1 2014-06-03 Richard Sandiford <rdsandiford@googlemail.com>
3 * defaults.h (USE_MD_CONSTRAINTS, EXTRA_MEMORY_CONSTRAINT)
4 (EXTRA_ADDRESS_CONSTRAINT, DEFAULT_CONSTRAINT_LEN, CONSTRAINT_LEN)
5 (CONST_OK_FOR_CONSTRAINT_P, CONST_DOUBLE_OK_FOR_LETTER_P)
6 (REG_CLASS_FROM_CONSTRAINT, EXTRA_CONSTRAINT_STR): Delete definitions
8 (REG_CLASS_FROM_LETTER, CONST_OK_FOR_LETTER_P)
9 (CONST_DOUBLE_OK_FOR_LETTER_P, EXTRA_CONSTRAINT): Move poising to...
10 * system.h: ...here and make it unconditional.
11 * target.def (conditional_register_usage): Mention
12 define_register_constraint instead of old-style constraint macros.
13 * doc/tm.texi.in: Remove documentation for old-style constraint macros.
14 * doc/tm.texi: Regenerate.
15 * genoutput.c: Remove USE_MD_CONSTRAINTS conditions and all code
16 protected by !USE_MD_CONSTRAINTS.
17 * config/frv/frv.md: Remove quote from old version of documentation.
18 * config/frv/frv.c (frv_conditional_register_usage): Likewise.
19 * config/m32r/m32r.c (easy_di_const, easy_df_const): Avoid mentioning
20 CONST_DOUBLE_OK_FOR_LETTER.
21 * config/sh/constraints.md: Likewise EXTRA_CONSTRAINT.
23 2014-06-02 Andrew Pinski <apinski@cavium.com>
25 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER):
26 /lib/ld-linux32-aarch64.so.1 is used for ILP32.
27 (LINUX_TARGET_LINK_SPEC): Update linker script for ILP32.
28 file whose name depends on -mabi= and -mbig-endian.
29 * config/aarch64/t-aarch64-linux (MULTILIB_OSDIRNAMES):
30 Handle LP64 better and handle ilp32 too.
31 (MULTILIB_OPTIONS): Delete.
32 (MULTILIB_DIRNAMES): Delete.
34 2014-06-02 Andrew MacLeod <amacleod@redhat.com>
36 * expr.h: Remove prototypes of functions defined in builtins.c.
37 * tree.h: (build_call_expr_*, build_string_literal): Add prototypes.
38 Remove prototypes of functions defined in builtins.c.
39 * builtins.h: Update prototype list to include all exported functions.
40 * builtins.c: (default_libc_has_function, gnu_libc_has_function,
41 no_c99_libc_has_function): Move to targhooks.c
42 (build_string_literal, build_call_expr_loc_array,
43 build_call_expr_loc_vec, build_call_expr_loc, build_call_expr): Move
45 (expand_builtin_object_size, fold_builtin_object_size): Make static.
46 * targhooks.c (default_libc_has_function, gnu_libc_has_function,
47 no_c99_libc_has_function): Relocate from builtins.c.
48 * tree.c: Include builtins.h.
49 (build_call_expr_loc_array, build_call_expr_loc_vec,
50 build_call_expr_loc, build_call_expr, build_string_literal): Relocate
52 * fold-const.h (fold_fma): Move prototype to builtins.h.
53 * realmpfr.h (do_mpc_arg2): Move prototype to builtins.h.
54 * asan.c: Include builtins.h.
55 * cfgexpand.c: Likewise.
56 * convert.c: Likewise.
57 * emit-rtl.c: Likewise.
60 * fold-const.c: Likewise.
61 * gimple-fold.c: Likewise.
62 * gimple-ssa-strength-reduction.c: Likewise.
63 * gimplify.c: Likewise.
64 * ipa-inline.c: Likewise.
65 * ipa-prop.c: Likewise.
66 * lto-streamer-out.c: Likewise.
68 * tree-inline.c: Likewise.
69 * tree-object-size.c: Likewise.
70 * tree-sra.c: Likewise.
71 * tree-ssa-ccp.c: Likewise.
72 * tree-ssa-forwprop.c: Likewise.
73 * tree-ssa-loop-ivcanon.c: Likewise.
74 * tree-ssa-loop-ivopts.c: Likewise.
75 * tree-ssa-math-opts.c: Likewise.
76 * tree-ssa-reassoc.c: Likewise.
77 * tree-ssa-threadedge.c: Likewise.
78 * tree-streamer-in.c: Likewise.
79 * tree-vect-data-refs.c: Likewise.
80 * tree-vect-patterns.c: Likewise.
81 * tree-vect-stmts.c: Likewise.
82 * config/aarch64/aarch64.c: Likewise.
83 * config/alpha/alpha.c: Likewise.
84 * config/arc/arc.c: Likewise.
85 * config/arm/arm.c: Likewise.
86 * config/avr/avr.c: Likewise.
87 * config/bfin/bfin.c: Likewise.
88 * config/c6x/c6x.c: Likewise.
89 * config/cr16/cr16.c: Likewise.
90 * config/cris/cris.c: Likewise.
91 * config/epiphany/epiphany.c: Likewise.
92 * config/fr30/fr30.c: Likewise.
93 * config/frv/frv.c: Likewise.
94 * config/h8300/h8300.c: Likewise.
95 * config/i386/i386.c: Likewise.
96 * config/i386/winnt.c: Likewise.
97 * config/ia64/ia64.c: Likewise.
98 * config/iq2000/iq2000.c: Likewise.
99 * config/lm32/lm32.c: Likewise.
100 * config/m32c/m32c.c: Likewise.
101 * config/m32r/m32r.c: Likewise.
102 * config/m68k/m68k.c: Likewise.
103 * config/mcore/mcore.c: Likewise.
104 * config/mep/mep.c: Likewise.
105 * config/microblaze/microblaze.c: Likewise.
106 * config/mips/mips.c: Likewise.
107 * config/mmix/mmix.c: Likewise.
108 * config/mn10300/mn10300.c: Likewise.
109 * config/moxie/moxie.c: Likewise.
110 * config/msp430/msp430.c: Likewise.
111 * config/nds32/nds32.c: Likewise.
112 * config/pa/pa.c: Likewise.
113 * config/pdp11/pdp11.c: Likewise.
114 * config/picochip/picochip.c: Likewise.
115 * config/rl78/rl78.c: Likewise.
116 * config/rs6000/rs6000.c: Likewise.
117 * config/rx/rx.c: Likewise.
118 * config/s390/s390.c: Likewise.
119 * config/score/score.c: Likewise.
120 * config/sh/sh.c: Likewise.
121 * config/sparc/sparc.c: Likewise.
122 * config/spu/spu.c: Likewise.
123 * config/stormy16/stormy16.c: Likewise.
124 * config/tilegx/tilegx.c: Likewise.
125 * config/tilepro/tilepro.c: Likewise.
126 * config/v850/v850.c: Likewise.
127 * config/vax/vax.c: Likewise.
128 * config/xtensa/xtensa.c: Likewise.
130 2014-06-02 Jeff Law <law@redhat.com>
132 PR rtl-optimization/61094
133 * ree.c (combine_reaching_defs): Do not reextend an insn if it
134 was marked as do_no_reextend. If a copy is needed to eliminate
135 an extension, then mark it as do_not_reextend.
137 2014-06-02 Marcus Shawcroft <marcus.shawcroft@arm.com>
139 * config/aarch64/aarch64.md (set_fpcr): Drop ISB after FPCR write.
141 2014-06-02 Richard Henderson <rth@redhat.com>
144 * config/alpha/alpha.c (print_operand_address): Allow symbolic
145 addresses inside asms. Use output_operand_lossage instead of
148 2014-06-02 Uros Bizjak <ubizjak@gmail.com>
151 * config/i386/i386.c (ix86_expand_vec_perm) [case V32QImode]: Use
152 GEN_INT (-128) instead of GEN_INT (128) to set MSB of QImode constant.
154 2014-06-02 Tom de Vries <tom@codesourcery.com>
156 * config/aarch64/aarch64.c (aarch64_float_const_representable_p): Handle
157 case that x has VOIDmode.
159 2014-06-02 Bernd Schmidt <bernds@codesourcery.com>
161 * varasm.c (copy_constant): Delete function.
162 (build_constant_desc): Don't call it.
164 2014-06-02 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
167 * config/arm/arm.h (TARGET_SUPPORTS_WIDE_INT): Define.
168 * config/arm/arm.md (mov64 splitter): Replace const_double_operand
169 with immediate_operand.
171 2014-06-02 Andreas Schwab <schwab@suse.de>
174 (ia64_first_cycle_multipass_dfa_lookahead_guard): Check
175 pending_data_specs first.
177 2014-06-02 Richard Biener <rguenther@suse.de>
179 PR tree-optimization/61378
180 * tree-ssa-sccvn.c (vn_reference_lookup_3): Initialize
183 2014-06-01 Uros Bizjak <ubizjak@gmail.com>
185 * config/i386/constraints.md (Bw): Rename from 'w'.
186 (Bz): Rename from 'z'.
187 * config/i386/i386.md: Change 'w' to 'Bw' and 'z' to 'Bz' globally.
189 2014-06-01 Kai Tietz <ktietz@redhat.com>
192 * config/i386/constrains.md (define_constrain): New 'Bs' constraint.
193 * config/i386/i386.md (sibcall_insn_operand): Use Bs
194 instead of m constraint.
196 2014-05-31 Andreas Schwab <schwab@linux-m68k.org>
198 * config/m68k/m68k.md (beq0_di, bne0_di): Make the "o" constraint
199 a separate alternative where the scratch operand 2 is marked as
202 2014-05-31 Kugan Vivekanandarajah <kuganv@linaro.org>
204 * config/arm/arm.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New define.
205 (arm_builtins) : Add ARM_BUILTIN_GET_FPSCR and ARM_BUILTIN_SET_FPSCR.
206 (bdesc_2arg) : Add description for builtins __builtins_arm_set_fpscr
207 and __builtins_arm_get_fpscr.
208 (arm_init_builtins) : Initialize builtins __builtins_arm_set_fpscr and
209 __builtins_arm_get_fpscr.
210 (arm_expand_builtin) : Expand builtins __builtins_arm_set_fpscr and
211 __builtins_arm_ldfpscr.
212 (arm_atomic_assign_expand_fenv): New function.
213 * config/arm/vfp.md (set_fpscr): New pattern.
214 (get_fpscr) : Likewise.
215 * config/arm/unspecs.md (unspecv): Add VUNSPEC_GET_FPSCR and
217 * doc/extend.texi (AARCH64 Built-in Functions) : Document
218 __builtins_arm_set_fpscr, __builtins_arm_get_fpscr.
220 2014-05-30 Jakub Jelinek <jakub@redhat.com>
222 * asan.c (report_error_func): Add SLOW_P argument, use
223 BUILT_IN_ASAN_*_N if set.
224 (build_check_stmt): Likewise.
225 (instrument_derefs): If T has insufficient alignment,
226 force same handling as for odd sizes.
228 * sanitizer.def (BUILT_IN_ASAN_REPORT_LOAD_N,
229 BUILT_IN_ASAN_REPORT_STORE_N): New.
230 * asan.c (struct asan_mem_ref): Change access_size type to
232 (asan_mem_ref_init, asan_mem_ref_new, get_mem_refs_of_builtin_call,
233 update_mem_ref_hash_table): Likewise.
234 (asan_mem_ref_hasher::hash): Hash in a HWI.
235 (report_error_func): Change size_in_bytes argument to HWI.
236 Use *_N builtins if size_in_bytes is larger than 16 or not power of
238 (build_shadow_mem_access): New function.
239 (build_check_stmt): Use it. Change size_in_bytes argument to HWI.
240 Handle size_in_bytes not power of two or larger than 16.
241 (instrument_derefs): Don't give up if size_in_bytes is not
242 power of two or is larger than 16.
244 2014-05-30 Kai Tietz <ktietz@redhat.com>
247 * config/i386/i386.c (x86_output_mi_thunk): Add memory case
248 for sibling-tail-calls.
249 * config/i386/i386.md (sibcall_insn_operand): Add memory-constrain
251 * config/i386/predicates.md (sibcall_memory_operand): New predicate.
252 (sibcall_insn_operand): Add check for sibcall_memory_operand.
254 2014-05-30 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
256 * config/avr/avr-mcus.def: Change ATA6289 ISA to AVR4
257 * config/avr/avr-tables.opt: Regenerate.
258 * config/avr/t-multilib: Regenerate.
259 * doc/avr-mmcu.texi: Regenerate.
261 2014-05-30 Ian Lance Taylor <iant@google.com>
263 * config/i386/xmmintrin.h (_mm_pause): Move out of scope of pragma
266 2014-05-30 Tom de Vries <tom@codesourcery.com>
268 * config/i386/i386.c (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS):
271 2014-05-30 Tom de Vries <tom@codesourcery.com>
273 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
274 * lra.c (initialize_lra_reg_info_element): Add init of
275 actual_call_used_reg_set field.
276 (lra): Call lra_create_live_ranges before lra_inheritance for
278 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
280 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
281 instead of call_used_reg_set for -fuse-caller-save.
282 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
284 2014-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
286 * config/arm/thumb2.md (*thumb2_movhi_insn): Set type of movw
288 * config/arm/vfp.md (*thumb2_movsi_vfp): Likewise.
290 2014-05-30 Richard Sandiford <rdsandiford@googlemail.com>
292 * ira.c (ira_get_dup_out_num): Check for output operands at
293 the start of the loop. Handle cases where an included alternative
294 follows an excluded one.
296 2014-05-29 Mike Stump <mikestump@comcast.net>
299 * collect2.c (maybe_run_lto_and_relink): Be sure to always run
300 post ld passes when lto is used.
302 2014-05-29 Vladimir Makarov <vmakarov@redhat.com>
304 PR rtl-optimization/61325
305 * lra-constraints.c (process_address): Rename to process_address_1.
306 (process_address): New function.
308 2014-05-29 Alan Lawrence <alan.lawrence@arm.com>
310 * config/aarch64/aarch64-builtins.c (aarch64_types_binopv_qualifiers,
311 TYPES_BINOPV): New static data.
312 * config/aarch64/aarch64-simd-builtins.def (im_lane_bound):
314 * config/aarch64/aarch64-simd.md (aarch64_ext,
315 aarch64_im_lane_boundsi): New patterns.
316 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_const_1): Match
318 (aarch64_evpc_ext): New function.
320 * config/aarch64/iterators.md (UNSPEC_EXT): New enum element.
322 * config/aarch64/arm_neon.h (vext_f32, vext_f64, vext_p8, vext_p16,
323 vext_s8, vext_s16, vext_s32, vext_s64, vext_u8, vext_u16, vext_u32,
324 vext_u64, vextq_f32, vextq_f64, vextq_p8, vextq_p16, vextq_s8,
325 vextq_s16, vextq_s32, vextq_s64, vextq_u8, vextq_u16, vextq_u32,
326 vextq_u64): Replace __asm with __builtin_shuffle and im_lane_boundsi.
328 2014-05-29 Tom de Vries <tom@codesourcery.com>
330 * rtl.h (BLOCK_SYMBOL_CHECK): Use SYMBOL_REF_FLAGS.
