguenther@suse.de>
9914 PR tree-optimization/60115
9915 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
9916 MEM_REF handling. Properly verify that the accesses are not
9917 out of the objects bound.
9919 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9921 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
9924 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
9926 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
9927 proper constants and fix formatting.
9928 (possible_polymorphic_call_targets): Fix formatting.
9930 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
9931 Ilya Tocar <ilya.tocar@intel.com>
9933 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
9934 (_mm512_loadu_epi32): Renamed into...
9935 (_mm512_loadu_si512): This.
9936 (_mm512_storeu_epi32): Renamed into...
9937 (_mm512_storeu_si512): This.
9938 (_mm512_maskz_ceil_ps): Removed.
9939 (_mm512_maskz_ceil_pd): Ditto.
9940 (_mm512_maskz_floor_ps): Ditto.
9941 (_mm512_maskz_floor_pd): Ditto.
9942 (_mm512_floor_round_ps): Ditto.
9943 (_mm512_floor_round_pd): Ditto.
9944 (_mm512_ceil_round_ps): Ditto.
9945 (_mm512_ceil_round_pd): Ditto.
9946 (_mm512_mask_floor_round_ps): Ditto.
9947 (_mm512_mask_floor_round_pd): Ditto.
9948 (_mm512_mask_ceil_round_ps): Ditto.
9949 (_mm512_mask_ceil_round_pd): Ditto.
9950 (_mm512_maskz_floor_round_ps): Ditto.
9951 (_mm512_maskz_floor_round_pd): Ditto.
9952 (_mm512_maskz_ceil_round_ps): Ditto.
9953 (_mm512_maskz_ceil_round_pd): Ditto.
9954 (_mm512_expand_pd): Ditto.
9955 (_mm512_expand_ps): Ditto.
9956 * config/i386/i386.c (ix86_builtins): Remove
9957 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
9958 (bdesc_args): Ditto.
9959 * config/i386/predicates.md (const1256_operand): New.
9960 (const_1_to_2_operand): Ditto.
9961 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9962 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9963 (*avx512pf_gatherpf<mode>sf): Ditto.
9964 (avx512pf_gatherpf<mode>df): Ditto.
9965 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9966 (*avx512pf_gatherpf<mode>df): Ditto.
9967 (avx512pf_scatterpf<mode>sf): Ditto.
9968 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9969 (*avx512pf_scatterpf<mode>sf): Ditto.
9970 (avx512pf_scatterpf<mode>df): Ditto.
9971 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9972 (*avx512pf_scatterpf<mode>df): Ditto.
9973 (avx512f_expand<mode>): Removed.
9974 (<shift_insn><mode>3<mask_name>): Change predicate type.
9976 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9978 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
9979 not at the end of datarefs vector use ordered_remove to avoid
9980 reordering datarefs vector.
9983 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
9984 mark local addressable non-static vars as GOVD_PRIVATE
9985 instead of GOVD_LOCAL.
9986 * omp-low.c (lower_omp_for): Move gimple_bind_vars
9987 and BLOCK_VARS of gimple_bind_block to new_stmt rather
9991 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
9992 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
9993 assume_aligned or alloc_align attributes.
9994 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
9995 arguments. Handle also assume_aligned and alloc_align attributes.
9996 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
9997 calls to functions with assume_aligned or alloc_align attributes.
9998 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
10000 2014-02-08 Terry Guo <terry.guo@arm.com>
10002 * doc/invoke.texi: Document ARM -march=armv7e-m.
10004 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10006 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
10007 flag on __cilkrts_rethrow builtin.
10010 * ipa-cp.c (determine_versionability): Fail at -O0
10011 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
10012 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
10015 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10018 * tree-inline.c (copy_forbidden): Fail for
10019 __attribute__((optimize (0))) functions.
10021 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10023 * varpool.c: Include pointer-set.h.
10024 (varpool_remove_unreferenced_decls): Variables in other partitions
10025 will not be output; be however careful to not lose information
10026 about partitioning.
10028 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10030 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
10031 lookup in the vtable constructor.
10033 2014-02-07 Jeff Law <law@redhat.com>
10036 * config/m68k/m68k.md (ashldi_extsi): Turn into a
10037 define_insn_and_split.
10039 * ipa-inline.c (inline_small_functions): Fix typos.
10041 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10043 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
10044 (s390_can_use_return_insn): Declare.
10045 * config/s390/s390.h (EPILOGUE_USES): Define.
10046 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
10048 (s390_chunkify_start): Handle return JUMP_LABELs.
10049 (s390_early_mach): Emit a main_pool instruction on the entry edge.
10050 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
10051 (s390_can_use_return_insn): New functions.
10052 (s390_fix_long_loop_prediction): Handle conditional returns.
10053 (TARGET_SET_UP_BY_PROLOGUE): Define.
10054 * config/s390/s390.md (ANY_RETURN): New code iterator.
10055 (*creturn, *csimple_return, return, simple_return): New patterns.
10057 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10059 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
10060 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
10061 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
10062 REG_CFA_RESTORE list when deciding not to restore a register.
10064 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10066 * config/s390/s390.c: Include tree-pass.h and context.h.
10067 (s390_early_mach): New function, split out from...
10068 (s390_emit_prologue): ...here.
10069 (pass_data_s390_early_mach): New pass structure.
10070 (pass_s390_early_mach): New class.
10071 (s390_option_override): Create and register early_mach pass.
10072 Move to end of file.
10074 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10076 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
10077 to match for the exit block.
10079 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10081 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
10082 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
10083 Reject misaligned operands.
10085 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10087 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
10089 2014-02-07 Richard Biener <rguenther@suse.de>
10091 PR middle-end/60092
10092 * gimple-low.c (lower_builtin_posix_memalign): New function.
10093 (lower_stmt): Call it to lower posix_memalign in a way
10094 to make alignment info accessible.
10096 2014-02-07 Jakub Jelinek <jakub@redhat.com>
10099 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
10100 __builtin_setjmp_receiver.
10102 2014-02-07 Richard Biener <rguenther@suse.de>
10104 PR middle-end/60092
10105 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
10106 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
10107 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
10108 Handle BUILT_IN_POSIX_MEMALIGN.
10109 (find_func_clobbers): Likewise.
10110 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
10111 (call_may_clobber_ref_p_1): Likewise.
10113 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10116 * ipa-devirt.c (record_target_from_binfo): Remove overactive
10119 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10122 * lto-cgraph.c (lto_output_node): Use
10123 symtab_get_symbol_partitioning_class.
10124 (lto_output_varpool_node): likewise.
10125 (symtab_get_symbol_partitioning_class): Move here from
10126 lto/lto-partition.c
10127 * cgraph.h (symbol_partitioning_class): Likewise.
10128 (symtab_get_symbol_partitioning_class): Declare.
10130 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10132 * ggc.h (ggc_internal_cleared_alloc): New macro.
10133 * vec.h (vec_safe_copy): Handle memory stats.
10134 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
10135 * target-globals.c (save_target_globals): Likewise.
10137 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10140 * expr.c (emit_move_resolve_push): Export; be bit more selective
10141 on when to clear alias set.
10142 * expr.h (emit_move_resolve_push): Declare.
10143 * function.h (struct function): Add tail_call_marked.
10144 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
10145 * config/i386/i386-protos.h (ix86_expand_push): Remove.
10146 * config/i386/i386.md (TImode move expander): De not call
10148 (FP push expanders): Preserve memory attributes.
10149 * config/i386/sse.md (push<mode>1): Remove.
10150 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
10151 (ix86_expand_push): Remove.
10152 * config/i386/mmx.md (push<mode>1): Remove.
10154 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10156 PR rtl-optimization/60030
10157 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
10158 lopart with paradoxical subreg before shifting it up by hprec.
10160 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10162 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
10163 Remove extra newline at end of file.
10164 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
10165 (arm_issue_rate): Handle cortexa57.
10166 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
10167 (cortex-a57.cortex-a53): Likewise.
10169 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10172 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
10173 don't record in REG_FRAME_RELATED_EXPR registers not set in that
10175 (arm_expand_prologue): Adjust all callers.
10176 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
10177 info, registers also at the lowest numbered registers side. Use
10178 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
10182 * var-tracking.c (adjust_mems): Before adding a SET to
10183 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
10185 2014-02-06 Alan Modra <amodra@gmail.com>
10188 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
10189 change SDmode to DDmode when lra_in_progress.
10191 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10193 PR middle-end/59150
10194 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
10195 free_data_ref on the dr first, and before goto again also set dr
10196 to the next dr. For simd_lane_access, free old datarefs[i] before
10197 overwriting it. For get_vectype_for_scalar_type failure, don't
10198 free_data_ref if simd_lane_access.
10200 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
10203 * tree.h (opts_for_fn): New inline function.
10204 (opt_for_fn): Define.
10205 * config/i386/i386.c (ix86_function_regparm): Use
10206 opt_for_fn (decl, optimize) instead of optimize.
10208 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
10210 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
10211 for SYMBOL_REF in large memory model.
10213 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10215 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
10216 and crypto support.
10217 (cortex-a57): Likewise.
10218 (cortex-a57.cortex-a53): Likewise.
10220 2014-02-06 Yury Gribov <y.gribov@samsung.com>
10221 Kugan Vivekanandarajah <kuganv@linaro.org>
10223 * config/arm/arm.c (arm_vector_alignment_reachable): Check
10225 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
10227 2014-02-06 Richard Biener <rguenther@suse.de>
10229 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
10230 set_loop_copy and initialize_original_copy_tables.
10232 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
10234 * config/aarch64/aarch64-simd.md
10235 (aarch64_ashr_simddi): Change QI to SI.
10237 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10238 Jakub Jelinek <jakub@redhat.com>
10240 PR middle-end/60013
10241 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
10244 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10246 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
10247 CODE_FOR_altivec_vpku[hw]um to
10248 CODE_FOR_altivec_vpku[hw]um_direct.
10249 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
10250 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
10251 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
10252 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
10254 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10256 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
10257 generation for -maltivec=be.
10258 (altivec_vsumsws): Simplify redundant test.
10260 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10262 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
10263 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
10264 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
10265 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
10266 gen_altivec_vpkuwum.
10267 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
10269 (altivec_vpks<VI_char>ss): Likewise.
10270 (altivec_vpks<VI_char>us): Likewise.
10271 (altivec_vpku<VI_char>us): Likewise.
10272 (altivec_vpku<VI_char>um): Likewise.
10273 (altivec_vpku<VI_char>um_direct): New (copy of
10274 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
10276 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
10277 target is little endian and -maltivec=be is not specified.
10278 (*altivec_vupkhs<VU_char>_direct): New (copy of
10279 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
10280 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
10281 target is little endian and -maltivec=be is not specified.
10282 (*altivec_vupkls<VU_char>_direct): New (copy of
10283 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
10284 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
10285 little endian and -maltivec=be is not specified.
10286 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
10287 little endian and -maltivec=be is not specified.
10289 2014-02-05 Richard Henderson <rth@redhat.com>
10292 * combine-stack-adj.c: Revert r206943.
10293 * sched-int.h (struct deps_desc): Add last_args_size.
10294 * sched-deps.c (init_deps): Initialize it.
10295 (sched_analyze_insn): Add OUTPUT dependencies between insns that
10296 contain REG_ARGS_SIZE notes.
10298 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10300 * lto-cgraph.c (asm_nodes_output): Make global.
10301 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
10302 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
10303 (driver_handle_option): Handle OPT_fwpa.