332 2014-05-29 Richard Earnshaw <rearnsha@arm.com>
333 Richard Sandiford <rdsandiford@googlemail.com>
335 * arm/iterators.md (shiftable_ops): New code iterator.
336 (t2_binop0, arith_shift_insn): New code attributes.
337 * arm/predicates.md (shift_nomul_operator): New predicate.
338 * arm/arm.md (insn_enabled): Delete.
339 (enabled): Remove insn_enabled test.
340 (*arith_shiftsi): Delete. Replace with ...
341 (*<arith_shift_insn>_multsi): ... new pattern.
342 (*<arith_shift_insn>_shiftsi): ... new pattern.
343 * config/arm/arm.c (arm_print_operand): Handle operand format 'b'.
345 2014-05-29 Radovan Obradovic <robradovic@mips.com>
346 Tom de Vries <tom@codesourcery.com>
348 * config/mips/mips.h (POST_CALL_TMP_REG): Define.
349 * config/mips/mips.c (mips_emit_call_insn): Add POST_CALL_TMP_REG
351 (mips_split_call): Use POST_CALL_TMP_REG.
352 (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): Redefine to true.
354 2014-05-29 Tom de Vries <tom@codesourcery.com>
356 * final.c (collect_fn_hard_reg_usage): Guard variable declaration
357 with #ifdef STACK_REGS.
359 2014-05-28 Jan Hubicka <hubicka@ucw.cz>
361 * varasm.c (get_variable_section): Walk aliases.
362 (place_block_symbol): Walk aliases.
364 2014-05-28 Tom de Vries <tom@codesourcery.com>
367 2014-05-28 Tom de Vries <tom@codesourcery.com>
369 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
370 * lra.c (initialize_lra_reg_info_element): Add init of
371 actual_call_used_reg_set field.
372 (lra): Call lra_create_live_ranges before lra_inheritance for
374 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
376 * lra-constraints.c (need_for_call_save_p): Use
377 actual_call_used_reg_set instead of call_used_reg_set for
379 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
381 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
383 * doc/md.texi: Document that the % constraint character must
384 be at the beginning of the string.
385 * genoutput.c (validate_insn_alternatives): Check that '=',
386 '+' and '%' only appear at the beginning of a constraint.
387 * ira.c (commutative_constraint_p): Delete.
388 (ira_get_dup_out_num): Expect the '%' commutativity marker to be
389 at the start of the string.
390 * config/alpha/alpha.md (*movmemdi_1, *clrmemdi_1): Remove
392 * config/arm/neon.md (bicdi3_neon): Likewise.
393 * config/iq2000/iq2000.md (addsi3_internal, subsi3_internal, sgt_si)
394 (slt_si, sltu_si): Likewise.
395 * config/vax/vax.md (sbcdi3): Likewise.
396 * config/h8300/h8300.md (*cmpstz): Remove duplicate '+'.
397 * config/arc/arc.md (mulsi_600, mulsidi_600, umulsidi_600)
398 (mul64): Move '%' to beginning of constraint.
399 * config/arm/arm.md (*xordi3_insn): Likewise.
400 * config/nds32/nds32.md (add<mode>3, mulsi3, andsi3, iorsi3)
403 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
405 * doc/md.texi: Document the restrictions on the "enabled" attribute.
407 2014-05-28 Jason Merrill <jason@redhat.com>
410 * cgraph.h (symtab_node::get_comdat_group_id): New.
411 * cgraphunit.c (analyze_functions): Call it.
412 * symtab.c (dump_symtab_node): Likewise.
413 * tree.c (decl_comdat_group_id): New.
414 * tree.h: Declare it.
415 * lto-streamer-out.c (write_symbol): Use it.
416 * trans-mem.c (ipa_tm_create_version_alias): Likewise.
418 2014-05-28 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
421 * wide-int.cc: Do not include longlong.h when compiling with clang.
423 2014-05-28 Richard Biener <rguenther@suse.de>
425 * tree-ssa-propagate.c (add_control_edge): Print less vertical space.
426 * tree-vrp.c (extract_range_from_ssa_name): Also copy VR_UNDEFINED.
427 (vrp_visit_assignment_or_call): Print less vertical space.
428 (vrp_visit_stmt): Likewise.
429 (vrp_visit_phi_node): Likewise. For a PHI argument with
430 VR_VARYING range consider recording it as copy.
432 2014-05-28 Richard Biener <rguenther@suse.de>
435 2014-05-28 Richard Biener <rguenther@suse.de>
437 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
439 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
441 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
442 sufficiently aligned and an offset is used at the same time.
443 (expand_expr_real_1): Likewise.
445 2014-05-28 Richard Biener <rguenther@suse.de>
448 * fold-const.c (fold_comparison): When folding
449 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
450 the sign of the remaining constant operand stays the same.
452 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
454 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
455 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
456 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
458 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES is true.
459 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
460 (m32bit-doubles) Likewise.
461 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
462 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
465 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
467 * configure.ac ($gcc_cv_ld_clearcap): New test.
468 * configure: Regenerate.
469 * config.in: Regenerate.
470 * config/sol2.opt (mclear-hwcap): New option.
471 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
472 * config/sol2-clearcap.map: Moved here from
473 testsuite/gcc.target/i386/clearcap.map.
474 * config/sol2-clearcapv2.map: Move here from
475 gcc.target/i386/clearcapv2.map.
476 * config/t-sol2 (install): Depend on install-clearcap-map.
477 (install-clearcap-map): New target.
478 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
481 2014-05-28 Richard Biener <rguenther@suse.de>
483 * hwint.h (*_HALF_WIDE_INT*): Move to ...
484 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
485 ... here and remove the rest.
486 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
488 2014-05-28 Richard Biener <rguenther@suse.de>
490 PR tree-optimization/61335
491 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
492 new range fails, drop to varying.
494 2014-05-28 Tom de Vries <tom@codesourcery.com>
496 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
497 * lra.c (initialize_lra_reg_info_element): Add init of
498 actual_call_used_reg_set field.
499 (lra): Call lra_create_live_ranges before lra_inheritance for
501 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
503 * lra-constraints.c (need_for_call_save_p): Use
504 actual_call_used_reg_set instead of call_used_reg_set for
506 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
508 2014-05-28 Radovan Obradovic <robradovic@mips.com>
509 Tom de Vries <tom@codesourcery.com>
511 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save
513 (@item -fuse-caller-save): New item.
515 2014-05-28 Radovan Obradovic <robradovic@mips.com>
516 Tom de Vries <tom@codesourcery.com>
518 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
519 OPT_fuse_caller_save.
521 2014-05-28 Radovan Obradovic <robradovic@mips.com>
522 Tom de Vries <tom@codesourcery.com>
524 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
525 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
526 get_call_reg_set_usage.
527 * resource.c (mark_set_resources, mark_target_live_regs): Use
528 get_call_reg_set_usage.
529 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
531 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
532 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
533 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
534 * ira-build.c (ira_create_allocno): Init
535 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
536 (create_cap_allocno, propagate_allocno_info)
537 (propagate_some_info_from_allocno)
538 (copy_info_to_removed_store_destinations): Handle
539 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
540 * ira-costs.c (ira_tune_allocno_costs): Use
541 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
543 2014-05-28 Radovan Obradovic <robradovic@mips.com>
544 Tom de Vries <tom@codesourcery.com>
546 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
547 and function_used_regs_valid fields.
548 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
549 find_all_hard_reg_sets.
550 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
551 (get_call_reg_set_usage): New function.
552 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
553 * regs.h (get_call_reg_set_usage): Declare.
555 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
558 * config/dbx.h (License): Add Runtime Library Exception.
559 * config/newlib-stdint.h (License): Same.
560 * config/rtems.h (License): Same
561 * config/initfini-array.h (License): Same
562 * config/v850/v850.h (License): Same.
563 * config/v850/v850-opts.h (License): Same
564 * config/v850/rtems.h (License): Same.
566 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
569 * doc/extend.texi (Local Labels): Note that label differences are
570 not supported for AVR.
572 2014-05-27 Andrew Pinski <apinski@cavium.com>
574 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
575 Use <w> for the register in assembly template.
576 (stack_protect_test): Use the mode of operands[0] for the result.
577 (stack_protect_test_<mode>): Use <w> for the register
578 in assembly template.
580 2014-05-27 DJ Delorie <dj@redhat.com>
582 * config/rx/rx.c (add_vector_labels): New.
583 (rx_output_function_prologue): Call it.
584 (rx_handle_func_attribute): Don't require empty arguments.
585 (rx_handle_vector_attribute): New.
586 (rx_attribute_table): Add "vector" attribute.
587 * doc/extend.texi (interrupt, vector): Document new/changed
588 RX-specific attributes.
590 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
592 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
594 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
595 predicate to detect a negative quotient.
597 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
599 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
600 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
601 Add X - Y CMP 0 to X CMP Y transformation.
602 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
604 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
606 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
609 2014-05-27 Steve Ellcey <sellcey@mips.com>
611 * config/mips/mips.c: Add include of cgraph.h.
613 2014-05-27 Richard Biener <rguenther@suse.de>
615 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
617 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
620 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
621 * config/arm/arm-cores.def (License): Same.
622 * config/arm/arm-opts.h (License): Same.
623 * config/arm/aout.h (License): Same.
624 * config/arm/bpabi.h (License): Same.
625 * config/arm/elf.h (License): Same.
626 * config/arm/linux-elf.h (License): Same.
627 * config/arm/linux-gas.h (License): Same.
628 * config/arm/netbsd-elf.h (License): Same.
629 * config/arm/uclinux-eabi.h (License): Same.
630 * config/arm/uclinux-elf.h (License): Same.
631 * config/arm/vxworks.h (License): Same.
633 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
635 * config/arm/neon.md (neon_bswap<mode>): New pattern.
636 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
637 (arm_init_neon_builtins): Handle NEON_BSWAP.
638 Define required type nodes.
639 (arm_expand_neon_builtin): Handle NEON_BSWAP.
640 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
641 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
642 * config/arm/iterators.md (VDQHSD): New mode iterator.
644 2014-05-27 Richard Biener <rguenther@suse.de>
646 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
647 Try using literal operands when comparing value-ranges failed.
649 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
651 * ira.c (commutative_operand): Adjust for change to recog_data.
652 [Missing from previous commit.]
654 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
656 * system.h (TEST_BIT): New macro.
657 * recog.h (alternative_mask): New type.
658 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
659 (recog_data_d): Replace alternative_enabled_p array with
660 enabled_alternatives.
661 (target_recog): New structure.
662 (default_target_recog, this_target_recog): Declare.
663 (get_enabled_alternatives, recog_init): Likewise.
664 * recog.c (default_target_recog, this_target_recog): New variables.
665 (get_enabled_alternatives): New function.
666 (extract_insn): Use it.
667 (recog_init): New function.
668 (preprocess_constraints, constrain_operands): Adjust for change to
670 * postreload.c (reload_cse_simplify_operands): Likewise.
671 * reload.c (find_reloads): Likewise.
672 * ira-costs.c (record_reg_classes): Likewise.
673 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
674 all alternatives after a disabled one would be skipped.
675 (ira_implicitly_set_insn_hard_regs): Likewise.
676 * ira.c (ira_setup_alts): Adjust for change to recog_data.
677 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
678 with enabled_alternatives.
679 * lra.c (free_insn_recog_data): Update accordingly.
680 (lra_update_insn_recog_data): Likewise.
681 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
682 * lra-constraints.c (process_alt_operands): Likewise. Handle
683 only_alternative as part of the enabled mask.
684 * target-globals.h (this_target_recog): Declare.
685 (target_globals): Add a recog field.
686 (restore_target_globals): Restore this_target_recog.
687 * target-globals.c: Include recog.h.
688 (default_target_globals): Initialize recog field.
689 (save_target_globals): Likewise.
690 * reginfo.c (reinit_regs): Call recog_init.
691 * toplev.c (backend_init_target): Likewise.
693 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
695 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
696 rather than any named insn's code.
698 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
701 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
702 * config/arm/arm-cores.def (License): Same.
704 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
706 * tree.h (decl_comdat_group): Declare.
707 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
708 * tree.c (decl_comdat_group): Here.
710 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
712 PR rtl-optimization/61222
713 * combine.c (simplify_shift_const_1): When moving a PLUS outside
714 the shift, truncate the PLUS operand to the result mode.
716 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
719 * config/i386/i386.c (ix86_rtx_costs)
720 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
723 2014-05-26 Martin Jambor <mjambor@suse.cz>
725 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
728 2014-05-26 Richard Biener <rguenther@suse.de>
730 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
731 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
732 Provide specializations.
733 (wi::int_traits <HOST_WIDE_INT>,
734 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
736 2014-05-26 Alan Modra <amodra@gmail.com>
739 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
740 params and return a bool. Remove dead code. Update comment.
741 Assert we have a const_int source. Remove bogus code from
742 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
743 handling of constants > 2G and reg_equal note, from..
744 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
745 return value. Update comment. If we can, use a new pseudo
746 for intermediate calculations.
747 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
749 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
750 call to rs6000_emit_set_const in splitter.
751 (movdi_internal64+2, +3): Likewise.
753 2014-05-26 Richard Biener <rguenther@suse.de>
755 * system.h: Define __STDC_FORMAT_MACROS before
756 including inttypes.h.
757 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
758 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
759 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
760 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
761 HOST_WIDEST_INT_C): Remove.
762 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
763 if C99 inttypes.h is not available.
764 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
765 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
766 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
767 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
768 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
769 (struct output_info): Likewise.
770 (print_statistics): Adjust.
771 (dump_bitmap_statistics): Likewise.
772 * bt-load.c (migrate_btr_defs): Print with PRId64.
773 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
774 (MAX_SAFE_MULTIPLIER): Adjust.
775 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
776 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
777 dump_cgraph_node): Likewise.
778 * final.c (dump_basic_block_info): Likewise.
779 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
780 * gcov.c (format_gcov): Likewise.
781 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
783 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
784 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
785 (inline_small_functions, dump_overall_stats, dump_inline_stats):
786 Use PRId64 for dumping.
787 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
788 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
789 (add_allocno_hard_regs): Adjust.
790 * loop-doloop.c (doloop_modify): Print using PRId64.
791 * loop-iv.c (inverse): Compute in uint64_t.
792 (determine_max_iter, iv_number_of_iterations): Likewise.
793 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
795 * lto-streamer-out.c (write_symbol): Use uint64_t.
796 * mcf.c (CAP_INFINITY): Use int64_t maximum.
797 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
798 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
799 * modulo-sched.c (const_iteration_count): Use int64_t.
800 (sms_schedule): Dump using PRId64.
801 * predict.c (dump_prediction): Likewise.
802 * pretty-print.h (pp_widest_integer): Remove.
803 * profile.c (get_working_sets, is_edge_inconsistent,
804 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
805 * tree-pretty-print.c (pp_double_int): Remove case handling
806 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
807 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
809 (pass_optimize_bswap::execute): Remove restriction on hosts.
810 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
811 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
812 * tree.c (widest_int_cst_value): Remove.
813 * tree.h (widest_int_cst_value): Likewise.
814 * value-prof.c (dump_histogram_value): Print using PRId64.
815 * gengtype.c (main): Also inject int64_t.
816 * ggc-page.c (struct max_alignment): Use int64_t.
817 * alloc-pool.c (struct allocation_object_def): Likewise.