10305 2014-02-05 Jakub Jelinek <jakub@redhat.com>
10308 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
10309 a comment typo and formatting issue. If odr_hash hasn't been
10310 created, return vNULL and set *completep to false.
10312 PR middle-end/57499
10313 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
10314 bb with no successors.
10316 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
10319 * doc/invoke.texi (-march): Clarify documentation for ARM.
10320 (-mtune): Likewise.
10323 2014-02-05 Richard Biener <rguenther@suse.de>
10325 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
10326 when not vectorizing because of too many alias checks.
10327 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
10328 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
10330 2014-02-05 Nick Clifton <nickc@redhat.com>
10332 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
10333 accept extended registers in any mode when compiling for the MN10300.
10335 2014-02-05 Yury Gribov <y.gribov@samsung.com>
10337 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
10338 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
10339 sanitization attributes.
10340 (can_inline_edge_p): Likewise.
10341 (sanitize_attrs_match_for_inline_p): New function.
10343 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10345 * ipa-prop.c (detect_type_change): Shor circuit testing of
10346 type changes on THIS pointer.
10348 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
10351 * config/pa/pa.c (legitimize_tls_address): Return original address
10352 if not passed a SYMBOL_REF rtx.
10353 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
10355 (pa_emit_move_sequence): Simplify TLS source operands.
10356 (pa_legitimate_constant_p): Reject all TLS constants.
10357 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
10358 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
10360 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10362 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
10363 groups when we know they are controlled by LTO.
10364 * varasm.c (default_binds_local_p_1): If object is in other partition,
10365 it will be resolved locally.
10367 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10369 * config/host-linux.c (linux_gt_pch_use_address): Don't
10370 use SSIZE_MAX because it is not always defined.
10372 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
10375 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
10376 threshold for pseudo splitting.
10377 (update_ebb_live_info): Process call argument hard registers and
10378 hard registers from insn definition too.
10379 (max_small_class_regs_num): New constant.
10380 (inherit_in_ebb): Update live hard regs through EBBs. Update
10381 reloads_num only for small register classes. Don't split for
10384 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
10387 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
10390 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10392 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
10393 visibility is safe.
10395 2014-02-04 Marek Polacek <polacek@redhat.com>
10397 * gdbinit.in (pel): Define.
10399 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10401 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
10404 2014-02-04 Richard Biener <rguenther@suse.de>
10407 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
10408 in function context local.
10409 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
10410 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
10411 similar to LTO_imported_decl_ref.
10413 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10415 PR tree-optimization/60002
10416 * cgraphclones.c (build_function_decl_skip_args): Clear
10417 DECL_LANG_SPECIFIC.
10419 PR tree-optimization/60023
10420 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
10421 false to gsi_replace.
10422 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
10423 has been in some EH region and vec_stmt could throw, add
10424 vec_stmt into the same EH region.
10425 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
10426 has no lhs, ignore it.
10427 * internal-fn.c (expand_MASK_LOAD): Likewise.
10430 * tree-inline.c (copy_forbidden): Fail for
10431 __attribute__((optimize (0))) functions.
10434 * omp-low.c (simd_clone_struct_copy): If from->inbranch
10435 is set, copy one less argument.
10436 (expand_simd_clones): Don't subtract clone_info->inbranch
10437 from simd_clone_struct_alloc argument.
10439 PR rtl-optimization/57915
10440 * recog.c (simplify_while_replacing): If all unary/binary/relational
10441 operation arguments are constant, attempt to simplify those.
10443 PR middle-end/59261
10444 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
10445 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
10447 2014-02-04 Richard Biener <rguenther@suse.de>
10449 PR tree-optimization/60012
10450 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
10451 TBAA disambiguation to all DDRs.
10453 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10456 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
10457 (LINK_SPEC): Use it for -shared, -shared-libgcc.
10459 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10462 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
10464 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10466 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
10467 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
10469 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10472 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
10473 to figure out targets of polymorphic calls with known decl.
10474 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10475 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
10476 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
10477 (get_polymorphic_call_info): ... here.
10478 (get_polymorphic_call_info_from_invariant): New function.
10480 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10482 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
10483 lookup via vtable pointer; check for type consistency
10484 and turn inconsitent facts into UNREACHABLE.
10485 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10486 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
10487 type inconsistent querries; return UNREACHABLE instead.
10489 2014-02-03 Richard Henderson <rth@twiddle.net>
10492 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
10493 already processed this node.
10494 (normalize_one_pred_1): Pass along mark_set.
10495 (normalize_one_pred): Create and destroy a pointer_set_t.
10496 (normalize_one_pred_chain): Likewise.
10498 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
10500 PR gcov-profile/58602
10501 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
10503 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10506 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
10507 -fno-devirtualize; try to devirtualize by the knowledge of
10508 virtual table pointer given by aggregate propagation.
10509 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10510 (ipa_print_node_jump_functions): Dump also offset that
10511 is relevant for polymorphic calls.
10512 (determine_known_aggregate_parts): Add arg_type parameter; use it
10513 instead of determining the type from pointer type.
10514 (ipa_compute_jump_functions_for_edge): Update call of
10515 determine_known_aggregate_parts.
10516 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
10517 (gimple_get_virt_method_for_binfo): ... here; simplify using
10518 vtable_pointer_value_to_vtable.
10519 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
10520 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
10521 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
10522 (vtable_pointer_value_to_vtable): Break out from ...; handle also
10524 (vtable_pointer_value_to_binfo): ... here.
10525 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
10527 2014-02-03 Teresa Johnson <tejohnson@google.com>
10529 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10530 redef of outer loop index variable.
10532 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10536 * doc/extend.texi (Function Attributes): Typo.
10538 2014-02-03 Cong Hou <congh@google.com>
10540 PR tree-optimization/60000
10541 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10542 if the vectorized statement is a store. A store statement can only
10543 appear at the end of pattern statements.
10545 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10547 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10548 (ix86_option_override_internal): Default long double to 64-bit for
10549 32-bit Bionic and to 128-bit for 64-bit Bionic.
10551 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10552 TARGET_LONG_DOUBLE_128 is true.
10553 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10555 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10556 (mlong-double-64): Negate -mlong-double-128.
10557 (mlong-double-128): New option.
10559 * config/i386/i386-c.c (ix86_target_macros): Define
10560 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10562 * doc/invoke.texi: Document -mlong-double-128.
10564 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10566 PR rtl-optimization/60024
10567 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10569 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10571 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10573 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10575 PR rtl-optimization/57662
10576 * sel-sched.c (code_motion_path_driver): Do not mark already not
10577 existing blocks in the visiting bitmap.
10579 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10581 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10582 on the insn being emitted.
10584 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10585 Will Deacon <will.deacon@arm.com>
10587 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
10589 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10591 * config/arm/arm-tables.opt: Regenerate.
10593 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10595 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
10596 for vector types other than V16QImode.
10597 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
10598 define_expand, and call altivec_expand_vec_perm_le when producing
10599 code with little endian element order.
10600 (*altivec_vperm_<mode>_internal): New insn having previous
10601 behavior of altivec_vperm_<mode>.
10602 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
10603 altivec_expand_vec_perm_le when producing code with little endian
10605 (*altivec_vperm_<mode>_uns_internal): New insn having previous
10606 behavior of altivec_vperm_<mode>_uns.
10608 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10610 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
10611 (altivec_vsumsws): Add handling for -maltivec=be with a little
10613 (altivec_vsumsws_direct): New.
10614 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
10615 gen_altivec_vsumsws.
10617 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10619 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
10620 vtable_pointer_value_to_binfo): New functions.
10621 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
10622 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
10624 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
10626 * config/nios2/nios2.md (load_got_register): Initialize GOT
10627 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
10628 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
10630 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10632 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
10633 preserverd by passthrough, do not propagate the type.
10635 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10637 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
10638 (mips_atomic_assign_expand_fenv): New function.
10639 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
10641 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10643 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
10644 (__builtin_mips_set_fcsr): Likewise.
10645 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
10646 MIPS_USI_FTYPE_VOID.
10647 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
10648 (mips16_expand_set_fcsr): Likewise.
10649 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
10650 (mips16_set_fcsr_stub): Likewise.
10651 (mips16_get_fcsr_one_only_stub): New class.
10652 (mips16_set_fcsr_one_only_stub): Likewise.
10653 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
10654 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
10655 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
10656 (hard_float): New availability predicate.
10657 (mips_builtins): Add get_fcsr and set_fcsr.
10658 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
10659 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
10660 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
10661 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
10662 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
10665 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10667 * config/mips/mips.c (mips_one_only_stub): New class.
10668 (mips_need_mips16_rdhwr_p): Replace with...
10669 (mips16_rdhwr_stub): ...this new variable.
10670 (mips16_stub_call_address): New function.
10671 (mips16_rdhwr_one_only_stub): New class.
10672 (mips_expand_thread_pointer): Use mips16_stub_call_address.
10673 (mips_output_mips16_rdhwr): Delete.
10674 (mips_finish_stub): New function.
10675 (mips_code_end): Use it to handle rdhwr stubs.
10677 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
10680 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
10681 when calculating size of integer atomic types.
10683 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
10685 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
10687 2014-02-01 Jakub Jelinek <jakub@redhat.com>
10689 PR tree-optimization/60003
10690 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
10691 * profile.c (branch_prob): Use gimple_call_builtin_p
10692 to check for BUILT_IN_SETJMP_RECEIVER.
10693 * tree-inline.c (copy_bb): Call notice_special_calls.
10695 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
10698 * lra-constraints.c (process_alt_operands): Update reload_sum only
10701 2014-01-31 Richard Henderson <rth@redhat.com>
10703 PR middle-end/60004
10704 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
10705 until after else_eh is processed.
10707 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10709 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
10710 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
10711 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
10712 in smmintrin.h, remove them.
10713 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
10714 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
10715 * config/i386/i386.md (ROUND_SAE): Fix value.
10716 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
10717 (const48_operand): New.
10718 * config/i386/subst.md (round), (round_expand): Use
10719 const_4_or_8_to_11_operand.
10720 (round_saeonly), (round_saeonly_expand): Use const48_operand.
10722 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10724 * config/i386/constraints.md (Yk): Swap meaning with k.
10725 * config/i386/i386.md (movhi_internal): Change Yk to k.
10726 (movqi_internal): Ditto.
10727 (*k<logic><mode>): Ditto.
10730 (kandn<mode>): Ditto.
10731 (*<code>hi_1): Ditto.
10732 (*<code>qi_1): Ditto.
10733 (kxnor<mode>): Ditto.
10734 (kortestzhi): Ditto.
10735 (kortestchi): Ditto.
10737 (*one_cmplhi2_1): Ditto.
10738 (*one_cmplqi2_1): Ditto.
10739 * config/i386/sse.md (): Change k to Yk.
10740 (avx512f_load<mode>_mask): Ditto.
10741 (avx512f_blendm<mode>): Ditto.
10742 (avx512f_store<mode>_mask): Ditto.
10743 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
10744 (avx512f_storedqu<mode>_mask): Ditto.
10745 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
10747 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
10748 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
10749 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
10750 (avx512f_maskcmp<mode>3): Ditto.
10751 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
10752 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
10753 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
10754 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
10755 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
10756 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
10757 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
10758 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
10759 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
10760 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
10761 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
10762 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
10763 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
10764 (vec_extract_lo_<mode>_maskm): Ditto.
10765 (vec_extract_hi_<mode>_maskm): Ditto.
10766 (avx512f_vternlog<mode>_mask): Ditto.
10767 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
10768 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
10769 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
10770 (avx512f_<code>v8div16qi2_mask): Ditto.
10771 (avx512f_<code>v8div16qi2_mask_store): Ditto.
10772 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
10773 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
10774 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
10775 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
10776 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10777 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10778 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10779 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10780 (avx512cd_maskb_vec_dupv8di): Ditto.
10781 (avx512cd_maskw_vec_dupv16si): Ditto.
10782 (avx512f_vpermi2var<mode>3_maskz): Ditto.
10783 (avx512f_vpermi2var<mode>3_mask): Ditto.
10784 (avx512f_vpermi2var<mode>3_mask): Ditto.
10785 (avx512f_vpermt2var<mode>3_maskz): Ditto.
10786 (*avx512f_gathersi<mode>): Ditto.
10787 (*avx512f_gathersi<mode>_2): Ditto.
10788 (*avx512f_gatherdi<mode>): Ditto.
10789 (*avx512f_gatherdi<mode>_2): Ditto.
10790 (*avx512f_scattersi<mode>): Ditto.
10791 (*avx512f_scatterdi<mode>): Ditto.
10792 (avx512f_compress<mode>_mask): Ditto.
10793 (avx512f_compressstore<mode>_mask): Ditto.
10794 (avx512f_expand<mode>_mask): Ditto.
10795 * config/i386/subst.md (mask): Change k to Yk.
10796 (mask_scalar_merge): Ditto.
10799 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
10801 * doc/extend.texi (Vector Extensions): Document ?: in C++.
10803 2014-01-31 Richard Biener <rguenther@suse.de>
10805 PR middle-end/59990
10806 * builtins.c (fold_builtin_memory_op): Make sure to not
10807 use a floating-point mode or a boolean or enumeral type for
10808 the copy operation.
10810 2014-01-30 DJ Delorie <dj@redhat.com>
10812 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
10813 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
10814 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
10815 whenever main() has an epilogue.
10817 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10819 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
10820 unused variable "field".
10821 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
10822 (vsx_mergeh_<mode>): Likewise.
10823 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
10824 (altivec_vmrghh): Likewise.
10825 (altivec_vmrghw): Likewise.
10826 (altivec_vmrglb): Likewise.
10827 (altivec_vmrglh): Likewise.
10828 (altivec_vmrglw): Likewise.
10829 (altivec_vspltb): Add missing uses.
10830 (altivec_vsplth): Likewise.
10831 (altivec_vspltw): Likewise.
10832 (altivec_vspltsf): Likewise.
10834 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10837 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
10838 frame related instructions.
10840 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
10842 PR rtl-optimization/59959
10843 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
10844 any reload of register whose subreg is invalid.
10846 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10848 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
10849 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
10850 Add missing return type - void.
10852 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10854 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
10855 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
10856 remove element index adjustment for endian (now handled in vsx.md
10858 (altivec_expand_vec_perm_const): Use
10859 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
10860 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
10861 (vsx_xxspltw_<mode>): Adjust element index for little endian.
10862 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
10863 define_expand and a new define_insn *altivec_vspltb_internal;
10864 adjust for -maltivec=be on a little endian target.
10865 (altivec_vspltb_direct): New.
10866 (altivec_vsplth): Divide into a define_expand and a new
10867 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
10868 little endian target.
10869 (altivec_vsplth_direct): New.
10870 (altivec_vspltw): Divide into a define_expand and a new
10871 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
10872 little endian target.
10873 (altivec_vspltw_direct): New.
10874 (altivec_vspltsf): Divide into a define_expand and a new
10875 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
10876 a little endian target.
10878 2014-01-30 Richard Biener <rguenther@suse.de>
10880 PR tree-optimization/59993
10881 * tree-ssa-forwprop.c (associate_pointerplus): Check we
10882 can propagate form the earlier stmt and avoid the transform
10883 when the intermediate result is needed.
10885 2014-01-30 Alangi Derick <alangiderick@gmail.com>
10887 * README.Portability: Fix typo.
10889 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
10891 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
10892 comparison_operator with ordered_comparison_operator.
10894 2014-01-30 Nick Clifton <nickc@redhat.com>
10896 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
10897 Rename to mn10300_store_multiple_regs.
10898 * config/mn10300/mn10300.c: Likewise.
10899 * config/mn10300/mn10300.md (store_movm): Fix typo: call
10900 store_multiple_regs.
10901 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
10902 Call mn10300_store_multiple_regs.
10904 2014-01-30 Nick Clifton <nickc@redhat.com>
10905 DJ Delorie <dj@redhat.com>
10907 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
10908 %fp 2 to keep registers after it properly word-aligned.
10909 (rl78_alloc_physical_registers_umul): Handle the case where both
10910 input operands are the same.
10912 2014-01-30 Richard Biener <rguenther@suse.de>
10914 PR tree-optimization/59903
10915 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
10918 2014-01-30 Jason Merrill <jason@redhat.com>
10921 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
10924 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
10926 2014-01-30 Richard Biener <rguenther@suse.de>
10928 PR tree-optimization/59951
10929 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
10931 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
10934 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
10935 SFmode to DFmode case.
10937 2014-01-29 DJ Delorie <dj@redhat.com>
10939 * config/msp430/msp430.opt (-minrt): New.
10940 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
10942 (ENDFILE_SPEC): Likewise.
10944 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
10946 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
10947 (estimate_function_body_sizes): Use it.
10949 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
10952 * dwarf2out.c (is_cxx_auto): New.
10953 (is_base_type): Use it.
10954 (gen_type_die_with_usage): Likewise.
10956 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10958 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
10959 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
10960 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
10961 -maltivec=be with LE targets.
10962 (vsx_mergeh_<mode>): Likewise.
10963 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
10964 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
10965 (altivec_vmrghb): Replace with define_expand and new
10966 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
10967 (altivec_vmrghb_direct): New define_insn.
10968 (altivec_vmrghh): Replace with define_expand and new
10969 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
10970 (altivec_vmrghh_direct): New define_insn.
10971 (altivec_vmrghw): Replace with define_expand and new
10972 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
10973 (altivec_vmrghw_direct): New define_insn.
10974 (*altivec_vmrghsf): Adjust for endianness.
10975 (altivec_vmrglb): Replace with define_expand and new
10976 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
10977 (altivec_vmrglb_direct): New define_insn.
10978 (altivec_vmrglh): Replace with define_expand and new
10979 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
10980 (altivec_vmrglh_direct): New define_insn.
10981 (altivec_vmrglw): Replace with define_expand and new
10982 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
10983 (altivec_vmrglw_direct): New define_insn.
10984 (*altivec_vmrglsf): Adjust for endianness.
10985 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10986 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10987 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10988 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10989 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10990 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10991 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10992 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10994 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
10996 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
10997 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
11000 2014-01-29 Richard Biener <rguenther@suse.de>
11002 PR tree-optimization/58742
11003 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
11004 associate_pointerplus_align.
11005 (associate_pointerplus_diff): New function.
11006 (associate_pointerplus): Likewise. Call associate_pointerplus_align
11007 and associate_pointerplus_diff.
11009 2014-01-29 Richard Biener <rguenther@suse.de>
11011 * lto-streamer.h (LTO_major_version): Bump to 3.
11012 (LTO_minor_version): Reset to 0.
11014 2014-01-29 Renlin Li <Renlin.Li@arm.com>
11016 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
11017 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
11018 (arm_file_start): Generate correct asm header for armv7ve.
11019 * config/arm/bpabi.h: Add multilib support for armv7ve.
11020 * config/arm/driver-arm.c: Change the architectures of cortex-a7
11021 and cortex-a15 to armv7ve.
11022 * config/arm/t-aprofile: Add multilib support for armv7ve.
11023 * doc/invoke.texi: Document -march=armv7ve.
11025 2014-01-29 Richard Biener <rguenther@suse.de>
11027 PR tree-optimization/58742
11028 * tree-ssa-forwprop.c (associate_plusminus): Return true
11029 if we changed sth, defer EH cleanup to ...
11030 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
11031 (simplify_mult): New function.
11033 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11035 PR middle-end/59917
11036 PR tree-optimization/59920
11037 * tree.c (build_common_builtin_nodes): Remove
11038 __builtin_setjmp_dispatcher initialization.
11039 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
11040 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
11041 instead of gsi_after_labels + manually skipping debug stmts.
11042 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
11043 ignore bbs with IFN_ABNORMAL_DISPATCHER.
11044 * tree-inline.c (copy_edges_for_bb): Remove
11045 can_make_abnormal_goto argument, instead add abnormal_goto_dest
11046 argument. Ignore computed_goto_p stmts. Don't call
11047 make_abnormal_goto_edges. If a call might need abnormal edges
11048 for non-local gotos, see if it already has an edge to
11049 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
11050 with true argument, don't do anything then, otherwise add
11051 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
11052 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
11054 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
11055 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
11056 (lower_stmt): Don't set data->calls_builtin_setjmp.
11057 (lower_builtin_setjmp): Adjust comment.
11058 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
11059 * tree-cfg.c (found_computed_goto): Remove.
11060 (factor_computed_gotos): Remove.
11061 (make_goto_expr_edges): Return bool, true for computed gotos.
11062 Don't call make_abnormal_goto_edges.
11063 (build_gimple_cfg): Don't set found_computed_goto, don't call
11064 factor_computed_gotos.
11065 (computed_goto_p): No longer static.
11066 (make_blocks): Don't set found_computed_goto.
11067 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
11068 (make_edges): If make_goto_expr_edges returns true, push bb
11069 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
11070 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
11071 vector. Record mapping between bbs and OpenMP regions if there
11072 are any, adjust make_gimple_omp_edges caller. Call
11073 handle_abnormal_edges.
11074 (make_abnormal_goto_edges): Remove.
11075 * tree-cfg.h (make_abnormal_goto_edges): Remove.
11076 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
11077 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
11078 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
11079 * internal-fn.def (ABNORMAL_DISPATCHER): New.
11080 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
11081 filling *region also set *region_idx to (*region)->entry->index.
11084 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
11085 For REGs set ORIGINAL_REGNO.
11087 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
11089 * doc/md.texi: Mention that a target shouldn't implement
11090 vec_widen_(s|u)mul_even/odd pair if it is less efficient
11093 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11095 PR tree-optimization/59594
11096 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
11097 a copy of the datarefs vector rather than the vector itself.
11099 2014-01-28 Jason Merrill <jason@redhat.com>
11102 * dwarf2out.c (auto_die): New static.
11103 (gen_type_die_with_usage): Handle C++1y 'auto'.
11104 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
11107 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
11110 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
11111 (SPEC_X32): Likewise.
11112 (SPEC_64): Likewise.
11113 * config/i386/i386.c (ix86_option_override_internal): Turn off
11114 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
11116 (x86_file_start): Output .code16gcc for TARGET_16BIT.
11117 * config/i386/i386.h (TARGET_16BIT): New macro.
11118 (TARGET_16BIT_P): Likewise.
11119 * config/i386/i386.opt: Add m16.
11120 * doc/invoke.texi: Document -m16.
11122 2014-01-28 Jakub Jelinek <jakub@redhat.com>
11124 PR preprocessor/59935
11125 * input.c (location_get_source_line): Bail out on when line number
11126 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
11128 2014-01-28 Richard Biener <rguenther@suse.de>
11130 PR tree-optimization/58742
11131 * tree-ssa-forwprop.c (associate_plusminus): Handle
11132 pointer subtraction of the form (T)(P + A) - (T)P.