818 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
820 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
821 * doc/tm.texi: Regenerated.
822 * gengtype-lex.l (IWORD): Handle [u]int64_t.
823 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
824 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
825 mmix_output_register_setting): Use [u]int64_t in prototypes.
826 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
827 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
828 mmix_output_octa, mmix_output_shifted_value): Adjust.
829 (mmix_intval): Adjust. Remove unreachable case.
830 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
832 2014-05-26 Richard Biener <rguenther@suse.de>
834 * configure.ac: Drop __int64 type check. Insist that we
835 found uint64_t and int64_t.
836 * hwint.h (HOST_BITS_PER___INT64): Remove.
837 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
838 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
839 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
840 (HOST_WIDEST_FAST_INT): Remove __int64 case.
841 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
842 for dst_q_src_df_rms_cdt.
843 * configure: Regenerate.
844 * config.in: Likewise.
846 2014-05-26 Michael Tautschnig <mt@debian.org>
849 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
850 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
852 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
854 PR rtl-optimization/61278
855 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
857 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
859 PR rtl-optimization/61220
860 Part of PR rtl-optimization/61225
861 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
862 insn; skip split_edge for a block with only one successor.
864 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
866 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
869 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
871 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
872 (update_vtable_references): New function.
873 (function_and_variable_visibility): Rewrite also vtable initializers.
874 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
876 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
878 * ggc.h (ggc_grow): New function.
879 * ggc-none.c (ggc_grow): New function.
880 * ggc-page.c (ggc_grow): Likewise.
882 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
884 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
885 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
886 comdat_can_be_unshared_p, cgraph_externally_visible_p,
887 varpool_externally_visible_p, can_replace_by_local_alias,
888 update_visibility_by_resolution_info, function_and_variable_visibility,
889 pass_data_ipa_function_and_variable_visibility,
890 make_pass_ipa_function_and_variable_visibility,
891 whole_program_function_and_variable_visibility,
892 pass_data_ipa_whole_program_visibility,
893 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
894 * cgraph.h (cgraph_local_node_p): Declare.
895 * ipa-visibility.c: New file.
896 * Makefile.in (OBJS): Add ipa-visiblity.o
898 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
900 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
901 that var decl is available.
903 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
905 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
907 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
908 (find_decls_types_r): Do not walk COMDAT_GROUP.
909 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
910 * varasm.c (make_decl_one_only): Use set_comdat_group;
911 create node if needed.
912 * ipa-inline-transform.c (save_inline_function_body): Update
913 way we decl->symtab mapping.
914 * symtab.c (symtab_hash, hash_node, eq_node
915 symtab_insert_node_to_hashtable): Remove.
916 (symtab_register_node): Update.
917 (symtab_unregister_node): Update.
918 (symtab_get_node): Reimplement as inline function.
919 (symtab_add_to_same_comdat_group): Update.
920 (symtab_dissolve_same_comdat_group_list): Update.
921 (dump_symtab_base): Update.
922 (verify_symtab_base): Update.
923 (symtab_make_decl_local): Update.
924 (fixup_same_cpp_alias_visibility): Update.
925 (symtab_nonoverwritable_alias): Update.
926 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
927 * ipa.c (update_visibility_by_resolution_info): UPdate.
928 * bb-reorder.c: Include cgraph.h
929 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
931 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
932 * cgraph.c (cgraph_get_create_node): Update.
933 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
935 (symtab_get_node): Make inline.
936 (symtab_insert_node_to_hashtable): Remove.
937 (symtab_can_be_discarded): Update.
938 (decl_comdat_group): New function.
939 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
941 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
943 (read_comdat_group): New function.
944 (input_node, input_varpool_node): Use it.
945 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
947 * mips.c (mips_start_unique_function): Likewise.
948 (ix86_code_end): Likewise.
949 (rs6000_code_end): Likweise.
950 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
952 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
954 * gengtype-state.c (fatal_reading_state): Bring offline.
955 * optabs.c (widening_optab_handler): Bring offline.
956 * optabs.h (widening_optab_handler): Likewise.
957 * final.c (get_attr_length_1): Likewise.
959 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
961 * sched-int.h (sd_iterator_cond): Manually tail recurse.
963 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
965 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
966 (ppc440-compare): Include shift with dot.
967 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
968 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
970 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
972 (e6500_sfx2): Include it.
973 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
974 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
975 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
976 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
977 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
978 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
979 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
980 *lshiftrt_internal1le, *lshiftrt_internal1be,
981 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
982 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
983 *rotldi3_internal10le, *rotldi3_internal10be,
984 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
985 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
986 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
987 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
988 define_insns): Use type "shift" in the appropriate alternatives.
990 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
992 * config/rs6000/rs6000.md (type): Add "logical". Delete
994 (dot): Adjust comment.
995 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
996 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
997 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
998 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
999 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
1000 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
1001 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
1002 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1004 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1005 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1006 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1007 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1008 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1009 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1010 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1011 * config/rs6000/8540.md (ppc8540_su): Adjust.
1012 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1013 cell-cmp-microcoded): Adjust.
1014 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1015 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1016 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1017 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1018 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1019 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1020 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1021 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1022 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1024 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1025 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
1026 Adjust. Adjust comment.
1027 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1028 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1030 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1032 * config/rs6000/rs6000.md (type): Add "add".
1033 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
1034 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
1035 define_insns): Use it.
1036 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1038 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1039 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1040 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1041 * config/rs6000/601.md (ppc601-integer): Adjust.
1042 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1043 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1044 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1045 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1046 * config/rs6000/8540.md (ppc8540_su): Adjust.
1047 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1048 cell-cmp-microcoded): Adjust.
1049 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1050 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1051 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1052 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1053 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1054 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1055 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1056 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1057 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1059 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1060 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
1061 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1062 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1064 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1066 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
1067 "delayed_compare", "var_delayed_compare".
1068 (var_shift): New attribute.
1069 (cell_micro): Adjust.
1070 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
1071 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
1072 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
1073 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
1074 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
1075 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
1076 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
1077 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
1078 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
1079 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
1080 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
1081 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
1082 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
1083 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
1084 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
1085 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
1086 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
1087 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
1088 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
1089 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
1090 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
1091 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
1092 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
1093 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1094 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1096 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1097 * config/rs6000/440.md (ppc440-integer): Adjust.
1098 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1099 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
1100 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1101 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1102 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1103 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1104 * config/rs6000/8540.md (ppc8540_su): Adjust.
1105 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1106 cell-cmp-microcoded): Adjust.
1107 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
1108 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1109 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
1110 e500mc64_delayed): Adjust.
1111 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
1112 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
1113 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1114 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
1115 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
1116 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
1117 power6-delayed-compare, power6-var-delayed-compare): Adjust.
1118 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
1119 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
1121 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1122 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1124 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1126 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
1127 (bits): New mode_attr.
1128 (idiv_ldiv): Delete mode_attr.
1129 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
1130 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1131 rs6000_adjust_priority, is_nonpipeline_insn,
1132 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1134 * config/rs6000/40x.md (ppc403-idiv): Adjust.
1135 * config/rs6000/440.md (ppc440-idiv): Adjust.
1136 * config/rs6000/476.md (ppc476-idiv): Adjust.
1137 * config/rs6000/601.md (ppc601-idiv): Adjust.
1138 * config/rs6000/603.md (ppc603-idiv): Adjust.
1139 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
1140 ppc620-ldiv): Adjust.
1141 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
1142 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
1143 * config/rs6000/8540.md (ppc8540_divide): Adjust.
1144 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
1145 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
1146 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
1147 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
1148 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
1149 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
1150 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
1151 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
1152 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
1153 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
1154 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
1155 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
1156 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
1157 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
1158 * config/rs6000/titan.md (titan_fxu_div): Adjust.
1160 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1162 * config/rs6000/rs6000.md (type): Delete "insert_word",
1163 "insert_dword". Add "insert".
1164 (size): Update comment.
1165 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1166 insn_must_be_first_in_group): Adjust.
1167 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
1168 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
1169 *insvsi_internal6, insvdi_internal): Adjust.
1171 * config/rs6000/40x.md (ppc403-integer): Adjust.
1172 * config/rs6000/440.md (ppc440-integer): Adjust.
1173 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
1174 * config/rs6000/601.md (ppc601-integer): Adjust.
1175 * config/rs6000/603.md (ppc603-integer): Adjust.
1176 * config/rs6000/6xx.md (ppc604-integer): Adjust.
1177 * config/rs6000/7450.md (ppc7450-integer): Adjust.
1178 * config/rs6000/7xx.md (ppc750-integer): Adjust.
1179 * config/rs6000/8540.md (ppc8540_su): Adjust.
1180 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
1181 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
1182 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1183 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
1184 * config/rs6000/e5500.md (e5500_sfx): Adjust.
1185 * config/rs6000/e6500.md (e6500_sfx): Adjust.
1186 * config/rs6000/mpc.md (mpccore-integer): Adjust.
1187 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
1188 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
1189 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
1190 * config/rs6000/power7.md (power7-integer): Adjust.
1191 * config/rs6000/power8.md (power8-1cyc): Adjust.
1192 * config/rs6000/rs64.md (rs64a-integer): Adjust.
1193 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1195 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1197 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
1198 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
1199 (size): New attribute.
1200 (dot): New attribute.
1201 (cell_micro): Adjust.
1202 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
1203 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
1204 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
1205 umuldi3_highpart): Adjust.
1206 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1207 rs6000_adjust_priority, is_nonpipeline_insn,
1208 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1210 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
1211 ppc405-imul3): Adjust.
1212 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
1213 * config/rs6000/476.md (ppc476-imul): Adjust.
1214 * config/rs6000/601.md (ppc601-imul): Adjust.
1215 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
1216 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
1217 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
1218 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
1219 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
1221 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
1222 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
1223 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
1225 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
1226 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
1227 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
1228 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
1229 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
1230 * config/rs6000/mpc.md (mpccore-imul): Adjust.
1231 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
1232 power4-lmul, power4-imul, power4-imul3): Adjust.
1233 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
1234 power5-lmul, power5-imul, power5-imul3): Adjust.
1235 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
1236 power6-lmul, power6-imul, power6-imul3): Adjust.
1237 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
1238 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
1240 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
1241 rs64a-lmul): Adjust.
1242 * config/rs6000/titan.md (titan_imul): Adjust.
1244 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1246 * config/rs6000/rs6000.md (type): Add new value "halfmul".
1247 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
1248 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
1249 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
1250 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
1251 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
1252 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
1253 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
1254 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
1255 * config/rs6000/titan.md: Delete nonsensical comment.
1256 (titan_imul): Add type imul3.
1257 (titan_mulhw): Remove type imul3; add type halfmul.
1259 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1261 * config/rs6000/rs6000.md (type): Reorder, reformat.
1263 2014-05-23 Martin Jambor <mjambor@suse.cz>
1265 PR tree-optimization/53787
1266 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
1267 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
1268 analysis_done, update all uses.
1269 * ipa-prop.c: Include domwalk.h
1270 (param_analysis_info): Removed.
1271 (param_aa_status): New type.
1272 (ipa_bb_info): Likewise.
1273 (func_body_info): Likewise.
1274 (ipa_get_bb_info): New function.
1275 (aa_overwalked): Likewise.
1276 (find_dominating_aa_status): Likewise.
1277 (parm_bb_aa_status_for_bb): Likewise.
1278 (parm_preserved_before_stmt_p): Changed to use new param AA info.
1279 (load_from_unmodified_param): Accept func_body_info as a parameter
1280 instead of parms_ainfo.
1281 (parm_ref_data_preserved_p): Changed to use new param AA info.
1282 (parm_ref_data_pass_through_p): Likewise.
1283 (ipa_load_from_parm_agg_1): Likewise. Update callers.
1284 (compute_complex_assign_jump_func): Changed to use new param AA info.
1285 (compute_complex_ancestor_jump_func): Likewise.
1286 (ipa_compute_jump_functions_for_edge): Likewise.
1287 (ipa_compute_jump_functions): Removed.
1288 (ipa_compute_jump_functions_for_bb): New function.
1289 (ipa_analyze_indirect_call_uses): Likewise, moved variable
1291 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
1292 and info, moved variable declarations down.
1293 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
1295 (ipa_analyze_stmt_uses): Likewise.
1296 (ipa_analyze_params_uses): Removed.
1297 (ipa_analyze_params_uses_in_bb): New function.
1298 (ipa_analyze_controlled_uses): Likewise.
1299 (free_ipa_bb_info): Likewise.
1300 (analysis_dom_walker): New class.
1301 (ipa_analyze_node): Handle node-specific forbidden analysis,
1302 initialize and free func_body_info, use dominator walker.
1303 (ipcp_modif_dom_walker): New class.
1304 (ipcp_transform_function): Create and free func_body_info, use
1305 ipcp_modif_dom_walker, moved a lot of functionality there.
1307 2014-05-23 Marek Polacek <polacek@redhat.com>
1308 Jakub Jelinek <jakub@redhat.com>
1310 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
1311 * gcc.c (sanitize_spec_function): Likewise.
1312 * convert.c (convert_to_integer): Include "ubsan.h". Add
1313 floating-point to integer instrumentation.
1314 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
1315 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
1316 SANITIZE_NONDEFAULT.
1317 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
1318 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
1319 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
1320 * ubsan.c: Include "realmpfr.h" and "dfp.h".
1321 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
1322 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
1323 float/double/long double.
1324 (ubsan_instrument_float_cast): New function.
1325 * ubsan.h (ubsan_instrument_float_cast): Declare.
1327 2014-05-23 Jiong Wang <jiong.wang@arm.com>
1329 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
1331 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
1332 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
1333 Adjust for tailcalling through registers.
1334 * config/aarch64/aarch64.h (enum reg_class): New caller save
1336 (REG_CLASS_NAMES): Likewise.
1337 (REG_CLASS_CONTENTS): Likewise.
1338 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
1339 Allow tailcalling without decls.
1341 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1343 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1344 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
1346 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
1347 gsi, and variables v_* to v*.
1349 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
1351 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
1353 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1355 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
1356 * omp-low.c: Update accordingly.
1358 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
1359 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
1360 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
1361 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
1362 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
1363 GF_OMP_TARGET_KIND_UPDATE.
1365 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1366 Explicitly enumerate the expected region types.
1368 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
1371 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
1372 documentation; the old documentation didn't clearly state the
1373 constraints on the contents of the pointed-to storage.
1375 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1377 Fix bootstrap error on ia64
1378 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
1379 Return default value.
1381 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
1383 PR tree-optimization/54733
1384 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
1386 (find_bswap_or_nop_load): New.
1387 (find_bswap_1): Renamed to ...
1388 (find_bswap_or_nop_1): This. Also add support for memory source.
1389 (find_bswap): Renamed to ...
1390 (find_bswap_or_nop): This. Also add support for memory source and
1391 detection of bitwise operations equivalent to load in host endianness.
1392 (execute_optimize_bswap): Likewise. Also move its leading comment back
1393 in place and split statement transformation into ...
1394 (bswap_replace): This.
1396 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1398 PR rtl-optimization/61215
1399 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
1400 simplify_gen_subreg until final substitution.
1402 2014-05-23 Alan Modra <amodra@gmail.com>
1405 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
1406 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
1407 Use "Y" constraint rather than "m".
1409 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1411 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
1413 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
1414 New function declaration.
1415 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
1416 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
1417 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
1418 (aarch64_init_builtins) : Initialize builtins
1419 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1420 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1421 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
1422 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
1423 and __builtins_aarch64_set_fpsr.
1424 (aarch64_atomic_assign_expand_fenv): New function.
1425 * config/aarch64/aarch64.md (set_fpcr): New pattern.
1426 (get_fpcr) : Likewise.
1427 (set_fpsr) : Likewise.
1428 (get_fpsr) : Likewise.
1429 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
1430 and UNSPECV_SET_FPSR.
1431 * doc/extend.texi (AARCH64 Built-in Functions) : Document
1432 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1433 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1435 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1437 PR rtl-optimization/60969
1438 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
1439 constraints. Set up mem cost for NO_REGS case.
1441 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
1443 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
1445 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
1447 * config/darwin.c: Include "lto-section-names.h".
1448 (LTO_SEGMENT_NAME): Don't define.
1449 * config/i386/winnt.c: Include "lto-section-names.h".
1450 * lto-streamer.c: Include "lto-section-names.h".
1451 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
1452 * lto-wrapper.c: Include "lto-section-names.h".
1453 (LTO_SECTION_NAME_PREFIX): Don't define.
1454 * lto-section-names.h: New file.
1455 * cgraphunit.c: Include "lto-section-names.h".
1457 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
1459 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
1461 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
1464 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
1466 2014-05-22 Nick Clifton <nickc@redhat.com>
1468 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
1470 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
1472 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
1473 -> (T)A transformation to integer types.
1475 2014-05-22 Teresa Johnson <tejohnson@google.com>
1477 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
1478 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
1479 (gcov_rewrite): Use gcov_nonruntime_assert.
1481 (gcov_write_words): Ditto.
1482 (gcov_write_length): Ditto.
1483 (gcov_read_words): Use gcov_nonruntime_assert, and remove
1484 gcc_assert from IN_LIBGCOV code.
1485 (gcov_read_summary): Use gcov_error to flag profile corruption.
1486 (gcov_sync): Use gcov_nonruntime_assert.
1487 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
1488 (gcov_histo_index): Use gcov_nonruntime_assert.
1489 (static void gcov_histogram_merge): Ditto.
1490 (compute_working_sets): Ditto.
1491 * gcov-io.h (gcov_nonruntime_assert): Define.
1492 (gcov_error): Define for !IN_LIBGCOV
1494 2014-05-22 Richard Biener <rguenther@suse.de>
1496 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1497 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
1498 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
1499 and deallocation site.
1500 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1501 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
1502 passing through the incoming points-to set.
1503 (handle_lhs_call): Use flags argument instead of recomputing it.
1504 (find_func_aliases_for_call): Call handle_lhs_call with proper
1507 2014-05-22 Jakub Jelinek <jakub@redhat.com>
1509 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
1510 all padding bits in REAL_VALUE_TYPE are cleared.
1512 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1514 Cleanup and improve multipass_dfa_lookahead_guard
1515 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
1516 (core2i7_first_cycle_multipass_begin,)
1517 (core2i7_first_cycle_multipass_issue,)
1518 (core2i7_first_cycle_multipass_backtrack): Update signature.
1519 * config/ia64/ia64.c
1520 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
1521 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
1522 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
1524 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
1525 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
1527 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
1529 * doc/tm.texi: Regenerate.
1531 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
1532 * haifa-sched.c (ready_try): Make signed to allow negative values.
1533 (rebug_ready_list_1): Update.
1534 (choose_ready): Simplify.
1535 (sched_extend_ready_list): Update.
1537 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1539 Remove IA64 speculation tweaking flags
1540 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
1541 speculation tuning flags.
1542 (msched-prefer-non-data-spec-insns,)
1543 (msched-prefer-non-control-spec-insns): Obsolete options.
1544 * haifa-sched.c (choose_ready): Remove handling of
1545 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1546 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
1547 and PREFER_NON_DATA_SPEC.
1548 * sel-sched.c (process_spec_exprs): Remove handling of
1549 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1551 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1553 Improve scheduling debug output
1554 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
1555 (advance_one_cycle): Update.
1556 (schedule_insn, queue_to_ready): Add debug printouts.
1557 (debug_ready_list_1): New static function.
1558 (debug_ready_list): Update.
1559 (max_issue): Add debug printouts.
1560 (dump_insn_stream): New static function.
1561 (schedule_block): Use it. Also better indent printouts.
1563 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1565 Fix sched_insn debug counter
1566 * haifa-sched.c (schedule_insn): Update.
1567 (struct haifa_saved_data): Add nonscheduled_insns_begin.
1568 (save_backtrack_point, restore_backtrack_point): Update.
1569 (first_nonscheduled_insn): New static function.
1570 (queue_to_ready, choose_ready): Use it.
1571 (schedule_block): Init nonscheduled_insns_begin.
1572 (sched_emit_insn): Update.
1575 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
1577 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
1579 (aarch64_secondary_reload) : LikeWise.
1580 (aarch64_class_max_nregs) : Remove CORE_REGS.
1581 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
1582 (REG_CLASS_NAMES) : Likewise.
1583 (REG_CLASS_CONTENTS) : LikeWise.
1584 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
1586 2014-05-21 Guozhi Wei <carrot@google.com>
1589 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
1591 (vqdmulhq_n_s16): Likewise.
1593 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
1595 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
1597 2014-05-21 Marek Polacek <polacek@redhat.com>
1600 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
1602 2014-05-21 Martin Jambor <mjambor@suse.cz>
1604 * doc/invoke.texi (Optimize Options): Document parameters
1605 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
1606 ipa-cp-array-index-hint-bonus.
1608 2014-05-21 Mark Wielaard <mjw@redhat.com>
1611 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
1612 version >= 3 or not strict DWARF.
1613 * langhooks.h (struct lang_hooks_for_types): Add
1614 enum_underlying_base_type.
1615 * langhooks.c (lhd_enum_underlying_base_type): New function.
1616 * gcc/langhooks.h (struct lang_hooks_for_types): Add
1617 enum_underlying_base_type.
1618 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
1619 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
1620 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
1622 2014-05-21 Richard Biener <rguenther@suse.de>
1624 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
1626 2014-05-21 John Marino <gnugcc@marino.st>
1628 * config.gcc (*-*-dragonfly*): New target.
1629 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
1630 * configure: Regenerate.
1631 * config/dragonfly-stdint.h: New.
1632 * config/dragonfly.h: New.
1633 * config/dragonfly.opt: New.
1634 * config/i386/dragonfly.h: New.
1635 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
1637 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1639 * tree.def (VOID_CST): New.
1640 * tree-core.h (TI_VOID): New.
1641 * tree.h (void_node): New.
1642 * tree.c (tree_node_structure_for_code, tree_code_size)
1643 (iterative_hash_expr): Handle VOID_CST.
1644 (build_common_tree_nodes): Initialize void_node.
1646 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
1648 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
1650 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
1652 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
1655 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
1656 flag_reorder_blocks_and_partition.
1657 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
1659 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
1662 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
1664 (*addc_r_t): Add new insn_and_split.
1666 2014-05-21 Jakub Jelinek <jakub@redhat.com>
1669 * omp-low.c (handle_simd_reference): New function.
1670 (lower_rec_input_clauses): Use it. Defer adding reference
1671 initialization even for reduction without placeholder if in simd,
1672 handle it properly later on.
1674 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1676 PR tree-optimization/60899
1677 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
1678 assume all static symbols will have definition wile parsing and
1679 check the do have definition later in compilation; check that
1680 variable referring symbol will be output before concluding that
1681 reference is safe; be conservative for referring local statics;
1682 be more precise about when comdat is output in other partition.
1684 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1687 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
1689 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
1690 (ipa_inline): Loop inline_to_all_callers until no more aliases
1693 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1695 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
1696 set writeonly flag only for vars actually written to.
1698 2014-05-20 Dehao Chen <dehao@google.com>
1700 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
1701 and callee count to get clone count.
1702 * tree-inline.c (expand_call_inline): Use callee count instead of bb
1705 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
1707 PR rtl-optimization/61243
1708 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
1710 2014-05-20 Xinliang David Li <davidxl@google.com>
1712 * cgraphunit.c (walk_polymorphic_call_targets): Add
1713 dbgcnt and fopt-info support.
1714 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1715 * ipa-devirt.c (ipa_devirt): Ditto.
1716 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1717 * ipa.c (walk_polymorphic_call_targets): Ditto.
1718 * gimple-fold.c (fold_gimple_assign): Ditto.
1719 (gimple_fold_call): Ditto.
1720 * dbgcnt.def: New counter.
1722 2014-05-20 DJ Delorie <dj@redhat.com>
1724 * config/msp430/msp430.md (split): Don't allow subregs when
1725 splitting SImode adds.
1726 (andneghi): Fix subtraction logic.
1727 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1729 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1731 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1733 * except.c (switch_to_exception_section, resolve_unique_section,
1734 get_named_text_section, default_function_rodata_section,
1735 align_variable, get_block_for_decl, default_section_type_flags):
1736 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1737 * symtab.c (symtab_add_to_same_comdat_group,
1738 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1739 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1741 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1742 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1743 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1744 (c6x_function_in_section_p): Likewise.
1745 * config/darwin.c (machopic_select_section): Likewise.
1746 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1747 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1748 * config/mep/mep.c (mep_select_section): LIkewise.
1749 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1751 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1753 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1754 EH region of calls to pure functions that can throw an exception.
1755 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1756 (copy_reference_ops_from_call): Also copy the EH region of the call if
1757 it can throw an exception.
1759 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1761 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1762 nested VEC_SELECTs that are inverses of each other.
1764 2014-05-20 Richard Biener <rguenther@suse.de>
1766 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1767 (extract_and_process_scc_for_name): not here.
1768 (cond_dom_walker::before_dom_children): Only process
1769 stmts that end the BB in interesting ways.
1770 (run_scc_vn): Mark param uses as visited.
1772 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1774 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1777 2014-05-20 Nick Clifton <nickc@redhat.com>
1779 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1780 (msp430_gimplify_va_arg_expr): New function.
1781 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
1783 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
1784 operand 0 in order to prevent confusion about the number of
1787 2014-05-20 Richard Biener <rguenther@suse.de>
1789 PR tree-optimization/61221
1790 * tree-ssa-pre.c (el_to_update): Remove.
1791 (eliminate_dom_walker::before_dom_children): Handle released
1792 VDEFs by value-numbering them to the associated VUSE. Update
1793 stmt immediately for substituted call address.
1794 (eliminate): Remove delayed stmt updating code.
1795 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
1796 possibly late re-numbered vuses.
1797 (vn_reference_lookup_2): Adjust.
1798 (vn_reference_lookup_pieces): Likewise.
1799 (vn_reference_lookup): Likewise.
1801 2014-05-20 Richard Biener <rguenther@suse.de>
1803 * config.gcc: Remove need_64bit_hwint.
1804 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
1805 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
1807 * config.in: Regenerate.
1808 * configure: Likewise.
1810 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
1812 * doc/extend.texi: Create Label Attributes section,
1813 move all label attributes into it and reference it.
1815 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
1817 * arm.c (thumb1_reorg): When scanning backwards skip anything
1818 that's not a proper insn.
1820 2014-05-19 Richard Biener <rguenther@suse.de>
1822 PR tree-optimization/61221
1823 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
1824 Do nothing for unreachable blocks.
1825 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
1826 Improve unreachability detection.
1828 2014-05-19 Richard Biener <rguenther@suse.de>
1830 PR tree-optimization/61209
1831 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
1833 2014-05-19 Nick Clifton <nickc@redhat.com>
1835 * except.c (init_eh): Fix computation of builtin setjmp buffer
1836 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
1838 2014-05-19 Richard Biener <rguenther@suse.de>
1840 PR tree-optimization/61184
1841 * tree-vrp.c (is_negative_overflow_infinity): Use
1842 TREE_OVERFLOW_P and do that check first.
1843 (is_positive_overflow_infinity): Likewise.
1844 (is_overflow_infinity): Likewise.
1845 (vrp_operand_equal_p): Properly treat operands with
1846 differing overflow as not equal.
1848 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
1850 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
1851 shift simplification where it was intended.
1853 2014-05-19 Christian Bruel <christian.bruel@st.com>
1856 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
1858 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
1861 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
1864 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1866 * reg-notes.def (CROSSING_JUMP): Likewise.
1867 * rtl.h (rtx_def): Update comment for jump flag.
1868 (CROSSING_JUMP_P): Define.
1869 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
1870 of a REG_CROSSING_JUMP note.
1871 * cfghooks.c (tidy_fallthru_edges): Likewise.
1872 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
1873 * emit-rtl.c (try_split): Likewise.
1874 * haifa-sched.c (sched_create_recovery_edges): Likewise.
1875 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
1876 * jump.c (redirect_jump_2): Likewise.
1877 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
1878 (relax_delay_slots): Likewise.
1879 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
1880 (bbit_di): Likewise.
1881 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
1882 * config/sh/sh.md (jump_compact): Likewise.
1883 * bb-reorder.c (rotate_loop): Likewise.
1884 (pass_duplicate_computed_gotos::execute): Likewise.
1885 (add_reg_crossing_jump_notes): Rename to...
1886 (update_crossing_jump_flags): ...this.
1887 (pass_partition_blocks::execute): Update accordingly.
1889 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1891 * tree.h: Remove extraneous template <>.
1893 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1895 * ipa.c (symtab_remove_unreachable_nodes): Remove
1896 symbol from comdat group if its body was eliminated.
1897 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
1898 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
1899 (symtab_unregister_node): ... this one.
1900 (verify_symtab_base): More strict checking of comdats.
1901 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
1903 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1905 * tree-pass.h (make_pass_ipa_comdats): New pass.
1906 * timevar.def (TV_IPA_COMDATS): New timevar.
1907 * passes.def (pass_ipa_comdats): Add.
1908 * Makefile.in (OBJS): Add ipa-comdats.o
1909 * ipa-comdats.c: New file.
1911 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1913 * ipa.c (update_visibility_by_resolution_info): New function.
1914 (function_and_variable_visibility): Use it.
1916 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1918 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
1920 (FOR_EACH_DEFINED_SYMBOL): New macro.
1921 (varpool_first_static_initializer, varpool_next_static_initializer,
1922 varpool_first_defined_variable, varpool_next_defined_variable):
1924 (symtab_in_same_comdat_p): Correctly deal with inline functions.
1926 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1928 * ggc-page.c (ggc_handle_finalizers): Add comment.
1930 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1932 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
1933 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
1934 (ggc_internal_cleared_alloc): Likewise.
1935 * ggc-page.c (finalizer): New class.
1936 (vec_finalizer): Likewise.
1937 (globals::finalizers): New member.
1938 (globals::vec_finalizers): Likewise.
1939 (ggc_internal_alloc): Record the finalizer if any for the block being
1941 (ggc_handle_finalizers): New function.
1942 (ggc_collect): Call ggc_handle_finalizers.
1943 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
1945 (ggc_internal_cleared_alloc): Likewise.
1946 (finalize): New function.
1947 (need_finalization_p): Likewise.
1948 (ggc_alloc): Install the type's destructor as the finalizer if it
1950 (ggc_cleared_alloc): Likewise.
1951 (ggc_vec_alloc): Likewise.
1952 (ggc_cleared_vec_alloc): Likewise.