11134 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11136 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
11139 2014-01-28 Richard Biener <rguenther@suse.de>
11142 2014-01-28 Richard Biener <rguenther@suse.de>
11144 PR rtl-optimization/45364
11145 PR rtl-optimization/59890
11146 * var-tracking.c (local_get_addr_clear_given_value): Handle
11147 already cleared slot.
11148 (val_reset): Handle not allocated local_get_addr_cache.
11149 (vt_find_locations): Use post-order on the inverted CFG.
11151 2014-01-28 Richard Biener <rguenther@suse.de>
11153 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
11155 2014-01-28 Richard Biener <rguenther@suse.de>
11157 PR rtl-optimization/45364
11158 PR rtl-optimization/59890
11159 * var-tracking.c (local_get_addr_clear_given_value): Handle
11160 already cleared slot.
11161 (val_reset): Handle not allocated local_get_addr_cache.
11162 (vt_find_locations): Use post-order on the inverted CFG.
11164 2014-01-28 Alan Modra <amodra@gmail.com>
11166 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
11167 * configure.ac <recursive call for build != host>: Define
11168 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
11169 and LD_FOR_BUILD too.
11170 * configure: Regenerate.
11172 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
11174 * config/i386/i386.c (get_builtin_code_for_version): Separate
11175 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
11176 Broadwell from Haswell.
11178 2014-01-27 Steve Ellcey <sellcey@mips.com>
11180 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
11181 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
11182 * config/mips/mips.c (mips_option_override): Change setting
11184 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
11185 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
11186 Change from Mask to Var.
11188 2014-01-27 Jeff Law <law@redhat.com>
11190 * ipa-inline.c (inline_small_functions): Fix typo.
11192 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
11194 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
11195 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
11196 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
11197 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
11198 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
11199 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
11200 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
11201 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
11202 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
11203 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
11204 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
11205 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
11206 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
11207 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
11208 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
11209 (_mm512_storeu_epi64): Ditto.
11210 (_mm512_cmpge_epi32_mask): Ditto.
11211 (_mm512_cmpge_epu32_mask): Ditto.
11212 (_mm512_cmpge_epi64_mask): Ditto.
11213 (_mm512_cmpge_epu64_mask): Ditto.
11214 (_mm512_cmple_epi32_mask): Ditto.
11215 (_mm512_cmple_epu32_mask): Ditto.
11216 (_mm512_cmple_epi64_mask): Ditto.
11217 (_mm512_cmple_epu64_mask): Ditto.
11218 (_mm512_cmplt_epi32_mask): Ditto.
11219 (_mm512_cmplt_epu32_mask): Ditto.
11220 (_mm512_cmplt_epi64_mask): Ditto.
11221 (_mm512_cmplt_epu64_mask): Ditto.
11222 (_mm512_cmpneq_epi32_mask): Ditto.
11223 (_mm512_cmpneq_epu32_mask): Ditto.
11224 (_mm512_cmpneq_epi64_mask): Ditto.
11225 (_mm512_cmpneq_epu64_mask): Ditto.
11226 (_mm512_expand_pd): Ditto.
11227 (_mm512_expand_ps): Ditto.
11228 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
11229 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
11230 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
11231 * config/i386/i386.c (ix86_builtins): Add
11232 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
11233 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
11234 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
11235 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
11236 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
11237 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
11238 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
11239 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
11240 IX86_BUILTIN_PMOVUSQW512_MEM.
11241 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
11242 __builtin_ia32_pmovsqd512mem_mask,
11243 __builtin_ia32_pmovqd512mem_mask,
11244 __builtin_ia32_pmovusqw512mem_mask,
11245 __builtin_ia32_pmovsqw512mem_mask,
11246 __builtin_ia32_pmovqw512mem_mask,
11247 __builtin_ia32_pmovusdw512mem_mask,
11248 __builtin_ia32_pmovsdw512mem_mask,
11249 __builtin_ia32_pmovdw512mem_mask,
11250 __builtin_ia32_pmovqb512mem_mask,
11251 __builtin_ia32_pmovusqb512mem_mask,
11252 __builtin_ia32_pmovsqb512mem_mask,
11253 __builtin_ia32_pmovusdb512mem_mask,
11254 __builtin_ia32_pmovsdb512mem_mask,
11255 __builtin_ia32_pmovdb512mem_mask.
11256 (bdesc_args): Add __builtin_ia32_expanddf512,
11257 __builtin_ia32_expandsf512.
11258 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
11259 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
11260 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
11261 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
11262 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
11263 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
11264 (avx512f_<code>v8div16qi2_mask_store): This.
11265 (avx512f_expand<mode>): New.
11267 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
11269 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
11271 (_mm512_mask_prefetch_i64gather_pd): Ditto.
11272 (_mm512_prefetch_i32scatter_pd): Ditto.
11273 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
11274 (_mm512_prefetch_i64scatter_pd): Ditto.
11275 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
11276 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
11277 (_mm512_mask_prefetch_i64gather_ps): Ditto.
11278 (_mm512_prefetch_i32scatter_ps): Ditto.
11279 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
11280 (_mm512_prefetch_i64scatter_ps): Ditto.
11281 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
11282 * config/i386/i386-builtin-types.def: Define
11283 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
11284 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
11285 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
11286 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
11287 IX86_BUILTIN_SCATTERPFQPD.
11288 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
11289 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
11290 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
11291 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
11292 __builtin_ia32_scatterpfqps.
11293 (ix86_expand_builtin): Expand new built-ins.
11294 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
11295 fix memory access data type.
11296 (*avx512pf_gatherpf<mode>_mask): Ditto.
11297 (*avx512pf_gatherpf<mode>): Ditto.
11298 (avx512pf_scatterpf<mode>): Ditto.
11299 (*avx512pf_scatterpf<mode>_mask): Ditto.
11300 (*avx512pf_scatterpf<mode>): Ditto.
11301 (GATHER_SCATTER_SF_MEM_MODE): New.
11302 (avx512pf_gatherpf<mode>df): Ditto.
11303 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11304 (*avx512pf_scatterpf<mode>df): Ditto.
11306 2014-01-27 Jakub Jelinek <jakub@redhat.com>
11309 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
11310 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
11313 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11315 * common/config/arm/arm-common.c
11316 (arm_rewrite_mcpu): Handle multiple names.
11318 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11320 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11322 * gimple-builder.h (create_gimple_tmp): Delete.
11324 2014-01-27 Christian Bruel <christian.bruel@st.com>
11326 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
11329 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
11331 * config/pa/pa.md (call): Generate indirect long calls to non-local
11332 functions when outputing 32-bit code.
11333 (call_value): Likewise except for special call to buggy powf function.
11335 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
11336 portable runtime and PIC indirect calls.
11337 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
11338 and PIC call sequences. Use ldo instead of blr to set return register
11339 in PIC call sequence.
11341 2014-01-25 Walter Lee <walt@tilera.com>
11343 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
11344 avoid clobbering a live register.
11346 2014-01-25 Walter Lee <walt@tilera.com>
11348 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
11349 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
11350 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
11351 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
11353 2014-01-25 Walter Lee <walt@tilera.com>
11355 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
11356 arguments on even registers.
11357 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
11359 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
11360 (BIGGEST_ALIGNMENT): Ditto.
11361 (BIGGEST_FIELD_ALIGNMENT): Ditto.
11363 2014-01-25 Walter Lee <walt@tilera.com>
11365 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
11366 insns before bundling.
11367 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
11369 2014-01-25 Walter Lee <walt@tilera.com>
11371 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
11372 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
11373 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
11375 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11377 * config/mips/constraints.md (kl): Delete.
11378 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
11379 define expands, using...
11380 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
11381 instructions for MIPS16.
11382 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
11383 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
11385 2014-01-25 Walter Lee <walt@tilera.com>
11387 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
11391 2014-01-25 Walter Lee <walt@tilera.com>
11393 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
11394 (TARGET_EXPAND_TO_RTL_HOOK): Define.
11396 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11398 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
11401 2014-01-25 Jakub Jelinek <jakub@redhat.com>
11403 * print-rtl.c (in_call_function_usage): New var.
11404 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
11405 EXPR_LIST mode as mode and not as reg note name.
11407 PR middle-end/59561
11408 * cfgloopmanip.c (copy_loop_info): If
11409 loop->warned_aggressive_loop_optimizations, make sure
11410 the flag is set in target loop too.
11412 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
11414 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
11416 * builtins.def: Likewise.
11417 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
11418 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
11419 * ira.c (ira_setup_eliminable_regset): Likewise.
11420 * omp-low.c (gate_expand_omp): Likewise.
11421 (execute_lower_omp): Likewise.
11422 (diagnose_sb_0): Likewise.
11423 (gate_diagnose_omp_blocks): Likewise.
11424 (simd_clone_clauses_extract): Likewise.
11427 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11429 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
11430 correction for little endian...
11431 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
11434 2014-01-24 Jeff Law <law@redhat.com>
11436 PR tree-optimization/59919
11437 * tree-vrp.c (find_assert_locations_1): Do not register asserts
11438 for non-returning calls.
11440 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
11442 * common/config/aarch64/aarch64-common.c
11443 (aarch64_rewrite_mcpu): Handle multiple names.
11444 * config/aarch64/aarch64.h
11445 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11447 2014-01-24 Dodji Seketeli <dodji@redhat.com>
11449 * input.c (add_file_to_cache_tab): Handle the case where fopen
11452 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
11455 * config/i386/i386.md (pushsf splitter): Get stack adjustment
11456 from push operand if code of push isn't PRE_DEC.
11458 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11461 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
11462 -mquad-memory-atomic. Update -mquad-memory documentation to say
11463 it is only used for non-atomic loads/stores.
11465 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
11466 -mquad-memory or -mquad-memory-atomic switches.
11468 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
11469 -mquad-memory-atomic to ISA 2.07 support.
11471 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
11472 to separate support of normal quad word memory operations (ldq, stq)
11473 from the atomic quad word memory operations.
11475 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
11476 support to separate non-atomic quad word operations from atomic
11477 quad word operations. Disable non-atomic quad word operations in
11478 little endian mode so that we don't have to swap words after the
11479 load and before the store.
11480 (quad_load_store_p): Add comment about atomic quad word support.
11481 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
11482 options printed with -mdebug=reg.
11484 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
11485 -mquad-memory-atomic as the test for whether we have quad word
11486 atomic instructions.
11487 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
11488 or -mp8-vector are used, allow byte/half-word atomic operations.
11490 * config/rs6000/sync.md (load_lockedti): Insure that the address
11491 is a proper indexed or indirect address for the lqarx instruction.
11492 On little endian systems, swap the hi/lo registers after the lqarx
11494 (load_lockedpti): Use indexed_or_indirect_operand predicate to
11495 insure the address is valid for the lqarx instruction.
11496 (store_conditionalti): Insure that the address is a proper indexed
11497 or indirect address for the stqcrx. instruction. On little endian
11498 systems, swap the hi/lo registers before doing the stqcrx.
11500 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
11501 insure the address is valid for the stqcrx. instruction.
11503 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11504 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
11505 type of quad memory support is available.
11507 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
11509 PR regression/59915
11510 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
11511 there is a danger of looping.
11513 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
11515 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11516 force flag_ira_loop_pressure if set via command line.
11518 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11520 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
11521 (ashr_simd): New builtin handling DI mode.
11522 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
11523 (aarch64_sshr_simddi): New match pattern.
11524 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
11525 (vshrd_n_s64): Likewise.
11526 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11528 2014-01-23 Nick Clifton <nickc@redhat.com>
11530 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11531 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11532 favour of mcu specific scripts.
11533 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11536 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11537 Alex Velenko <Alex.Velenko@arm.com>
11539 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11540 (vaddv_s16): Likewise.
11541 (vaddv_s32): Likewise.
11542 (vaddv_u8): Likewise.
11543 (vaddv_u16): Likewise.
11544 (vaddv_u32): Likewise.
11545 (vaddvq_s8): Likewise.
11546 (vaddvq_s16): Likewise.
11547 (vaddvq_s32): Likewise.
11548 (vaddvq_s64): Likewise.
11549 (vaddvq_u8): Likewise.
11550 (vaddvq_u16): Likewise.
11551 (vaddvq_u32): Likewise.
11552 (vaddvq_u64): Likewise.
11553 (vaddv_f32): Likewise.
11554 (vaddvq_f32): Likewise.
11555 (vaddvq_f64): Likewise.
11556 (vmaxv_f32): Likewise.
11557 (vmaxv_s8): Likewise.
11558 (vmaxv_s16): Likewise.
11559 (vmaxv_s32): Likewise.
11560 (vmaxv_u8): Likewise.
11561 (vmaxv_u16): Likewise.
11562 (vmaxv_u32): Likewise.
11563 (vmaxvq_f32): Likewise.
11564 (vmaxvq_f64): Likewise.
11565 (vmaxvq_s8): Likewise.
11566 (vmaxvq_s16): Likewise.
11567 (vmaxvq_s32): Likewise.
11568 (vmaxvq_u8): Likewise.
11569 (vmaxvq_u16): Likewise.
11570 (vmaxvq_u32): Likewise.
11571 (vmaxnmv_f32): Likewise.
11572 (vmaxnmvq_f32): Likewise.
11573 (vmaxnmvq_f64): Likewise.
11574 (vminv_f32): Likewise.
11575 (vminv_s8): Likewise.
11576 (vminv_s16): Likewise.
11577 (vminv_s32): Likewise.
11578 (vminv_u8): Likewise.
11579 (vminv_u16): Likewise.
11580 (vminv_u32): Likewise.
11581 (vminvq_f32): Likewise.
11582 (vminvq_f64): Likewise.
11583 (vminvq_s8): Likewise.
11584 (vminvq_s16): Likewise.
11585 (vminvq_s32): Likewise.
11586 (vminvq_u8): Likewise.
11587 (vminvq_u16): Likewise.
11588 (vminvq_u32): Likewise.
11589 (vminnmv_f32): Likewise.
11590 (vminnmvq_f32): Likewise.
11591 (vminnmvq_f64): Likewise.
11593 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11595 * config/aarch64/aarch64-simd.md
11596 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
11597 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
11598 (*aarch64_mul3_elt<mode>): Likewise.
11599 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
11600 (*aarch64_mul3_elt_to_64v2df): Likewise.
11601 (*aarch64_mla_elt<mode>): Likewise.
11602 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11603 (*aarch64_mls_elt<mode>): Likewise.
11604 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11605 (*aarch64_fma4_elt<mode>): Likewise.
11606 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11607 (*aarch64_fma4_elt_to_64v2df): Likewise.
11608 (*aarch64_fnma4_elt<mode>): Likewise.
11609 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11610 (*aarch64_fnma4_elt_to_64v2df): Likewise.
11611 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
11612 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
11613 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11614 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11615 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
11616 (aarch64_sqdmull_lane<mode>_internal): Likewise.
11617 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
11619 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
11621 * config/aarch64/aarch64-simd.md
11622 (aarch64_be_checked_get_lane<mode>): New define_expand.
11623 * config/aarch64/aarch64-simd-builtins.def
11624 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
11625 New builtin definition.
11626 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
11627 Use new safe be builtin.
11629 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11631 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
11633 (aarch64_be_st1<mode>): Likewise.
11634 (aarch_ld1<VALL:mode>): Define_expand modified.
11635 (aarch_st1<VALL:mode>): Likewise.
11636 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
11637 (UNSPEC_ST1): Likewise.
11639 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
11641 * config/microblaze/microblaze.md: Add trap insn and attribute
11643 2014-01-23 Dodji Seketeli <dodji@redhat.com>
11645 PR preprocessor/58580
11646 * input.h (location_get_source_line): Take an additional line_size
11648 (void diagnostics_file_cache_fini): Declare new function.
11649 * input.c (struct fcache): New type.
11650 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
11651 New static constants.
11652 (diagnostic_file_cache_init, total_lines_num)
11653 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
11654 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
11655 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
11656 (get_next_line, read_next_line, goto_next_line, read_line_num):
11657 New static function definitions.
11658 (diagnostic_file_cache_fini): New function.
11659 (location_get_source_line): Take an additional output line_len
11660 parameter. Re-write using lookup_or_add_file_to_cache_tab and
11662 * diagnostic.c (diagnostic_finish): Call
11663 diagnostic_file_cache_fini.
11664 (adjust_line): Take an additional input parameter for the length
11665 of the line, rather than calculating it with strlen.
11666 (diagnostic_show_locus): Adjust the use of
11667 location_get_source_line and adjust_line with respect to their new
11668 signature. While displaying a line now, do not stop at the first
11669 null byte. Rather, display the zero byte as a space and keep
11670 going until we reach the size of the line.
11671 * Makefile.in: Add vec.o to OBJS-libcommon
11673 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11674 Ilya Tocar <ilya.tocar@intel.com>
11676 * config/i386/avx512fintrin.h (_mm512_kmov): New.
11677 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
11678 (__builtin_ia32_kmov16): Ditto.
11679 * config/i386/i386.md (UNSPEC_KMOV): New.
11682 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11684 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
11685 (_mm512_storeu_si512): Ditto.
11687 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
11690 * rtl.h (get_referenced_operands): Declare.
11691 * recog.c (get_referenced_operands): New function.
11692 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
11693 operands have been referenced when recording LO_SUM references.
11695 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
11697 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
11699 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11701 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
11702 Enable for generic and recent AMD targets.
11704 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11706 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
11707 ARG_SIZE note when adjustment was eliminated.
11709 2014-01-22 Jeff Law <law@redhat.com>
11711 PR tree-optimization/59597
11712 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
11713 in file. Accept new argument REGISTERING and use it to modify
11714 dump output appropriately.
11715 (register_jump_thread): Corresponding changes.
11716 (mark_threaded_blocks): Reinstate code to cancel unprofitable
11717 thread paths involving joiner blocks. Add code to dump cancelled
11718 jump threading paths.
11720 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
11722 PR rtl-optimization/59477
11723 * lra-constraints.c (inherit_in_ebb): Process call for living hard
11724 regs. Update reloads_num and potential_reload_hard_regs for all insns.
11726 2014-01-22 Tom Tromey <tromey@redhat.com>
11728 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
11730 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
11732 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11734 PR rtl-optimization/59896
11735 * lra-constraints.c (process_alt_operands): Check unused note for
11736 matched operands of insn with no output reloads.
11738 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
11740 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
11741 (mips_move_from_gpr_cost): Likewise.
11743 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11745 PR rtl-optimization/59858
11746 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
11747 ira_class_hard_regs_num.
11748 (process_alt_operands): Increase reject for dying matched operand.
11750 2014-01-21 Jakub Jelinek <jakub@redhat.com>
11753 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
11754 smaller than size, perform several stores or loads and stores
11755 at dst + count - size to store or copy all of size bytes, rather
11756 than just last modesize bytes.
11758 2014-01-20 DJ Delorie <dj@redhat.com>
11760 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
11761 that CLOBBERs are REGs before propogating their values.
11763 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
11765 PR middle-end/59789
11766 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
11767 (cgraph_inline_failed_type): New function.
11768 * cgraph.h (DEFCIFCODE): Add type.
11769 (cgraph_inline_failed_type_t): New enum.
11770 (cgraph_inline_failed_type): New prototype.
11771 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
11772 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
11773 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
11774 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
11775 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
11776 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
11777 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
11778 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
11779 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
11780 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
11781 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
11782 OPTIMIZATION_MISMATCH.
11783 * tree-inline.c (expand_call_inline): Emit errors during
11784 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
11786 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11789 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
11790 mode attribute in insn output.
11792 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
11794 * output.h (output_constant): Delete.
11795 * varasm.c (output_constant): Make private.
11797 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
11799 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
11801 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11803 PR middle-end/59860
11804 * tree.h (fold_builtin_strcat): New prototype.
11805 * builtins.c (fold_builtin_strcat): No longer static. Add len
11806 argument, if non-NULL, don't call c_strlen. Optimize
11807 directly into __builtin_memcpy instead of __builtin_strcpy.
11808 (fold_builtin_2): Adjust fold_builtin_strcat caller.
11809 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
11811 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11813 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11814 for SImode_address_operand operands, having only a REG argument.
11816 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
11818 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
11819 loader name using mbig-endian.
11820 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
11822 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11824 * doc/invoke.texi (-march): Clarify documentation for AArch64.
11825 (-mtune): Likewise.
11828 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
11830 * config/aarch64/aarch64-protos.h
11831 (aarch64_cannot_change_mode_class_ptr): Declare.
11832 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
11833 aarch64_cannot_change_mode_class_ptr): New.
11834 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
11835 backend hook aarch64_cannot_change_mode_class.
11837 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11839 * common/config/aarch64/aarch64-common.c
11840 (aarch64_handle_option): Don't handle any option order logic here.
11841 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
11842 selected_cpu, warn on architecture version mismatch.
11843 (aarch64_override_options): Fix parsing order for option strings.
11845 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11846 Iain Sandoe <iain@codesourcery.com>
11849 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
11850 warning. Amend comment to reflect current functionality.
11852 2014-01-20 Richard Biener <rguenther@suse.de>
11854 PR middle-end/59860
11855 * builtins.c (fold_builtin_strcat): Remove case better handled
11856 by tree-ssa-strlen.c.
11858 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
11860 * config/aarch64/aarch64.opt
11861 (mcpu, march, mtune): Make case-insensitive.
11863 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11866 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11867 if operands[1] is a REG or ZERO_EXTEND of a REG.
11869 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
11871 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
11873 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
11875 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
11876 long non-pic millicode calls.
11878 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11880 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
11882 2014-01-19 Kito Cheng <kito@0xlab.org>
11884 * builtins.c (expand_movstr): Check movstr expand done or fail.
11886 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11887 H.J. Lu <hongjiu.lu@intel.com>
11890 * config/i386/i386.md (*lea<mode>): Zero-extend return register
11891 to DImode for zero-extended addresses.
11893 2014-01-19 Jakub Jelinek <jakub@redhat.com>
11895 PR rtl-optimization/57763
11896 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
11897 on the new indirect jump_insn and increment LABEL_NUSES (label).
11899 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
11903 * config.gcc (x86_archs): New variable.
11904 (x86_64_archs): Likewise.
11905 (x86_cpus): Likewise.
11906 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
11907 --with-arch/--with-cpu= options.
11908 Support --with-arch=/--with-cpu={nehalem,westmere,
11909 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
11911 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11913 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
11914 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
11916 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11918 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
11920 2014-01-18 Jakub Jelinek <jakub@redhat.com>
11923 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
11924 clear cpp_get_options (parse_in)->warn_unused_macros for
11925 ix86_target_macros_internal with cpp_define.
11927 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
11929 * jump.c (delete_related_insns): Keep (use (insn))s.
11930 * reorg.c (redundant_insn): Check for barriers too.
11932 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11934 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
11936 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
11938 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
11939 call to $$dyncall when TARGET_LONG_CALLS is true.