1954 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1956 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
1958 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1960 * alias.c (record_alias_subset): Adjust.
1961 * bitmap.c (bitmap_element_allocate): Likewise.
1962 (bitmap_gc_alloc_stat): Likewise.
1963 * cfg.c (init_flow): Likewise.
1964 (alloc_block): Likewise.
1965 (unchecked_make_edge): Likewise.
1966 * cfgloop.c (alloc_loop): Likewise.
1967 (flow_loops_find): Likewise.
1968 (rescan_loop_exit): Likewise.
1969 * cfgrtl.c (init_rtl_bb_info): Likewise.
1970 * cgraph.c (insert_new_cgraph_node_version): Likewise.
1971 (cgraph_allocate_node): Likewise.
1972 (cgraph_create_edge_1): Likewise.
1973 (cgraph_allocate_init_indirect_info): Likewise.
1974 * cgraphclones.c (cgraph_clone_edge): Likewise.
1975 * cgraphunit.c (add_asm_node): Likewise.
1976 (init_lowered_empty_function): Likewise.
1977 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
1978 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
1979 (alpha_use_linkage): Likewise.
1980 * config/arc/arc.c (arc_init_machine_status): Likewise.
1981 * config/arm/arm.c (arm_init_machine_status): Likewise.
1982 * config/avr/avr.c (avr_init_machine_status): Likewise.
1983 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
1984 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
1985 * config/cris/cris.c (cris_init_machine_status): Likewise.
1986 * config/darwin.c (machopic_indirection_name): Likewise.
1987 (darwin_build_constant_cfstring): Likewise.
1988 (darwin_enter_string_into_cfstring_table): Likewise.
1989 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
1990 * config/frv/frv.c (frv_init_machine_status): Likewise.
1991 * config/i386/i386.c (get_dllimport_decl): Likewise.
1992 (ix86_init_machine_status): Likewise.
1993 (assign_386_stack_local): Likewise.
1994 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
1995 (i386_pe_maybe_record_exported_symbol): Likewise.
1996 (i386_pe_record_stub): Likewise.
1997 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
1998 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
1999 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
2000 (m32c_note_pragma_address): Likewise.
2001 * config/mep/mep.c (mep_init_machine_status): Likewise.
2002 (mep_note_pragma_flag): Likewise.
2003 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
2004 (mips16_local_alias): Likewise.
2005 (mips_init_machine_status): Likewise.
2006 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
2007 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
2008 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
2009 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
2010 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
2011 * config/pa/pa.c (pa_init_machine_status): Likewise.
2012 (pa_get_deferred_plabel): Likewise.
2013 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
2014 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
2015 (rs6000_init_machine_status): Likewise.
2016 (output_toc): Likewise.
2017 * config/s390/s390.c (s390_init_machine_status): Likewise.
2018 * config/score/score.c (score_output_external): Likewise.
2019 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
2020 * config/spu/spu.c (spu_init_machine_status): Likewise.
2021 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
2022 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
2023 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
2024 * coverage.c (coverage_end_function): Likewise.
2025 * dbxout.c (dbxout_init): Likewise.
2026 * doc/gty.texi: Don't mention variable_size attribute.
2027 * dwarf2cfi.c (new_cfi): Adjust.
2028 (new_cfi_row): Likewise.
2029 (copy_cfi_row): Likewise.
2030 (create_cie_data): Likewise.
2031 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
2032 (new_loc_descr): Likewise.
2033 (find_AT_string_in_table): Likewise.
2034 (add_addr_table_entry): Likewise.
2035 (new_die): Likewise.
2036 (add_var_loc_to_decl): Likewise.
2037 (clone_die): Likewise.
2038 (clone_as_declaration): Likewise.
2039 (break_out_comdat_types): Likewise.
2040 (new_loc_list): Likewise.
2041 (add_loc_descr_to_each): Likewise.
2042 (add_location_or_const_value_attribute): Likewise.
2043 (add_linkage_name): Likewise.
2044 (lookup_filename): Likewise.
2045 (dwarf2out_var_location): Likewise.
2046 (new_line_info_table): Likewise.
2047 (dwarf2out_init): Likewise.
2048 (mem_loc_descriptor): Likewise.
2049 (loc_descriptor): Likewise.
2050 (add_const_value_attribute): Likewise.
2051 (tree_add_const_value_attribute): Likewise.
2052 (comp_dir_string): Likewise.
2053 (dwarf2out_vms_debug_main_pointer): Likewise.
2054 (string_cst_pool_decl): Likewise.
2055 * emit-rtl.c (set_mem_attrs): Likewise.
2056 (get_reg_attrs): Likewise.
2057 (start_sequence): Likewise.
2058 (init_emit): Likewise.
2059 (init_emit_regs): Likewise.
2060 * except.c (init_eh_for_function): Likewise.
2061 (gen_eh_region): Likewise.
2062 (gen_eh_region_catch): Likewise.
2063 (gen_eh_landing_pad): Likewise.
2064 (add_call_site): Likewise.
2065 * function.c (add_frame_space): Likewise.
2066 (insert_temp_slot_address): Likewise.
2067 (assign_stack_temp_for_type): Likewise.
2068 (get_hard_reg_initial_val): Likewise.
2069 (allocate_struct_function): Likewise.
2070 (prepare_function_start): Likewise.
2071 (types_used_by_var_decl_insert): Likewise.
2072 * gengtype.c (variable_size_p): Remove function.
2073 (enum alloc_quantity): Remove enum.
2074 (write_typed_alloc_def): Remove function.
2075 (write_typed_struct_alloc_def): Likewise.
2076 (write_typed_typedef_alloc_def): Likewise.
2077 (write_typed_alloc_defns): Likewise.
2079 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
2080 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
2081 * ggc.h (ggc_alloc): new function.
2082 (ggc_cleared_alloc): Likewise.
2083 (ggc_vec_alloc): Template on type of vector element, and remove
2084 element size argument.
2085 (ggc_cleared_vec_alloc): Likewise.
2086 * gimple.c (gimple_build_omp_for): Adjust.
2087 (gimple_copy): Likewise.
2088 * ipa-cp.c (get_replacement_map): Likewise.
2089 (find_aggregate_values_for_callers_subset): Likewise.
2090 (known_aggs_to_agg_replacement_list): Likewise.
2091 * ipa-devirt.c (get_odr_type): Likewise.
2092 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
2093 (read_agg_replacement_chain): Likewise.
2094 * loop-iv.c (get_simple_loop_desc): Likewise.
2095 * lto-cgraph.c (input_node_opt_summary): Likewise.
2096 * lto-section-in.c (lto_new_in_decl_state): Likewise.
2097 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
2098 (input_eh_region): Likewise.
2099 (input_eh_lp): Likewise.
2100 (input_cfg): Likewise.
2101 * optabs.c (set_optab_libfunc): Likewise.
2102 (init_tree_optimization_optabs): Likewise.
2103 (set_conv_libfunc): Likewise.
2104 * passes.c (do_per_function_toporder): Likewise.
2105 * rtl.h: Don't use variable_size gty attribute.
2106 * sese.c (if_region_set_false_region): Adjust.
2107 * stringpool.c (gt_pch_save_stringpool): Likewise.
2108 * target-globals.c (save_target_globals): Likewise.
2109 * toplev.c (general_init): Likewise.
2110 * trans-mem.c (record_tm_replacement): Likewise.
2111 (split_bb_make_tm_edge): Likewise.
2112 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2113 * tree-data-ref.h (lambda_vector_new): Likewise.
2114 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
2115 * tree-iterator.c (tsi_link_before): Likewise.
2116 (tsi_link_after): Likewise.
2117 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
2118 * tree-ssa-loop-niter.c (record_estimate): Likewise.
2119 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
2120 * tree-ssa-operands.h: Don't use variable_size gty attribute.
2121 * tree-ssa.c (init_tree_ssa): Adjust.
2122 * tree-ssanames.c (set_range_info): Likewise.
2123 (get_ptr_info): Likewise.
2124 (duplicate_ssa_name_ptr_info): Likewise.
2125 (duplicate_ssa_name_range_info): Likewise.
2126 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
2127 (unpack_ts_fixed_cst_value_fields): Likewise.
2128 * tree.c (build_fixed): Likewise.
2129 (build_real): Likewise.
2130 (build_string): Likewise.
2131 (decl_priority_info): Likewise.
2132 (decl_debug_expr_insert): Likewise.
2133 (decl_value_expr_insert): Likewise.
2134 (decl_debug_args_insert): Likewise.
2135 (type_hash_add): Likewise.
2136 (build_omp_clause): Likewise.
2137 * ubsan.c (decl_for_type_insert): Likewise.
2138 * varasm.c (get_unnamed_section): Likewise.
2139 (get_noswitch_section): Likewise.
2140 (get_section): Likewise.
2141 (get_block_for_section): Likewise.
2142 (create_block_symbol): Likewise.
2143 (build_constant_desc): Likewise.
2144 (create_constant_pool): Likewise.
2145 (force_const_mem): Likewise.
2146 (record_tm_clone_pair): Likewise.
2147 * varpool.c (varpool_create_empty_node): Likewise.
2149 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2151 * dwarf2out.c (tree_add_const_value_attribute): Call
2152 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
2153 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
2154 instead of ggc_internal_<x>alloc_stat.
2155 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
2156 (ggc_realloc): Likewise.
2157 * ggc-none.c (ggc_internal_alloc): Likewise.
2158 (ggc_internal_cleared_alloc): Likewise.
2159 * ggc-page.c: Likewise.
2160 * ggc.h (ggc_internal_alloc_stat): Likewise.
2161 (ggc_internal_alloc): Remove macro.
2162 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
2163 (ggc_internal_cleared_alloc): Remove macro.
2164 (GGC_RESIZEVEC): Adjust.
2165 (ggc_resizevar): Remove macro.
2166 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
2167 (ggc_internal_cleared_vec_alloc_stat): Likewise.
2168 (ggc_internal_vec_cleared_alloc): Remove macro.
2169 (ggc_alloc_atomic_stat): Drop _stat suffix.
2170 (ggc_alloc_atomic): Remove macro.
2171 (ggc_alloc_cleared_atomic): Remove macro.
2172 (ggc_alloc_string_stat): Drop _stat suffix.
2173 (ggc_alloc_string): Remove macro.
2174 (ggc_alloc_rtx_def_stat): Adjust.
2175 (ggc_alloc_tree_node_stat): Likewise.
2176 (ggc_alloc_cleared_tree_node_stat): Likewise.
2177 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
2178 (ggc_alloc_cleared_simd_clone_stat): Likewise.
2179 * gimple.c (gimple_build_omp_for): Likewise.
2180 (gimple_copy): Likewise.
2181 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
2182 * toplev.c (realloc_for_line_map): Adjust.
2183 * tree-data-ref.h (lambda_vector_new): Likewise.
2184 * tree-phinodes.c (allocate_phi_node): Likewise.
2185 * tree.c (grow_tree_vec_stat): Likewise.
2186 * vec.h (va_gc::reserve): Adjust.
2188 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
2190 * config/microblaze/microblaze.c (break_handler): New Declaration.
2191 (microblaze_break_function_p,microblaze_is_break_handler): New.
2192 (compute_frame_size): Use microblaze_break_function_p.
2193 Add the test of break_handler.
2194 (microblaze_function_prologue) : Add the test of variable
2195 break_handler. Check the fnname by BREAK_HANDLER_NAME.
2196 (microblaze_function_epilogue) : Add the test of break_handler.
2197 (microblaze_globalize_label) : Add the test of break_handler.
2198 Check the name by BREAK_HANDLER_NAME.
2200 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
2202 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
2203 microblaze_is_break_handler test.
2204 (call_internal1,call_value_intern): Use microblaze_break_function_p.
2205 Use SYMBOL_REF_DECL.
2207 * config/microblaze/microblaze-protos.h
2208 (microblaze_break_function_p,microblaze_is_break_handler):
2211 * doc/extend.texi (MicroBlaze break_handler Functions): Document
2212 new MicroBlaze break_handler functions.
2214 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2216 * doc/extend.texi (Size of an asm): Move node text according
2217 to its @menu entry position.
2219 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
2221 PR tree-optimization/61140
2222 PR tree-optimization/61150
2223 PR tree-optimization/61197
2224 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
2226 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2228 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
2230 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
2232 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
2233 __SIZEOF_INT128__ is defined.
2235 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2237 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
2238 (rs6000_delegitimize_address): Use it.
2240 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2242 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
2243 inplace argument. Store the new address in the original MEM when true.
2244 * emit-rtl.c (change_address_1): Likewise.
2245 (adjust_address_1, adjust_automodify_address_1, offset_address):
2247 * rtl.h (plus_constant): Add an inplace argument.
2248 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
2249 when true. Avoid generating (plus X (const_int 0)).
2250 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
2251 in-place. Pass true to plus_constant.
2252 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
2254 2014-05-16 Dehao Chen <dehao@google.com>
2256 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
2258 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2261 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
2263 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
2265 2014-05-16 Dehao Chen <dehao@google.com>
2267 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
2268 optimize_function_for_size_p.
2269 * regs.h (REG_FREQ_FROM_BB): Likewise.
2271 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2274 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
2275 negt_reg_operand cases.
2276 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
2278 * config/sh/predicates.md (cbranch_treg_value): Simplify.
2280 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2282 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
2285 2014-05-16 David Malcolm <dmalcolm@redhat.com>
2288 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2290 * tree-cfg.c (dump_function_to_file): Dump the return type of
2291 functions, in a line to itself before the function body, mimicking
2292 the layout of a C function.
2294 2014-05-16 Dehao Chen <dehao@google.com>
2296 * cfghooks.c (make_forwarder_block): Use direct computation to
2297 get fall-through edge's count and frequency.
2299 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
2301 * config/arc/arc.c (arc_init): Fix typo in error message.
2302 * config/i386/i386.c (ix86_expand_builtin): Likewise.
2303 (split_stack_prologue_scratch_regno): Likewise.
2304 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
2305 word from error message.
2307 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
2309 * ira-costs.c: Fix typo in comment.
2311 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
2313 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
2315 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
2317 * varpool.c (dump_varpool_node): Dump write-only flag.
2318 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
2320 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
2321 write-only variables.
2322 * ipa.c (process_references): New function.
2323 (set_readonly_bit): New function.
2324 (set_writeonly_bit): New function.
2325 (clear_addressable_bit): New function.
2326 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
2327 fix handling of aliases.
2328 * cgraph.h (struct varpool_node): Add writeonly flag.
2330 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
2332 PR rtl-optimization/60969
2333 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
2334 Calculate costs for this case.
2336 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
2338 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
2339 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
2341 2014-05-16 Richard Biener <rguenther@suse.de>
2343 PR tree-optimization/61194
2344 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
2345 bool patterns ending in a COND_EXPR.
2347 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2349 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
2351 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2353 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
2354 where we were unable to cost an RTX.
2356 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2358 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
2361 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2362 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2364 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
2366 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2367 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2369 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
2370 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
2372 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2373 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2375 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
2378 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2379 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2381 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2384 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2385 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2387 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
2388 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
2390 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2391 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2393 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2396 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2397 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2399 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
2400 ZERO_EXTEND and SIGN_EXTEND better.
2402 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2403 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2405 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
2408 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2409 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2411 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
2412 costs when costing loads and stores to memory.