11941 2014-01-17 Jeff Law <law@redhat.com>
11943 * ree.c (combine_set_extension): Temporarily disable test for
11944 changing number of hard registers.
11946 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11948 PR middle-end/58125
11949 * ipa-inline-analysis.c (inline_free_summary):
11950 Do not free summary of aliases.
11952 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11954 PR middle-end/59706
11955 * gimplify.c (gimplify_expr): Use create_tmp_var
11956 instead of create_tmp_var_raw. If cond doesn't have
11957 integral type, don't add the IFN_ANNOTATE builtin at all.
11959 2014-01-17 Martin Jambor <mjambor@suse.cz>
11962 * ipa-cp.c (prev_edge_clone): New variable.
11963 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
11964 Also resize prev_edge_clone vector.
11965 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
11966 (ipcp_edge_removal_hook): New function.
11967 (ipcp_driver): Register ipcp_edge_removal_hook.
11969 2014-01-17 Andrew Pinski <apinski@cavium.com>
11970 Steve Ellcey <sellcey@mips.com>
11973 * config/mips/mips.c (mips_print_operand): Check operand mode instead
11976 2014-01-17 Jeff Law <law@redhat.com>
11978 PR middle-end/57904
11979 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
11980 so that pass_ccp runs first.
11982 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11984 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
11985 (ix86_adjust_cost): Use !TARGET_XXX.
11986 (do_reorder_for_imul): Likewise.
11987 (swap_top_of_ready_list): Likewise.
11988 (ix86_sched_reorder): Likewise.
11990 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11992 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
11993 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
11994 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
11995 (intel_memset): New. Duplicate slm_memset.
11996 (intel_cost): New. Duplicate slm_cost.
11997 (m_INTEL): New macro.
11998 (processor_target_table): Add "intel".
11999 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
12000 with PROCESSOR_INTEL for "intel".
12001 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
12002 PROCESSOR_SILVERMONT.
12003 (ix86_issue_rate): Likewise.
12004 (ix86_adjust_cost): Likewise.
12005 (ia32_multipass_dfa_lookahead): Likewise.
12006 (swap_top_of_ready_list): Likewise.
12007 (ix86_sched_reorder): Likewise.
12008 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
12009 instead of TARGET_OPT_AGU.
12010 * config/i386/i386.h (TARGET_INTEL): New.
12011 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
12012 (processor_type): Add PROCESSOR_INTEL.
12013 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
12014 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
12016 2014-01-17 Marek Polacek <polacek@redhat.com>
12019 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
12022 2014-01-17 Richard Biener <rguenther@suse.de>
12024 PR tree-optimization/46590
12025 * opts.c (default_options_table): Add entries for
12026 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
12027 all enabled at -O1 but not for -Og.
12028 * common.opt (fbranch-count-reg): Remove Init(1).
12029 (fmove-loop-invariants): Likewise.
12030 (ftree-pta): Likewise.
12032 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12034 * config/i386/i386.c (ix86_data_alignment): For compatibility with
12035 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
12036 decls to at least the GCC 4.8 used alignments.
12039 * tree-nested.c (convert_nonlocal_reference_stmt,
12040 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
12041 of GIMPLE_BIND stmts, adjust associated decls.
12043 2014-01-17 Richard Biener <rguenther@suse.de>
12045 PR tree-optimization/46590
12046 * vec.h (vec<>::bseach): New member function implementing
12047 binary search according to C89 bsearch.
12048 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
12049 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
12050 bitmap pointer again. Make accesses_in_loop a flat array.
12051 (mem_ref_obstack): New global.
12052 (outermost_indep_loop): Adjust for mem_ref->stored changes.
12053 (mark_ref_stored): Likewise.
12054 (ref_indep_loop_p_2): Likewise.
12055 (set_ref_stored_in_loop): New helper function.
12056 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
12057 (memref_free): Adjust.
12058 (record_mem_ref_loc): Simplify.
12059 (gather_mem_refs_stmt): Adjust.
12060 (sort_locs_in_loop_postorder_cmp): New function.
12061 (analyze_memory_references): Sort accesses_in_loop after
12062 loop postorder number.
12063 (find_ref_loc_in_loop_cmp): New function.
12064 (for_all_locs_in_loop): Find relevant cluster of locs in
12065 accesses_in_loop and iterate without recursion.
12066 (execute_sm): Avoid uninit warning.
12067 (struct ref_always_accessed): Simplify.
12068 (ref_always_accessed::operator ()): Likewise.
12069 (ref_always_accessed_p): Likewise.
12070 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
12071 loop postorder numbers here.
12072 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
12075 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12078 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
12079 on decls for which assemble_alias has been called.
12081 2014-01-17 Nick Clifton <nickc@redhat.com>
12083 * config/msp430/msp430.opt: (mcpu): New option.
12084 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
12085 (msp430_option_override): Parse target_cpu. If the MCU name
12086 matches a generic string, clear target_mcu.
12087 (msp430_attr): Allow numeric interrupt values up to 63.
12088 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
12089 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
12091 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
12093 * config/msp430/msp430.md (popm): Use %J rather than %I.
12094 (addsi3): Use msp430_nonimmediate_operand for operand 2.
12095 (addhi_cy_i): Use immediate_operand for operand 2.
12096 * doc/invoke.texi: Document -mcpu option.
12098 2014-01-17 Richard Biener <rguenther@suse.de>
12100 PR rtl-optimization/38518
12101 * df.h (df_analyze_loop): Declare.
12102 * df-core.c: Include cfgloop.h.
12103 (df_analyze_1): Split out main part of df_analyze.
12104 (df_analyze): Adjust.
12105 (loop_inverted_post_order_compute): New function.
12106 (loop_post_order_compute): Likewise.
12107 (df_analyze_loop): New function avoiding whole-function
12108 postorder computes.
12109 * loop-invariant.c (find_defs): Use df_analyze_loop.
12110 (find_invariants): Adjust.
12111 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
12113 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
12115 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
12116 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
12118 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
12120 * ipa-ref.c (ipa_remove_stmt_references): Fix references
12121 traversal when removing references.
12123 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
12126 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
12128 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
12130 PR middle-end/56791
12131 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
12132 pushing a reload for an autoinc when we had previously reloaded an
12133 inner part of the address.
12135 2014-01-16 Jakub Jelinek <jakub@redhat.com>
12137 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
12139 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
12140 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
12141 when not giving up or versioning for alias only because of
12143 (vect_analyze_data_ref_dependences): Set to true.
12144 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
12146 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
12147 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
12150 PR middle-end/58344
12151 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
12154 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
12155 operand 0 predicate for gathers, use a new pseudo as subtarget.
12157 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12159 PR middle-end/59609
12160 * lra-constraints.c (process_alt_operands): Add printing debug info.
12161 Check absence of input/output reloads for matched operands too.
12163 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12165 PR rtl-optimization/59835
12166 * ira.c (ira_init_register_move_cost): Increase cost for
12169 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
12171 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
12173 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
12176 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
12177 non-register objects. Use gen_(high/low)part more consistently.
12180 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12183 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
12184 endian support, remove tests for WORDS_BIG_ENDIAN.
12185 (p8_mfvsrd_3_<mode>): Likewise.
12186 (reload_gpr_from_vsx<mode>): Likewise.
12187 (reload_gpr_from_vsxsf): Likewise.
12188 (p8_mfvsrd_4_disf): Likewise.
12190 2014-01-16 Richard Biener <rguenther@suse.de>
12192 PR rtl-optimization/46590
12193 * lcm.c (compute_antinout_edge): Use postorder iteration.
12194 (compute_laterin): Use inverted postorder iteration.
12196 2014-01-16 Nick Clifton <nickc@redhat.com>
12198 PR middle-end/28865
12199 * varasm.c (output_constant): Return the number of bytes actually
12201 (output_constructor_array_range): Update the field size with the
12202 number of bytes emitted by output_constant.
12203 (output_constructor_regular_field): Likewise. Also do not
12204 complain if the total number of bytes emitted is now greater
12205 than the expected fieldpos.
12206 * output.h (output_constant): Update prototype and descriptive comment.
12208 2014-01-16 Marek Polacek <polacek@redhat.com>
12210 PR middle-end/59827
12211 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
12212 it is error_mark_node.
12214 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
12216 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
12217 VALID_AVX256_REG_OR_OI_MODE.
12219 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
12221 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
12222 current procedure should be profiled.
12224 2014-01-15 Andrew Pinski <apinski@cavium.com>
12226 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
12227 of moving from/to the STACK_REG register class.
12229 2014-01-15 Richard Henderson <rth@redhat.com>
12232 * reginfo.c (global_regs_decl): Globalize.
12233 * rtl.h (global_regs_decl): Declare.
12234 * ira.c (do_reload): Diagnose frame_pointer_needed and it
12235 reserved via global_regs.
12237 2014-01-15 Teresa Johnson <tejohnson@google.com>
12239 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
12241 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12243 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
12244 and vmulosh rather than call gen_vec_widen_smult_*.
12245 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
12246 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
12247 (vec_widen_smult_even_v16qi): Likewise.
12248 (vec_widen_umult_even_v8hi): Likewise.
12249 (vec_widen_smult_even_v8hi): Likewise.
12250 (vec_widen_umult_odd_v16qi): Likewise.
12251 (vec_widen_smult_odd_v16qi): Likewise.
12252 (vec_widen_umult_odd_v8hi): Likewise.
12253 (vec_widen_smult_odd_v8hi): Likewise.
12254 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
12255 vmuloub rather than call gen_vec_widen_umult_*.
12256 (vec_widen_umult_lo_v16qi): Likewise.
12257 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
12258 vmulosb rather than call gen_vec_widen_smult_*.
12259 (vec_widen_smult_lo_v16qi): Likewise.
12260 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
12261 rather than call gen_vec_widen_umult_*.
12262 (vec_widen_umult_lo_v8hi): Likewise.
12263 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
12264 rather than call gen_vec_widen_smult_*.
12265 (vec_widen_smult_lo_v8hi): Likewise.
12267 2014-01-15 Jeff Law <law@redhat.com>
12269 PR tree-optimization/59747
12270 * ree.c (find_and_remove_re): Properly handle case where a second
12271 eliminated extension requires widening a copy created for elimination
12272 of a prior extension.
12273 (combine_set_extension): Ensure that the number of hard regs needed
12274 for a destination register does not change when we widen it.
12276 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
12278 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
12279 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
12280 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
12281 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
12282 (avr-*-rtems*): Likewise.
12283 (bfin*-rtems*): Likewise.
12284 (moxie-*-rtems*): Likewise.
12285 (h8300-*-rtems*): Likewise.
12286 (i[34567]86-*-rtems*): Likewise.
12287 (lm32-*-rtems*): Likewise.
12288 (m32r-*-rtems*): Likewise.
12289 (m68k-*-rtems*): Likewise.
12290 (microblaze*-*-rtems*): Likewise.
12291 (mips*-*-rtems*): Likewise.
12292 (powerpc-*-rtems*): Likewise.
12293 (sh-*-rtems*): Likewise.
12294 (sparc-*-rtems*): Likewise.
12295 (sparc64-*-rtems*): Likewise.
12296 (v850-*-rtems*): Likewise.
12297 (m32c-*-rtems*): Likewise.
12299 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
12301 PR rtl-optimization/59511
12302 * ira.c (ira_init_register_move_cost): Use memory costs for some
12303 cases of register move cost calculations.
12304 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
12305 instead of BB frequency.
12306 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
12307 * lra-assigns.c (find_hard_regno_for): Ditto.
12309 2014-01-15 Richard Biener <rguenther@suse.de>
12311 PR tree-optimization/59822
12312 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
12313 (vectorizable_load): Use it to hoist defs of uses of invariant
12314 loads out of the loop.
12316 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12317 Kugan Vivekanandarajah <kuganv@linaro.org>
12320 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
12323 2014-01-15 Richard Biener <rguenther@suse.de>
12325 PR rtl-optimization/59802
12326 * lcm.c (compute_available): Use inverted postorder to seed
12327 the initial worklist.
12329 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12332 * config/s390/s390.c (s390_preferred_reload_class): Don't return
12333 ADDR_REGS for invalid symrefs in non-PIC code.
12335 2014-01-15 Jakub Jelinek <jakub@redhat.com>
12338 * builtins.c (determine_block_size): Initialize *probable_max_size
12339 even if len_rtx is CONST_INT.
12341 2014-01-14 Andrew Pinski <apinski@cavium.com>
12343 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
12344 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
12345 (cortexa53_tunings): Likewise.
12346 (aarch64_sched_issue_rate): New function.
12347 (TARGET_SCHED_ISSUE_RATE): Define.
12349 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12351 * ira-costs.c (find_costs_and_classes): Add missed
12352 ira_init_register_move_cost_if_necessary.
12354 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12357 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
12359 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
12362 * config/i386/i386.c (type_natural_mode): Add a bool parameter
12363 to indicate if type is used for function return value. Warn ABI
12364 change if the vector mode isn't available for function return value.
12365 (ix86_function_arg_advance): Pass false to type_natural_mode.
12366 (ix86_function_arg): Likewise.
12367 (ix86_gimplify_va_arg): Likewise.
12368 (function_arg_32): Don't warn ABI change.
12369 (ix86_function_value): Pass true to type_natural_mode.
12370 (ix86_return_in_memory): Likewise.
12371 (ix86_struct_value_rtx): Removed.
12372 (TARGET_STRUCT_VALUE_RTX): Likewise.
12374 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12376 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
12377 converting a conditional jump into a conditional return.
12379 2014-01-14 Richard Biener <rguenther@suse.de>
12381 PR tree-optimization/58921
12382 PR tree-optimization/59006
12383 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
12384 hoisting invariant stmts.
12385 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
12386 invariant loads on the preheader edge if possible.
12388 2014-01-14 Joey Ye <joey.ye@arm.com>
12390 * doc/plugin.texi (Building GCC plugins): Update to C++.
12392 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
12394 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
12395 (_mm_rcp28_round_ss): Ditto.
12396 (_mm_rsqrt28_round_sd): Ditto.
12397 (_mm_rsqrt28_round_ss): Ditto.
12398 (_mm_rcp28_sd): Ditto.
12399 (_mm_rcp28_ss): Ditto.
12400 (_mm_rsqrt28_sd): Ditto.
12401 (_mm_rsqrt28_ss): Ditto.
12402 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
12403 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
12404 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
12405 (IX86_BUILTIN_RCP28SD): Ditto.
12406 (IX86_BUILTIN_RCP28SS): Ditto.
12407 (IX86_BUILTIN_RSQRT28SD): Ditto.
12408 (IX86_BUILTIN_RSQRT28SS): Ditto.
12409 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
12410 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
12411 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
12412 (ix86_expand_special_args_builtin): Expand new FTYPE.
12413 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
12414 (srcp14<mode>): Make insn unary.
12415 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
12416 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
12417 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
12418 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
12419 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
12420 Fix rounding: make it SAE only.
12421 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
12423 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
12425 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
12426 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
12427 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
12428 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
12429 (round_saeonly_mask_scalar_operand4): Ditto.
12430 (round_saeonly_mask_scalar_op3): Ditto.
12431 (round_saeonly_mask_scalar_op4): Ditto.
12433 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12435 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12436 Implement -maltivec=be for vec_insert and vec_extract.
12438 2014-01-10 DJ Delorie <dj@redhat.com>
12440 * config/msp430/msp430.md (call_internal): Don't allow memory
12441 references with SP as the base register.
12442 (call_value_internal): Likewise.
12443 * config/msp430/constraints.md (Yc): New. For memory references
12444 that don't use SP as a base register.
12446 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
12447 "an integer without a # prefix"
12448 * config/msp430/msp430.md (epilogue_helper): Use it.
12450 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12453 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
12454 AVX512F gather builtins.
12455 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
12456 on gather decls with INTEGER_TYPE masktype.
12457 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
12458 directly into the builtin rather than hoisting it before loop.
12460 PR tree-optimization/59387
12461 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
12462 (scev_const_prop): If folded_casts and type has undefined overflow,
12463 use force_gimple_operand instead of force_gimple_operand_gsi and
12464 for each added stmt if it is assign with
12465 arith_code_with_undefined_signed_overflow, call
12466 rewrite_to_defined_overflow.
12467 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
12468 gimple-fold.h instead.
12469 (arith_code_with_undefined_signed_overflow,
12470 rewrite_to_defined_overflow): Moved to ...
12471 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
12472 rewrite_to_defined_overflow): ... here. No longer static.
12473 Include gimplify-me.h.
12474 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
12475 rewrite_to_defined_overflow): New prototypes.
12477 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12479 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
12481 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
12483 * builtins.c (get_object_alignment_2): Minor tweak.
12484 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
12486 2014-01-13 Christian Bruel <christian.bruel@st.com>
12488 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
12489 optimized non constant lengths.
12491 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12494 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
12495 load as __atomic_load_N if possible.
12497 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12499 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
12501 (rs6000_expand_builtin): Adjust call.
12503 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12506 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
12507 * config/rs6000/rs6000.c: Include target-globals.h.
12508 (rs6000_set_current_function): Instead of doing target_reinit
12509 unconditionally, use save_target_globals_default_opts and
12510 restore_target_globals.
12512 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
12514 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
12515 (rs6000_expand_builtin): Handle mffs and mtfsf.
12516 (rs6000_init_builtins): Define mffs and mtfsf.
12517 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
12518 (rs6000_mffs): New pattern.
12519 (rs6000_mtfsf): New pattern.
12521 2014-01-11 Bin Cheng <bin.cheng@arm.com>
12523 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
12524 Start narrowing with START. Apply candidate-use pair
12525 and check overall cost in narrowing.
12526 (iv_ca_prune): Pass new argument.
12528 2014-01-10 Jeff Law <law@redhat.com>
12530 PR middle-end/59743
12531 * ree.c (combine_reaching_defs): Ensure the defining statement
12532 occurs before the extension when optimizing extensions with
12533 different source and destination hard registers.
12535 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12538 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12539 vtables into the type inheritance graph.
12541 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12543 PR rtl-optimization/59754
12544 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12545 modes in the REGNO != REGNO case.
12547 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12549 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12551 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12553 PR tree-optimization/59745
12554 * tree-predcom.c (tree_predictive_commoning_loop): Call
12555 free_affine_expand_cache if giving up because components is NULL.
12557 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12558 GC in payload of target_globals struct instead of allocating them on
12559 the heap and the larger structs separately using GC.
12560 * target-globals.h (struct target_globals): Make regs, hard_regs,
12561 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12562 of GTY((skip)) and change type to void *.
12563 (reset_target_globals): Cast loads from those fields to corresponding
12566 2014-01-10 Steve Ellcey <sellcey@mips.com>
12569 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12570 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12571 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12573 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12576 * aarch64-modes.def (CC_Zmode): New flags mode.
12577 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12578 represents an equality.
12579 (aarch64_get_condition_code): Handle CC_Zmode.
12580 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12582 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12584 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12585 extraction in good case.
12587 2014-01-10 Richard Biener <rguenther@suse.de>
12589 PR tree-optimization/59374
12590 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
12591 checking after SLP discovery. Mark stmts not participating
12592 in any SLP instance properly.
12594 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12596 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
12597 when handling a SET rtx.
12599 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12601 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
12602 (cortex-a57): Likewise.
12603 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
12605 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12607 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
12608 non-iwmmxt builtins.
12610 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12614 * ipa-devirt.c record_target_from_binfo): Take as argument
12615 stack of binfos and lookup matching one for virtual inheritance.
12616 (possible_polymorphic_call_targets_1): Update.
12618 2014-01-10 Huacai Chen <chenhc@lemote.com>
12620 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
12621 kernel strings for Loongson-2E/2F/3A.
12623 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12625 PR middle-end/59670
12626 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
12627 is_gimple_call before calling gimple_call_internal_p.
12629 2014-01-09 Steve Ellcey <sellcey@mips.com>
12631 * Makefile.in (TREE_FLOW_H): Remove.
12632 (TREE_SSA_H): Add file names from tree-flow.h.
12633 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
12634 * tree.h: Remove tree-flow.h reference.
12635 * hash-table.h: Remove tree-flow.h reference.
12636 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
12637 reference with tree-ssa-loop.h.
12639 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12641 * doc/invoke.texi: Add -maltivec={be,le} options, and document
12642 default element-order behavior for -maltivec.
12643 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
12644 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
12645 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
12646 when targeting big endian, at least for now.
12647 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
12649 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12651 PR middle-end/47735
12652 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
12653 var satisfies use_register_for_decl, just take into account type
12654 alignment, rather than decl alignment.
12656 PR tree-optimization/59622
12657 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
12658 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
12659 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
12660 Don't devirtualize for inplace at all. For targets.length () == 1,
12661 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
12663 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12665 * config/i386/i386.md (cpu): Remove the unused btver1.
12667 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12669 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
12671 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12674 * tree-core.h (struct target_globals): New forward declaration.
12675 (struct tree_target_option): Add globals field.
12676 * tree.h (TREE_TARGET_GLOBALS): Define.
12677 (prepare_target_option_nodes_for_pch): New prototype.
12678 * target-globals.h (struct target_globals): Define even if
12679 !SWITCHABLE_TARGET.
12680 * tree.c (prepare_target_option_node_for_pch,
12681 prepare_target_option_nodes_for_pch): New functions.
12682 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
12683 * config/i386/i386.c: Include target-globals.h.
12684 (ix86_set_current_function): Instead of doing target_reinit
12685 unconditionally, use save_target_globals_default_opts and
12686 restore_target_globals.
12688 2014-01-09 Richard Biener <rguenther@suse.de>
12690 PR tree-optimization/59715
12691 * tree-cfg.h (split_critical_edges): Declare.
12692 * tree-cfg.c (split_critical_edges): Export.
12693 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
12695 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
12697 * cfgexpand.c (expand_stack_vars): Optionally disable
12698 asan stack protection.
12699 (expand_used_vars): Likewise.
12700 (partition_stack_vars): Likewise.
12701 * asan.c (asan_emit_stack_protection): Optionally disable
12702 after return stack usage.
12703 (instrument_derefs): Optionally disable memory access instrumentation.
12704 (instrument_builtin_call): Likewise.
12705 (instrument_strlen_call): Likewise.
12706 (asan_protect_global): Optionally disable global variables protection.
12707 * doc/invoke.texi: Added doc for new options.
12708 * params.def: Added new options.
12709 * params.h: Likewise.
12711 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12713 PR rtl-optimization/59724
12714 * ifcvt.c (cond_exec_process_if_block): Don't call
12715 flow_find_head_matching_sequence with 0 longest_match.
12716 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
12717 non-active insns if !stop_after.
12718 (try_head_merge_bb): Revert 2014-01-07 changes.
12720 2014-01-08 Jeff Law <law@redhat.com>
12722 * ree.c (get_sub_rtx): New function, extracted from...