2414 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2415 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
2417 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
2420 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2422 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
2424 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2425 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2427 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
2429 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
2431 (aarch64_rtx_mult_cost): New.
2432 (aarch64_rtx_costs): Use it, refactor as appropriate.
2434 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2435 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2437 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
2438 emit instructions, return number of instructions which would
2440 (aarch64_add_constant): Update call to aarch64_build_constant.
2441 (aarch64_output_mi_thunk): Likewise.
2442 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
2445 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2447 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
2448 (TARGET_RTX_COSTS): Call it.
2450 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2452 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
2453 (cortexa57_vector_cost): Likewise.
2454 (cortexa57_tunings): Use them.
2456 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2458 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
2459 (cpu_addrcost_table): Use it.
2460 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
2461 (aarch64_address_cost): Rewrite using aarch64_classify_address,
2464 2014-05-16 Richard Biener <rguenther@suse.de>
2466 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
2467 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
2468 (visit_phi): Ignore edges marked as not executable.
2469 (class cond_dom_walker): New.
2470 (cond_dom_walker::before_dom_children): Value-number
2471 control statements and mark successor edges as not
2472 executable if possible.
2473 (run_scc_vn): First walk all control statements in
2474 dominator order, marking edges as not executable.
2475 * tree-inline.c (copy_edges_for_bb): Be not confused
2476 about random edge flags.
2478 2014-05-16 Richard Biener <rguenther@suse.de>
2480 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
2482 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
2485 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
2486 (__TM_simple_begin): Use it.
2487 (__TM_begin): Likewise.
2489 2014-05-15 Martin Jambor <mjambor@suse.cz>
2492 * ipa-prop.c (update_indirect_edges_after_inlining): Check
2493 type_preserved flag when the indirect edge is polymorphic.
2495 2014-05-15 Martin Jambor <mjambor@suse.cz>
2497 PR tree-optimization/61090
2498 * tree-sra.c (sra_modify_expr): Pass the current gsi to
2499 build_ref_for_model.
2501 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2503 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
2504 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
2506 2014-05-15 Jakub Jelinek <jakub@redhat.com>
2508 PR tree-optimization/61158
2509 * fold-const.c (fold_binary_loc): If X is zero-extended and
2510 shiftc >= prec, make sure zerobits is all ones instead of
2511 invoking undefined behavior.
2513 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2515 * regcprop.h: New file.
2516 * regcprop.c (skip_debug_insn_p): New decl.
2517 (replace_oldest_value_reg): Check skip_debug_insn_p.
2518 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
2519 * shrink-wrap.c: Include regcprop.h.
2520 (prepare_shrink_wrap): Call
2521 copyprop_hardreg_forward_bb_without_debug_insn.
2523 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2525 * shrink-wrap.h: Update comment.
2526 * shrink-wrap.c: Update comment.
2527 (next_block_for_reg): Rename to live_edge_for_reg.
2528 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
2529 (move_insn_for_shrink_wrap): Split live_edge.
2530 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
2532 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2534 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
2536 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
2537 * config/sparc/sparc.md (fptype_ut699): New attribute.
2538 (in_branch_delay): Return false if -mfix-ut699 is specified and
2539 fptype_ut699 is set to single.
2540 (truncdfsf2): Add fptype_ut699 attribute.
2541 (fix_truncdfsi2): Likewise.
2542 (floatsisf2): Change fptype attribute.
2543 (fix_truncsfsi2): Likewise.
2544 (negtf2_notv9): Delete.
2545 (negtf2_v9): Likewise.
2546 (negtf2_hq): New instruction.
2547 (negtf2): New instruction and splitter.
2548 (negdf2_notv9): Rewrite.
2549 (abstf2_notv9): Delete.
2550 (abstf2_hq_v9): Likewise.
2551 (abstf2_v9): Likewise.
2552 (abstf2_hq): New instruction.
2553 (abstf2): New instruction and splitter.
2554 (absdf2_notv9): Rewrite.
2556 2014-05-14 Cary Coutant <ccoutant@google.com>
2559 * opts.c (common_handle_option): Don't special-case "-g".
2560 (set_debug_level): Default to at least level 2 with "-g".
2562 2014-05-14 DJ Delorie <dj@redhat.com>
2564 * config/msp430/msp430.c (msp430_builtin): Add
2565 MSP430_BUILTIN_DELAY_CYCLES.
2566 (msp430_init_builtins): Register void __delay_cycles(long long).
2567 (msp430_builtin_decl): Add it.
2568 (cg_magic_constant): New.
2569 (msp430_expand_delay_cycles): New.
2570 (msp430_expand_builtin): Call it.
2571 (msp430_print_operand_raw): Change integer printing from "int" to
2573 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
2574 (delay_cycles_start): New.
2575 (delay_cycles_end): New.
2576 (delay_cycles_32): New.
2577 (delay_cycles_32x): New.
2578 (delay_cycles_16): New.
2579 (delay_cycles_16x): New.
2580 (delay_cycles_2): New.
2581 (delay_cycles_1): New.
2582 * doc/extend.texi: Document __delay_cycles().
2584 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
2586 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
2587 length attribute computation.
2589 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
2592 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
2594 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
2597 * config/sparc/sparc.md: Fix types of low and high in DI constant
2598 splitter. Use gen_int_mode in some other splitters.
2600 2014-05-14 Martin Jambor <mjambor@suse.cz>
2603 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
2605 2014-05-14 James Norris <jnorris@codesourcery.com>
2607 * omp-low.c (expand_parallel_call): Remove shadow variable.
2608 (expand_omp_taskreg): Likewise.
2610 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
2612 * common/config/i386/i386-common.c
2613 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
2614 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
2615 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
2616 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
2617 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
2618 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
2619 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
2620 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
2621 xsavecintrin.h, xsavesintrin.h.
2622 (x86_64-*-*): Ditto.
2623 * config/i386/clflushoptintrin.h: New.
2624 * config/i386/xsavecintrin.h: Ditto.
2625 * config/i386/xsavesintrin.h: Ditto.
2626 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
2627 (bit_XSAVES): Ditto.
2628 (bit_XSAVES): Ditto.
2629 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
2630 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
2632 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2633 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
2634 OPTION_MASK_ISA_XSAVES.
2635 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
2637 (PTA_CLFLUSHOPT) Define.
2638 (PTA_XSAVEC): Ditto.
2639 (PTA_XSAVES): Ditto.
2640 (ix86_option_override_internal): Handle new options.
2641 (ix86_valid_target_attribute_inner_p): Ditto.
2642 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
2643 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
2644 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
2645 (bdesc_special_args): Add __builtin_ia32_xsaves,
2646 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
2647 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
2648 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
2649 (ix86_expand_builtin): Handle new builtins.
2650 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
2651 (TARGET_CLFLUSHOPT_P): Ditto.
2652 (TARGET_XSAVEC): Ditto.
2653 (TARGET_XSAVEC_P): Ditto.
2654 (TARGET_XSAVES): Ditto.
2655 (TARGET_XSAVES_P): Ditto.
2656 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
2657 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
2658 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
2660 (ANY_XRSTOR64): Ditto.
2662 (xrstor): Change into <xrstor>.
2663 (xrstor_rex64): Change into <xrstor>_rex64.
2664 (xrstor64): Change into <xrstor>64
2666 * config/i386/i386.opt (mclflushopt): New.
2669 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
2671 * doc/invoke.texi: Document new options.
2673 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2675 PR rtl-optimization/60866
2676 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
2677 Default it to -1. Pass it down to init_simplejump_data.
2678 (init_simplejump_data): New parameter old_seqno. Pass it down
2679 to get_seqno_for_a_jump.
2680 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
2681 initializing new jump seqno as a last resort. Add comment.
2682 (sel_redirect_edge_and_branch): Save old seqno of the conditional
2683 jump and pass it down to sel_init_new_insn.
2684 (sel_redirect_edge_and_branch_force): Likewise.
2686 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
2688 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
2689 shifted values to avoid build warning.
2691 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2693 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
2694 * cfgrtl.c (rtl_merge_blocks): Fix comment.
2695 (cfg_layout_merge_blocks): Likewise.
2696 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
2698 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2700 PR rtl-optimization/60901
2701 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
2702 bb predecessor belongs to the same scheduling region. Adjust comment.
2704 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
2706 * doc/sourcebuild.texi: (dfp_hw): Document.
2707 (p8vector_hw): Likewise.
2708 (powerpc_eabi_ok): Likewise.
2709 (powerpc_elfv2): Likewise.
2710 (powerpc_htm_ok): Likewise.
2711 (ppc_recip_hw): Likewise.
2714 2014-05-13 Cary Coutant <ccoutant@google.com>
2716 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2718 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2720 * gengtype-parse.c (require3): Eliminate in favor of...
2722 (require_template_declaration): Update to support optional single *
2725 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2726 (create_user_defined_type): Handle a single level of explicit
2727 pointerness within template arguments.
2728 (struct write_types_data): Add field "kind".
2729 (filter_type_name): Handle "*" character.
2730 (write_user_func_for_structure_ptr): Require a write_types_data
2731 rather than just a prefix string, so that we can look up the kind
2732 of the wtd and use it as an index into wrote_user_func_for_ptr,
2733 ensuring that such functions are written at most once. Support
2734 subclasses by invoking the marking function of the ultimate base class.
2735 (write_user_func_for_structure_body): Require a write_types_data
2736 rather than just a prefix string, so that we can pass this to
2737 write_user_func_for_structure_ptr.
2738 (write_func_for_structure): Likewise.
2739 (ggc_wtd): Add initializer of new "kind" field.
2740 (pch_wtd): Likewise.
2742 * gengtype.h (enum write_types_kinds): New.
2743 (struct type): Add field wrote_user_func_for_ptr to the "s"
2746 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2748 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2749 instead of const_binop.
2750 (fold_binary_loc): Likewise.
2752 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2754 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2755 calculation to match get_ref_base_and_extent.
2757 2014-05-13 Catherine Moore <clm@codesourcery.com>
2758 Sandra Loosemore <sandra@codesourcery.com>
2760 * configure.ac: Fix assembly for explicit JALR relocation check.
2761 * configure: Regenerate.
2763 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2765 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2766 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2767 Remove associated type declarations and initialisations.
2768 (arm_expand_neon_builtin): Likewise.
2769 (neon_emit_pair_result_insn): Delete.
2770 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2771 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2772 (neon_vzip<mode>): Likewise.
2773 (neon_vuzp<mode>): Likewise.
2775 2014-05-13 Richard Biener <rguenther@suse.de>
2778 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2779 it needs revisiting whether the call still may be tail-called.
2781 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2783 * rtl.def (SYMBOL_REF): Remove middle "0" field.
2784 * rtl.h (block_symbol): Reduce number of fields to 2.
2785 (rtx_def): Add u2.symbol_ref_flags.
2786 (SYMBOL_REF_FLAGS): Use it.
2787 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
2788 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
2789 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
2790 Lower index of SYMBOL_REF_DATA.
2791 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
2792 Print SYMBOL_REF_FLAGS at the same time.
2793 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
2795 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2797 * rtl.def (VAR_LOCATION): Remove "i" field.
2798 * rtl.h (rtx_def): Add u2.var_location_status.
2799 (PAT_VAR_LOCATION_STATUS): Use it.
2800 (gen_rtx_VAR_LOCATION): Declare.
2801 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
2802 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
2803 * var-tracking.c (emit_note_insn_var_location): Remove casts.
2805 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2807 * rtl.def (scratch): Fix outdated comment and remove "0" field.
2808 * gengtype.c (adjust_field_rtx_def): Update accordingly.
2810 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2812 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
2813 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
2814 * rtl.h (rtx_def): Add insn_uid to u2 field.
2815 (RTX_FLAG_CHECK8): Delete in favor of...
2816 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
2817 (INSN_DELETED_P): Update accordingly.
2818 (INSN_UID): Use u2.insn_uid.
2819 (INSN_CHAIN_CODE_P): Define.
2820 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
2821 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
2822 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
2823 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
2824 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
2825 indices accordingly.
2826 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
2827 Update indices for insn-chain rtxes.
2828 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
2829 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
2830 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
2831 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
2832 * combine.c (try_combine): Likewise.
2833 * ira.c (setup_prohibited_mode_move_regs): Likewise.
2835 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2837 * rtl.def (REG): Remove middle field.
2838 * rtl.h (rtx_def): Add orignal_regno to u2.
2839 (ORIGINAL_REGNO): Use it instead of field 1.
2840 (REG_ATTRS): Lower field index accordingly.
2841 * gengtype.c (adjust_field_rtx_def): Remove handling of
2842 ORIGINAL_REGNO. Move REG_ATTRS index down.
2843 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
2844 code that prints the REGNO.
2846 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2848 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
2851 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2853 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
2855 2014-05-13 Bin Cheng <bin.cheng@arm.com>
2857 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
2858 (alloc_iv): Lower base expressions containing ADDR_EXPR.
2860 2014-05-13 Ian Bolton <ian.bolton@arm.com>
2862 * config/aarch64/aarch64-protos.h
2863 (aarch64_hard_regno_caller_save_mode): New prototype.
2864 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
2866 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
2868 2014-05-13 Christian Bruel <christian.bruel@st.com>
2870 * target.def (mode_switching): New hook vector.
2871 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
2872 (mode_exit, modepriority_to_mode): Likewise.
2873 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
2874 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2875 * target.h: Include tm.h and hard-reg-set.h.
2876 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
2877 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
2878 * doc/tm.texi Regenerate.
2879 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2880 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2881 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
2882 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
2883 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2884 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2885 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
2886 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
2887 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
2888 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
2889 (ix86_emit_mode_set): Hookify.
2890 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
2892 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2893 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
2894 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
2895 (epiphany_mode_priority_to_mode): Remove declaration.
2896 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
2897 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
2898 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
2900 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
2901 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
2902 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
2904 2014-05-13 Jakub Jelinek <jakub@redhat.com>
2907 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
2908 is const0_rtx, return immediately. Don't test count == 0 when
2911 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2913 * Makefile.in: add shrink-wrap.o.
2914 * config/i386/i386.c: include "shrink-wrap.h"
2915 * function.c: Likewise.
2916 (requires_stack_frame_p, next_block_for_reg,
2917 move_insn_for_shrink_wrap, prepare_shrink_wrap,
2918 dup_block_and_redirect): Move to shrink-wrap.c
2919 (thread_prologue_and_epilogue_insns): Extract three code segments
2920 as functions in shrink-wrap.c
2921 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
2923 * shrink-wrap.c: New file.
2924 * shrink-wrap.h: New file.
2926 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2928 * doc/extend.texi: Reflect current numbers of pragmas. Remove
2929 reference to Solaris.
2931 2014-05-12 Mike Stump <mikestump@comcast.net>
2934 * genattrtab.c (filename): Add.
2935 (convert_set_attr_alternative): Improve error message.
2936 (check_defs): Restore read_md_filename for error messages.
2937 (gen_insn): Save filename.
2939 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
2941 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
2942 -fno-local-ivars and -fivar-visibility.
2943 * c-family/c.opt: Make -Wshadow also implicitly enable
2946 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2948 * doc/tm.texi: Remove reference to deleted macro.
2949 * doc/tm.texi.in: Likewise.
2951 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
2954 * config/avr/avr.c (avr_out_store_psi): Use correct constant
2957 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
2960 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
2961 * config/arm/aout.h (License): Same.