12723 (merge_def_and_ext): Here.
12724 (combine_reaching_defs): Use get_sub_rtx.
12726 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
12728 * cgraph.h (varpool_variable_node): Do not choke on null node.
12730 2014-01-08 Catherine Moore <clm@codesourcery.com>
12732 * config/mips/mips.md (simple_return): Attempt to use JRC
12734 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
12736 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12738 PR rtl-optimization/59137
12739 * reorg.c (steal_delay_list_from_target): Call update_block for
12741 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
12743 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12745 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
12746 two duplicate entries.
12748 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12751 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
12753 * config/mips/mips.c (mips_truncated_op_cost): New function.
12754 (mips_rtx_costs): Adjust test for BADDU.
12755 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
12757 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
12759 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
12760 (*baddu_si): ...this new pattern.
12762 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12765 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
12767 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12769 PR middle-end/57748
12770 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
12772 (expand_expr, expand_normal): Adjust.
12773 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
12774 inner_reference_p. Use inner_reference_p to expand inner references.
12775 (store_expr): Adjust.
12776 * cfgexpand.c (expand_call_stmt): Adjust.
12778 2014-01-08 Rong Xu <xur@google.com>
12780 * gcov-io.c (gcov_var): Move from gcov-io.h.
12781 (gcov_position): Ditto.
12782 (gcov_is_error): Ditto.
12783 (gcov_rewrite): Ditto.
12784 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
12785 only part to libgcc/libgcov.h.
12787 2014-01-08 Marek Polacek <polacek@redhat.com>
12789 PR middle-end/59669
12790 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
12792 2014-01-08 Marek Polacek <polacek@redhat.com>
12795 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
12797 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12799 PR rtl-optimization/59649
12800 * stor-layout.c (get_mode_bounds): For BImode return
12801 0 and STORE_FLAG_VALUE.
12803 2014-01-08 Richard Biener <rguenther@suse.de>
12805 PR middle-end/59630
12806 * gimple.h (is_gimple_builtin_call): Remove.
12807 (gimple_builtin_call_types_compatible_p): New.
12808 (gimple_call_builtin_p): New overload.
12809 * gimple.c (is_gimple_builtin_call): Remove.
12810 (validate_call): Rename to ...
12811 (gimple_builtin_call_types_compatible_p): ... this and export. Also
12812 check return types.
12813 (validate_type): New static function.
12814 (gimple_call_builtin_p): New overload and adjust.
12815 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
12816 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
12817 (gimple_fold_stmt_to_constant_1): Likewise.
12818 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
12820 2014-01-08 Richard Biener <rguenther@suse.de>
12822 PR middle-end/59471
12823 * gimplify.c (gimplify_expr): Gimplify register-register type
12824 VIEW_CONVERT_EXPRs to separate stmts.
12826 2014-01-07 Jeff Law <law@redhat.com>
12828 PR middle-end/53623
12829 * ree.c (combine_set_extension): Handle case where source
12830 and destination registers in an extension insn are different.
12831 (combine_reaching_defs): Allow source and destination registers
12832 in extension to be different under limited circumstances.
12833 (add_removable_extension): Remove restriction that the
12834 source and destination registers in the extension are the same.
12835 (find_and_remove_re): Emit a copy from the extension's
12836 destination to its source after the defining insn if
12837 the source and destination registers are different.
12839 PR middle-end/59285
12840 * ifcvt.c (merge_if_block): If we are merging a block with more than
12841 one successor with a block with no successors, remove any BARRIER
12842 after the second block.
12844 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
12846 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
12848 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
12851 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
12852 for 14-bit register offsets when INT14_OK_STRICT is false.
12854 2014-01-07 Roland Stigge <stigge@antcom.de>
12855 Michael Meissner <meissner@linux.vnet.ibm.com>
12858 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
12859 Only check TFmode for SPE constants. Don't check TImode or TDmode.
12861 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
12863 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
12866 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
12868 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
12869 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
12870 rtx is const0_rtx or not.
12872 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
12875 * target-globals.c (save_target_globals): Remove this_fn_optab
12877 * toplev.c: Include optabs.h.
12878 (target_reinit): Temporarily restore the global options if another
12879 set of options are in force.
12881 2014-01-07 Jakub Jelinek <jakub@redhat.com>
12883 PR rtl-optimization/58668
12884 * cfgcleanup.c (flow_find_cross_jump): Don't count
12885 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
12886 to determine what is counted.
12887 (flow_find_head_matching_sequence): Use active_insn_p to determine
12889 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
12891 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
12892 determine what is counted.
12894 PR tree-optimization/59643
12895 * tree-predcom.c (split_data_refs_to_components): If one dr is
12896 read and one write, determine_offset fails and the write isn't
12897 in the bad component, just put the read into the bad component.
12899 2014-01-07 Mike Stump <mikestump@comcast.net>
12900 Jakub Jelinek <jakub@redhat.com>
12903 * tree-core.h (struct tree_optimization_option): Change optabs
12904 type from unsigned char * to void *.
12905 * optabs.c (init_tree_optimization_optabs): Adjust
12906 TREE_OPTIMIZATION_OPTABS initialization.
12908 2014-01-06 Jakub Jelinek <jakub@redhat.com>
12911 * config/i386/i386.h (struct machine_function): Add
12912 no_drap_save_restore field.
12913 * config/i386/i386.c (ix86_save_reg): Use
12914 !cfun->machine->no_drap_save_restore instead of
12915 crtl->stack_realign_needed.
12916 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
12917 this function clears frame_pointer_needed. Set
12918 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
12919 and DRAP reg is needed.
12921 2014-01-06 Marek Polacek <polacek@redhat.com>
12924 * doc/implement-c.texi: Mention that other integer types are
12925 permitted as bit-field types in strictly conforming mode.
12927 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
12929 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
12930 is newly allocated.
12932 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
12934 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
12936 2014-01-06 Martin Jambor <mjambor@suse.cz>
12939 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
12941 * ipa-prop.c (ipa_print_node_params): Fix indentation.
12943 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
12947 * var-tracking.c (add_stores): Preserve the value of the source even if
12948 we don't record the store.
12950 2014-01-06 Terry Guo <terry.guo@arm.com>
12952 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
12954 2014-01-05 Iain Sandoe <iain@codesourcery.com>
12957 * config/darwin.c (darwin_function_section): Adjust return values to
12958 correspond to optimisation changes made in r206070.
12960 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
12962 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
12963 from prefetch_block tune setting.
12964 (nocona_cost): Correct size of prefetch block to 64.
12966 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
12968 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
12969 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
12970 used to save the static chain register in the computation of the offset
12971 from which the FP registers need to be restored.
12973 2014-01-04 Jakub Jelinek <jakub@redhat.com>
12975 PR tree-optimization/59519
12976 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
12977 ICE if get_current_def (current_new_name) is already non-NULL, as long
12978 as it is a phi result of some other phi in *new_exit_bb that has
12981 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
12982 or vmovdqu* for misaligned_operand.
12983 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
12984 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
12985 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
12986 aligned_mem for AVX512F masked aligned load and store builtins and for
12987 non-temporal moves.
12989 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
12991 PR tree-optimization/59651
12992 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
12993 Address range for negative step should be added by TYPE_SIZE_UNIT.
12995 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
12997 * config/m68k/m68k.c (handle_move_double): Handle pushes with
12998 overlapping registers also for registers other than the stack pointer.
13000 2014-01-03 Marek Polacek <polacek@redhat.com>
13003 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
13006 2014-01-03 Jakub Jelinek <jakub@redhat.com>
13009 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
13012 * config/i386/i386.md (MODE_SIZE): New mode attribute.
13013 (push splitter): Use <P:MODE_SIZE> instead of
13014 GET_MODE_SIZE (<P:MODE>mode).
13015 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
13016 (mov -1, reg peephole2): Likewise.
13017 * config/i386/sse.md (*mov<mode>_internal,
13018 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
13019 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
13020 *<code><mode>3, *andnot<mode>3<mask_name>,
13021 <mask_codefor><code><mode>3<mask_name>): Likewise.
13022 * config/i386/subst.md (mask_mode512bit_condition,
13023 sd_mask_mode512bit_condition): Likewise.
13025 2014-01-02 Xinliang David Li <davidxl@google.com>
13027 PR tree-optimization/59303
13028 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
13029 (dump_predicates): Better output format.
13030 (pred_equal_p): New function.
13031 (is_neq_relop_p): Ditto.
13032 (is_neq_zero_form_p): Ditto.
13033 (pred_expr_equal_p): Ditto.
13034 (pred_neg_p): Ditto.
13035 (simplify_pred): Ditto.
13036 (simplify_preds_2): Ditto.
13037 (simplify_preds_3): Ditto.
13038 (simplify_preds_4): Ditto.
13039 (simplify_preds): Ditto.
13040 (push_pred): Ditto.
13041 (push_to_worklist): Ditto.
13042 (get_pred_info_from_cmp): Ditto.
13043 (is_degenerated_phi): Ditto.
13044 (normalize_one_pred_1): Ditto.
13045 (normalize_one_pred): Ditto.
13046 (normalize_one_pred_chain): Ditto.
13047 (normalize_preds): Ditto.
13048 (normalize_cond_1): Remove function.
13049 (normalize_cond): Ditto.
13050 (is_gcond_subset_of): Ditto.
13051 (is_subset_of_any): Ditto.
13052 (is_or_set_subset_of): Ditto.
13053 (is_and_set_subset_of): Ditto.
13054 (is_norm_cond_subset_of): Ditto.
13055 (pred_chain_length_cmp): Ditto.
13056 (convert_control_dep_chain_into_preds): Type change.
13057 (find_predicates): Ditto.
13058 (find_def_preds): Ditto.
13059 (destroy_predicates_vecs): Ditto.
13060 (find_matching_predicates_in_rest_chains): Ditto.
13061 (use_pred_not_overlap_with_undef_path_pred): Ditto.
13062 (is_pred_expr_subset): Ditto.
13063 (is_pred_chain_subset_of): Ditto.
13064 (is_included_in): Ditto.
13065 (is_superset_of): Ditto.
13067 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13069 Update copyright years.
13071 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13073 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
13074 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
13075 config/arc/arc.md, config/arc/arc.opt,
13076 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
13077 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
13078 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
13079 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
13080 config/linux-protos.h, config/linux.c, config/winnt-c.c,
13081 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
13082 vtable-verify.c, vtable-verify.h: Use the standard form for the
13085 2014-01-02 Tobias Burnus <burnus@net-b.de>
13087 * gcc.c (process_command): Update copyright notice dates.
13088 * gcov-dump.c: Ditto.
13090 * doc/cpp.texi: Bump @copying's copyright year.
13091 * doc/cppinternals.texi: Ditto.
13092 * doc/gcc.texi: Ditto.
13093 * doc/gccint.texi: Ditto.
13094 * doc/gcov.texi: Ditto.
13095 * doc/install.texi: Ditto.
13096 * doc/invoke.texi: Ditto.
13098 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13100 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
13102 2014-01-01 Jakub Jelinek <jakub@redhat.com>
13104 * config/i386/sse.md (*mov<mode>_internal): Guard
13105 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
13107 PR rtl-optimization/59647
13108 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
13109 new_rtx into UNSIGNED_FLOAT rtxes.
13111 Copyright (C) 2014 Free Software Foundation, Inc.
13113 Copying and distribution of this file, with or without modification,
13114 are permitted in any medium without royalty provided the copyright
13115 notice and this notice are preserved.