2962 * config/arm/bpabi.h (License): Same.
2963 * config/arm/elf.h (License): Same.
2964 * config/arm/linux-elf.h (License): Same.
2965 * config/arm/linux-gas.h (License): Same.
2966 * config/arm/netbsd-elf.h (License): Same.
2967 * config/arm/uclinux-eabi.h (License): Same.
2968 * config/arm/uclinux-elf.h (License): Same.
2969 * config/arm/vxworks.h (License): Same.
2971 2014-05-11 Jakub Jelinek <jakub@redhat.com>
2973 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
2974 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
2975 number of operands to 3.
2976 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
2977 * tree-nested.c (convert_nonlocal_omp_clauses,
2978 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
2979 * gimplify.c (gimplify_scan_omp_clauses): Handle
2980 OMP_CLAUSE_LINEAR_STMT.
2981 * omp-low.c (lower_rec_input_clauses): Fix typo.
2982 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
2983 cast between Fortran boolean_type_node and C _Bool if
2986 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
2988 PR tree-optimization/61136
2989 * wide-int.h (multiple_of_p): Define a version that doesn't return
2991 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
2992 integer_zerop/const_binop pair.
2993 (multiple_of_p): Likewise, converting both operands to widest_int
2996 2014-05-09 Teresa Johnson <tejohnson@google.com>
2998 * cgraphunit.c (analyze_functions): Use correct dump file.
3000 2014-05-09 Florian Weimer <fweimer@redhat.com>
3002 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
3004 (stack_protect_return_slot_p): New function.
3005 (expand_used_vars): Call stack_protect_decl_p and
3006 stack_protect_return_slot_p for -fstack-protector-strong.
3008 2014-05-09 David Wohlferd <LimeGreenSocks@yahoo.com>
3009 Andrew Haley <aph@redhat.com>
3010 Richard Sandiford <rdsandiford@googlemail.com>
3012 * doc/extend.texi: Rewrite inline asm page / re-org asm-related
3015 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
3018 * fold-const.c (fold_binary_loc): Changed width of mask.
3020 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3022 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
3023 unsigned int initializers for regno_in, regno_out.
3025 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3028 * config/avr/avr.md (cc): Add new attribute set_vzn.
3029 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
3030 Set cc insn attribute to set_vzn instead of set_zn for alternatives
3031 with INC, DEC or NEG.
3032 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
3033 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
3034 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
3036 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3039 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3041 * wide-int.cc (UTItype): Define.
3042 (UDWtype): Define for appropriate W_TYPE_SIZE.
3044 2014-05-09 Richard Biener <rguenther@suse.de>
3046 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
3047 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
3048 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
3049 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
3050 ssa_propagate): Adjust.
3052 2014-05-08 Jeff Law <law@redhat.com>
3054 PR tree-optimization/61009
3055 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
3056 tri-state rather than a boolean. When a block is too big to
3057 thread through, inform caller via negative return value.
3058 (thread_across_edge): If a block was too big for normal threading,
3059 then it's too big for a joiner too, so remove temporary equivalences
3060 and return immediately.
3062 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
3063 Matthias Klose <doko@ubuntu.com>
3066 * optc-gen.awk: Fix option handling for -Wunused-parameter.
3068 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3071 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
3073 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3076 * config/alpha/alpha.c: Include gimple-iterator.h.
3077 (alpha_gimple_fold_builtin): New function. Move
3078 ALPHA_BUILTIN_UMULH folding from ...
3079 (alpha_fold_builtin): ... here.
3080 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
3082 2014-05-08 Wei Mi <wmi@google.com>
3085 * config/i386/i386.c (ix86_compute_frame_layout): Update
3086 preferred_stack_boundary for call, expanded from tls descriptor.
3087 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
3088 to depend on SP register.
3089 (*tls_local_dynamic_base_32_gnu): Ditto.
3090 (*tls_local_dynamic_32_once): Ditto.
3091 (tls_global_dynamic_64_<mode>): Set
3092 ix86_tls_descriptor_calls_expanded_in_cfun.
3093 (tls_local_dynamic_base_64_<mode>): Ditto.
3094 (tls_global_dynamic_32): Set
3095 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
3096 to depend on SP register.
3097 (tls_local_dynamic_base_32): Ditto.
3099 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3101 * config/arm/arm_neon.h: Update comment.
3102 * config/arm/neon-docgen.ml: Delete.
3103 * config/arm/neon-gen.ml: Delete.
3104 * doc/arm-neon-intrinsics.texi: Update comment.
3106 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3108 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
3110 (vand, vorr, veor, vorn, vbic): Remove.
3111 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
3113 (neon_vsub_unspec): Likewise.
3114 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
3116 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3118 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
3119 (vadd_s16): Likewise.
3120 (vadd_s32): Likewise.
3121 (vadd_f32): Likewise.
3122 (vadd_u8): Likewise.
3123 (vadd_u16): Likewise.
3124 (vadd_u32): Likewise.
3125 (vadd_s64): Likewise.
3126 (vadd_u64): Likewise.
3127 (vaddq_s8): Likewise.
3128 (vaddq_s16): Likewise.
3129 (vaddq_s32): Likewise.
3130 (vaddq_s64): Likewise.
3131 (vaddq_f32): Likewise.
3132 (vaddq_u8): Likewise.
3133 (vaddq_u16): Likewise.
3134 (vaddq_u32): Likewise.
3135 (vaddq_u64): Likewise.
3136 (vmul_s8): Likewise.
3137 (vmul_s16): Likewise.
3138 (vmul_s32): Likewise.
3139 (vmul_f32): Likewise.
3140 (vmul_u8): Likewise.
3141 (vmul_u16): Likewise.
3142 (vmul_u32): Likewise.
3143 (vmul_p8): Likewise.
3144 (vmulq_s8): Likewise.
3145 (vmulq_s16): Likewise.
3146 (vmulq_s32): Likewise.
3147 (vmulq_f32): Likewise.
3148 (vmulq_u8): Likewise.
3149 (vmulq_u16): Likewise.
3150 (vmulq_u32): Likewise.
3151 (vsub_s8): Likewise.
3152 (vsub_s16): Likewise.
3153 (vsub_s32): Likewise.
3154 (vsub_f32): Likewise.
3155 (vsub_u8): Likewise.
3156 (vsub_u16): Likewise.
3157 (vsub_u32): Likewise.
3158 (vsub_s64): Likewise.
3159 (vsub_u64): Likewise.
3160 (vsubq_s8): Likewise.
3161 (vsubq_s16): Likewise.
3162 (vsubq_s32): Likewise.
3163 (vsubq_s64): Likewise.
3164 (vsubq_f32): Likewise.
3165 (vsubq_u8): Likewise.
3166 (vsubq_u16): Likewise.
3167 (vsubq_u32): Likewise.
3168 (vsubq_u64): Likewise.
3169 (vand_s8): Likewise.
3170 (vand_s16): Likewise.
3171 (vand_s32): Likewise.
3172 (vand_u8): Likewise.
3173 (vand_u16): Likewise.
3174 (vand_u32): Likewise.
3175 (vand_s64): Likewise.
3176 (vand_u64): Likewise.
3177 (vandq_s8): Likewise.
3178 (vandq_s16): Likewise.
3179 (vandq_s32): Likewise.
3180 (vandq_s64): Likewise.
3181 (vandq_u8): Likewise.
3182 (vandq_u16): Likewise.
3183 (vandq_u32): Likewise.
3184 (vandq_u64): Likewise.
3185 (vorr_s8): Likewise.
3186 (vorr_s16): Likewise.
3187 (vorr_s32): Likewise.
3188 (vorr_u8): Likewise.
3189 (vorr_u16): Likewise.
3190 (vorr_u32): Likewise.
3191 (vorr_s64): Likewise.
3192 (vorr_u64): Likewise.
3193 (vorrq_s8): Likewise.
3194 (vorrq_s16): Likewise.
3195 (vorrq_s32): Likewise.
3196 (vorrq_s64): Likewise.
3197 (vorrq_u8): Likewise.
3198 (vorrq_u16): Likewise.
3199 (vorrq_u32): Likewise.
3200 (vorrq_u64): Likewise.
3201 (veor_s8): Likewise.
3202 (veor_s16): Likewise.
3203 (veor_s32): Likewise.
3204 (veor_u8): Likewise.
3205 (veor_u16): Likewise.
3206 (veor_u32): Likewise.
3207 (veor_s64): Likewise.
3208 (veor_u64): Likewise.
3209 (veorq_s8): Likewise.
3210 (veorq_s16): Likewise.
3211 (veorq_s32): Likewise.
3212 (veorq_s64): Likewise.
3213 (veorq_u8): Likewise.
3214 (veorq_u16): Likewise.
3215 (veorq_u32): Likewise.
3216 (veorq_u64): Likewise.
3217 (vbic_s8): Likewise.
3218 (vbic_s16): Likewise.
3219 (vbic_s32): Likewise.
3220 (vbic_u8): Likewise.
3221 (vbic_u16): Likewise.
3222 (vbic_u32): Likewise.
3223 (vbic_s64): Likewise.
3224 (vbic_u64): Likewise.
3225 (vbicq_s8): Likewise.
3226 (vbicq_s16): Likewise.
3227 (vbicq_s32): Likewise.
3228 (vbicq_s64): Likewise.
3229 (vbicq_u8): Likewise.
3230 (vbicq_u16): Likewise.
3231 (vbicq_u32): Likewise.
3232 (vbicq_u64): Likewise.
3233 (vorn_s8): Likewise.
3234 (vorn_s16): Likewise.
3235 (vorn_s32): Likewise.
3236 (vorn_u8): Likewise.
3237 (vorn_u16): Likewise.
3238 (vorn_u32): Likewise.
3239 (vorn_s64): Likewise.
3240 (vorn_u64): Likewise.
3241 (vornq_s8): Likewise.
3242 (vornq_s16): Likewise.
3243 (vornq_s32): Likewise.
3244 (vornq_s64): Likewise.
3245 (vornq_u8): Likewise.
3246 (vornq_u16): Likewise.
3247 (vornq_u32): Likewise.
3248 (vornq_u64): Likewise.
3250 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3252 * wide-int.cc (UTItype): Define.
3253 (UDWtype): Define for appropriate W_TYPE_SIZE.
3255 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
3257 PR tree-optimization/59100
3258 * tree-ssa-phiopt.c: Include tree-inline.h.
3259 (neutral_element_p, absorbing_element_p): New functions.
3260 (value_replacement): Handle conditional binary operations with a
3261 neutral or absorbing element.
3263 2014-05-08 Richard Biener <rguenther@suse.de>
3265 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
3266 folding the expression.
3267 (valueize_expr): Remove.
3268 (visit_reference_op_load): Do not valueize the result of
3270 (simplify_binary_expression): Likewise.
3271 (simplify_unary_expression): Likewise.
3273 2014-05-08 Richard Biener <rguenther@suse.de>
3275 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
3276 looking at TYPE_ARG_TYPES.
3278 2014-05-08 Richard Biener <rguenther@suse.de>
3280 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
3281 pointer propagation special-case.
3283 2014-05-08 Bin Cheng <bin.cheng@arm.com>
3285 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
3286 core part of address expressions.
3288 2014-05-08 Alan Modra <amodra@gmail.com>
3291 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
3292 loads and stores when -mno-strict-align at any alignment.
3293 (expand_block_clear): Similarly. Also correct calculation of
3296 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
3299 * tree-complex.c (expand_complex_move): Keep line info when expanding
3301 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
3302 of complex expression. Use new argument to display correct location
3303 for values coming from phi statement.
3304 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
3305 (warn_uninitialized_phi): Pass location of phi argument to
3307 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
3308 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
3310 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
3312 * config/rs6000/predicates.md (indexed_address_mem): New.
3313 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
3314 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
3315 fpstore_ux, fpstore_u.
3316 (sign_extend, indexed, update): New.
3317 (cell_micro): Adjust.
3318 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
3319 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
3320 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
3321 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
3322 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
3323 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
3324 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
3325 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
3326 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
3327 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
3328 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
3329 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
3330 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
3331 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
3332 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
3334 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
3335 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
3336 *vsx_extract_<mode>_store): Adjust.
3337 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
3338 is_cracked_insn, insn_must_be_first_in_group,
3339 insn_must_be_last_in_group): Adjust.
3341 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
3343 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
3344 ppc440-fpstore): Adjust.
3345 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
3346 ppc476-fpstore): Adjust.
3347 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
3348 ppc601-fpstore): Adjust.
3349 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
3351 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
3353 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
3354 ppc7450-fpstore): Adjust.
3355 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
3356 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
3357 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
3359 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
3360 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
3361 cell-fpstore, cell-fpstore-update): Adjust.
3362 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
3363 ppce300c3_store, ppce300c3_fpstore): Adjust.
3364 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
3365 e500mc_fpstore): Adjust.
3366 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
3367 e500mc64_store, e500mc64_fpstore): Adjust.
3368 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
3369 e5500_fpstore): Adjust.
3370 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
3371 e6500_fpstore): Adjust.
3372 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
3374 * config/rs6000/power4.md (power4-load, power4-load-ext,
3375 power4-load-ext-update, power4-load-ext-update-indexed,
3376 power4-load-update-indexed, power4-load-update, power4-fpload,
3377 power4-fpload-update, power4-store, power4-store-update,
3378 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
3380 * config/rs6000/power5.md (power5-load, power5-load-ext,
3381 power5-load-ext-update, power5-load-ext-update-indexed,
3382 power5-load-update-indexed, power5-load-update, power5-fpload,
3383 power5-fpload-update, power5-store, power5-store-update,
3384 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
3386 * config/rs6000/power6.md (power6-load, power6-load-ext,
3387 power6-load-update, power6-load-update-indexed,
3388 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
3389 power6-fpload-update, power6-store, power6-store-update,
3390 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
3392 * config/rs6000/power7.md (power7-load, power7-load-ext,
3393 power7-load-update, power7-load-update-indexed,
3394 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
3395 power7-fpload-update, power7-store, power7-store-update,
3396 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
3398 * config/rs6000/power8.md (power8-load, power8-load-update,
3399 power8-load-ext, power8-load-ext-update, power8-fpload,
3400 power8-fpload-update, power8-store, power8-store-update-indexed,
3401 power8-fpstore, power8-fpstore-update): Adjust.
3402 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
3404 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
3405 titan_lsu_store, titan_lsu_fpstore): Adjust.
3406 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
3408 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
3411 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
3412 unrolled byte insns. Emit address increments after move insns.
3414 2014-05-07 David Malcolm <dmalcolm@redhat.com>
3416 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
3417 const_gimple, rather than a gimple.
3418 (gimple_call_builtin_p): Likewise, for the three variants.
3420 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
3421 (gimple_call_builtin_p): Likewise, for the three variants.
3423 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3425 PR tree-optimization/61095
3426 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
3428 2014-05-07 Richard Biener <rguenther@suse.de>
3430 PR tree-optimization/61034
3431 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
3432 (maybe_skip_until): Use translate to take into account
3433 lattices when trying to do disambiguations.
3434 (get_continuation_for_phi_1): Likewise.
3435 (get_continuation_for_phi): Adjust for added translate arguments.
3436 (walk_non_aliased_vuses): Likewise.
3437 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
3438 (walk_non_aliased_vuses): Likewise.
3439 (call_may_clobber_ref_p_1): Declare.
3440 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
3441 calls. Stop early if we are only supposed to disambiguate.
3442 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
3444 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
3446 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
3447 Emit an error when the function has arguments.
3449 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3451 * cfgloop.h (unswitch_loops): Remove.
3452 * doc/passes.texi: Remove references to loop-unswitch.c
3453 * timevar.def (TV_LOOP_UNSWITCH): Remove.
3455 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
3457 * tree-vect-data-refs.c (vect_grouped_load_supported): New
3458 check for loads group of length 3.
3459 (vect_permute_load_chain): New permutations for loads group of
3461 * tree-vect-stmts.c (vect_model_load_cost): Change cost
3462 of vec_perm_shuffle for the new permutations.
3464 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
3466 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
3467 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
3468 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
3469 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
3470 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
3471 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
3472 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
3473 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
3475 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3477 * loop-unswitch.c: Delete.
3479 2014-05-07 Richard Biener <rguenther@suse.de>
3481 * config.gcc: Always set need_64bit_hwint to yes.
3483 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
3485 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
3486 of using optimize_size.
3488 2014-05-06 Mike Stump <mikestump@comcast.net>
3490 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
3492 2014-05-06 Joseph Myers <joseph@codesourcery.com>
3494 * config/i386/sse.md (*mov<mode>_internal)
3495 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
3496 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
3497 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
3498 (*<code><mode>3, *andnot<mode>3<mask_name>)
3499 (<mask_codefor><code><mode>3<mask_name>): Only consider
3500 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
3502 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3505 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3507 * lra-constraints.c (valid_address_p): Move earlier in file.
3508 Add a constraint argument to the address_info version.
3509 (satisfies_memory_constraint_p): New function.
3510 (satisfies_address_constraint_p): Likewise.
3511 (process_alt_operands, curr_insn_transform): Use them.
3512 (process_address): Pass the constraint to valid_address_p when
3513 checking address operands.
3515 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
3517 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
3518 to their respective blocks. Fix inadvertent use of "node".
3520 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3522 * emit-rtl.c (init_derived_machine_modes): New functionm, split
3524 (init_emit_once): ...here.
3525 * rtl.h (init_derived_machine_modes): Declare.
3526 * toplev.c (do_compile): Call it even if no_backend.
3528 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
3529 Mike Stump <mikestump@comcast.net>
3530 Richard Sandiford <rdsandiford@googlemail.com>
3531 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3533 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
3534 (rtx_equal_for_memref_p): Update comment.
3535 (adjust_offset_for_component_ref): Use wide-int interfaces.
3536 * builtins.c (get_object_alignment_2): Likewise.
3537 (c_readstr): Likewise.
3538 (target_char_cast): Add comment.
3539 (determine_block_size): Use wide-int interfaces.
3540 (expand_builtin_signbit): Likewise.
3541 (fold_builtin_int_roundingfn): Likewise.
3542 (fold_builtin_bitop): Likewise.
3543 (fold_builtin_bswap): Likewise.
3544 (fold_builtin_logarithm): Use signop.
3545 (fold_builtin_pow): Likewise.
3546 (fold_builtin_memory_op): Use wide-int interfaces.
3547 (fold_builtin_object_size): Likewise.
3548 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
3549 nb_iterations_estimate.
3550 (record_niter_bound): Use wide-int interfaces.
3551 (get_estimated_loop_iterations_int): Likewise.
3552 (get_estimated_loop_iterations): Likewise.
3553 (get_max_loop_iterations): Likewise.
3554 * cfgloop.h: Include wide-int.h.
3555 (struct nb_iter_bound): Change bound to widest_int.
3556 (struct loop): Change nb_iterations_upper_bound and
3557 nb_iterations_estimate to widest_int.
3558 (record_niter_bound): Switch to use widest_int.
3559 (get_estimated_loop_iterations): Likewise.
3560 (get_max_loop_iterations): Likewise.
3561 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
3562 update for wide-int.
3563 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
3564 * combine.c (try_combine): Likewise.
3565 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
3566 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
3568 (aarch64_float_const_representable_p): Likewise.
3569 * config/arc/arc.c: Include wide-int.h.
3570 (arc_can_use_doloop_p): Use wide-int interfaces.
3571 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
3572 (vfp3_const_double_index): Likewise.
3573 * config/avr/avr.c (avr_out_round): Likewise.
3574 (avr_fold_builtin): Likewise.
3575 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
3576 (bfin_can_use_doloop_p): Likewise.
3577 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
3578 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
3579 * config/i386/i386.c: Include wide-int.h.
3580 (ix86_data_alignment): Use wide-int interfaces.
3581 (ix86_local_alignment): Likewise.
3582 (ix86_emit_swsqrtsf): Update real_from_integer.
3583 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
3584 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
3585 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
3586 (zero_constant): Likewise.
3587 (input_operand): Likewise.
3588 (splat_input_operand): Likewise.
3589 (non_logical_cint_operand): Change const_double to const_wide_int.
3590 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
3591 (easy_altivec_constant): Remove comment.
3592 (paired_expand_vector_init): Use CONSTANT_P.
3593 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
3594 (rs6000_emit_move): Update checks.
3595 (rs6000_aggregate_candidate): Use wide-int interfaces.
3596 (rs6000_expand_ternop_builtin): Likewise.
3597 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
3598 (rs6000_assemble_integer): Likewise.
3599 (rs6000_hash_constant): Likewise.
3600 (output_toc): Likewise.
3601 (rs6000_rtx_costs): Likewise.
3602 (rs6000_emit_swrsqrt); Update call to real_from_integer.
3603 * config/rs6000/rs6000-c.c: Include wide-int.h.
3604 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
3605 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
3606 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
3607 Handle CONST_WIDE_INT.
3608 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
3609 Use tree_fits_uhwi_p.
3610 * config/sparc/sparc.c: Include wide-int.h.
3611 (sparc_fold_builtin): Use wide-int interfaces.
3612 * config/vax/vax.c: Include wide-int.h.
3613 (vax_float_literal): Use real_from_integer.
3614 * coretypes.h (struct hwivec_def): New.
3616 (const_hwivec): New.
3617 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
3618 (equiv_constant): Handle CONST_WIDE_INT.
3619 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
3620 (cselib_hash_rtx): Handle CONST_WIDE_INT.
3621 * dbxout.c (stabstr_U): Use wide-int interfaces.
3622 (dbxout_type): Update to use cst_fits_shwi_p.
3623 * defaults.h (LOG2_BITS_PER_UNIT): Define.
3624 (TARGET_SUPPORTS_WIDE_INT): Add default.
3625 * dfp.c: Include wide-int.h.
3626 (decimal_real_to_integer2): Use wide-int interfaces and rename to
3627 decimal_real_to_integer.
3628 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
3629 decimal_real_to_integer.
3630 * doc/generic.texi (Constant expressions): Update for wide_int.
3631 * doc/rtl.texi (const_double): Likewise.
3632 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
3633 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
3634 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
3635 (REAL_VALUE_FROM_INT): Remove.
3636 (TARGET_SUPPORTS_WIDE_INT): New.
3637 * doc/tm.texi: Regenerate.
3638 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
3639 * double-int.h: Include wide-int.h.
3640 (struct wi::int_traits): New.
3641 * dwarf2out.c (get_full_len): New.
3642 (dw_val_equal_p): Add case dw_val_class_wide_int.
3643 (size_of_loc_descr): Likewise.
3644 (output_loc_operands): Likewise.
3645 (insert_double): Remove.
3646 (insert_wide_int): New.
3648 (print_die): Add case dw_val_class_wide_int.
3649 (attr_checksum): Likewise.
3650 (attr_checksum_ordered): Likewise.
3651 (same_dw_val_p): Likewise.
3652 (size_of_die): Likewise.
3653 (value_format): Likewise.
3654 (output_die): Likewise.
3655 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
3657 (clz_loc_descriptor): Use wide-int interfaces.
3658 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
3659 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
3660 (round_up_to_align): Use wide-int interfaces.
3661 (field_byte_offset): Likewise.
3662 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
3663 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
3664 CONST_DOUBLE handling. Use wide-int interfaces.
3665 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
3666 (gen_enumeration_type_die): Use add_AT_wide.
3667 (hash_loc_operands): Add case dw_val_class_wide_int.
3668 (compare_loc_operands): Likewise.
3669 * dwarf2out.h: Include wide-int.h.
3670 (wide_int_ptr): New.
3671 (enum dw_val_class): Add dw_val_class_wide_int.
3672 (struct dw_val_struct): Add val_wide.
3673 * emit-rtl.c (const_wide_int_htab): New.
3674 (const_wide_int_htab_hash): New.
3675 (const_wide_int_htab_eq): New.
3676 (lookup_const_wide_int): New.
3677 (const_double_htab_hash): Use wide-int interfaces.
3678 (const_double_htab_eq): Likewise.
3679 (rtx_to_double_int): Conditionally compile for wide-int.
3680 (immed_double_int_const): Rename to immed_wide_int_const and
3681 update for wide-int.
3682 (immed_double_const): Conditionally compile for wide-int.
3683 (init_emit_once): Use wide-int interfaces.
3684 * explow.c (plus_constant): Likewise.
3685 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
3686 (lshift_value): Use wide-int interfaces.
3687 (expand_mult): Likewise.
3688 (choose_multiplier): Likewise.
3689 (expand_smod_pow2): Likewise.
3690 (make_tree): Likewise.
3691 * expr.c (convert_modes): Consolidate handling of constants.
3692 Use wide-int interfaces.
3693 (emit_group_load_1): Add note.
3694 (store_expr): Update comment.
3695 (get_inner_reference): Use wide-int interfaces.
3696 (expand_constructor): Update comment.
3697 (expand_expr_real_2): Use wide-int interfaces.
3698 (expand_expr_real_1): Likewise.
3699 (reduce_to_bit_field_precision): Likewise.
3700 (const_vector_from_tree): Likewise.
3701 * final.c: Include wide-int-print.h.
3702 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
3703 * fixed-value.c: Include wide-int.h.
3704 (fixed_from_string): Use wide-int interfaces.
3705 (fixed_to_decimal): Likewise.
3706 (fixed_convert_from_real): Likewise.
3707 (real_convert_from_fixed): Likewise.
3708 * fold-const.h (mem_ref_offset): Return an offset_int.
3709 (div_if_zero_remainder): Remove code parameter.
3710 * fold-const.c (div_if_zero_remainder): Remove code parameter.
3711 Use wide-int interfaces.
3712 (may_negate_without_overflow_p): Use wide-int interfaces.
3713 (negate_expr_p): Likewise.
3714 (fold_negate_expr): Likewise.
3715 (int_const_binop_1): Likewise.
3716 (const_binop): Likewise.
3717 (fold_convert_const_int_from_int): Likewise.
3718 (fold_convert_const_int_from_real): Likewise.
3719 (fold_convert_const_int_from_fixed): Likewise.
3720 (fold_convert_const_fixed_from_int): Likewise.
3721 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
3722 (sign_bit_p): Use wide-int interfaces.
3723 (make_range_step): Likewise.
3724 (build_range_check): Likewise. Pass an integer of the correct type
3725 instead of using integer_one_node.
3726 (range_predecessor): Pass an integer of the correct type instead
3727 of using integer_one_node.
3728 (range_successor): Likewise.
3729 (merge_ranges): Likewise.
3730 (unextend): Use wide-int interfaces.
3731 (extract_muldiv_1): Likewise.
3732 (fold_div_compare): Likewise.
3733 (fold_single_bit_test): Likewise.
3734 (fold_sign_changed_comparison): Likewise.
3735 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3736 (fold_plusminus_mult_expr): Use wide-int interfaces.
3737 (native_encode_int): Likewise.
3738 (native_interpret_int): Likewise.
3739 (fold_unary_loc): Likewise.
3740 (pointer_may_wrap_p): Likewise.
3741 (size_low_cst): Likewise.
3742 (mask_with_tz): Likewise.
3743 (fold_binary_loc): Likewise.
3744 (fold_ternary_loc): Likewise.
3745 (multiple_of_p): Likewise.
3746 (tree_call_nonnegative_warnv_p): Update calls to
3747 tree_int_cst_min_precision and real_from_integer.
3748 (fold_negate_const): Use wide-int interfaces.
3749 (fold_abs_const): Likewise.
3750 (fold_relational_const): Use tree_int_cst_lt.
3751 (round_up_loc): Use wide-int interfaces.
3752 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3753 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3754 * gengtype.c: Remove include of double-int.h.
3755 (do_typedef): Use wide-int interfaces.
3756 (open_base_files): Add wide-int.h.
3757 (main): Add offset_int and widest_int typedefs.
3758 * gengtype-lex.l: Handle "^".
3759 (CXX_KEYWORD): Add "static".
3760 * gengtype-parse.c (require3): New.
3761 (require_template_declaration): Handle constant template arguments
3762 and nested templates.
3763 * gengtype-state.c: Don't include "double-int.h".
3764 * genpreds.c (write_one_predicate_function): Update comment.
3765 (write_tm_constrs_h): Add check for hval and lval use in
3767 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3768 (add_to_sequence): Likewise.
3769 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3770 and const_double_operand.
3771 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3773 * gimple-fold.c (get_base_constructor): Likewise.
3774 (fold_array_ctor_reference): Likewise.
3775 (fold_nonarray_ctor_reference): Likewise.
3776 (fold_const_aggregate_ref_1): Likewise.
3777 (gimple_val_nonnegative_real_p): Likewise.
3778 (gimple_fold_indirect_ref): Likewise.
3779 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3780 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
3781 (struct slsr_cand_d): Change index to be widest_int.
3782 (struct incr_info_d): Change incr to be widest_int.
3783 (alloc_cand_and_find_basis): Use wide-int interfaces.
3784 (slsr_process_phi): Likewise.
3785 (backtrace_base_for_ref): Likewise. Return a widest_int.
3786 (restructure_reference): Take a widest_int instead of a double_int.
3787 (slsr_process_ref): Use wide-int interfaces.
3788 (create_mul_ssa_cand): Likewise.
3789 (create_mul_imm_cand): Likewise.
3790 (create_add_ssa_cand): Likewise.
3791 (create_add_imm_cand): Take a widest_int instead of a double_int.
3792 (slsr_process_add): Use wide-int interfaces.
3793 (slsr_process_cast): Likewise.
3794 (slsr_process_copy): Likewise.
3795 (dump_candidate): Likewise.
3796 (dump_incr_vec): Likewise.
3797 (replace_ref): Likewise.
3798 (cand_increment): Likewise. Return a widest_int.
3799 (cand_abs_increment): Likewise.
3800 (replace_mult_candidate): Take a widest_int instead of a double_int.
3801 (replace_unconditional_candidate): Use wide-int interfaces.
3802 (incr_vec_index): Take a widest_int instead of a double_int.
3803 (create_add_on_incoming_edge): Likewise.
3804 (create_phi_basis): Use wide-int interfaces.
3805 (replace_conditional_candidate): Likewise.
3806 (record_increment): Take a widest_int instead of a double_int.
3807 (record_phi_increments): Use wide-int interfaces.
3808 (phi_incr_cost): Take a widest_int instead of a double_int.
3809 (lowest_cost_path): Likewise.
3810 (total_savings): Likewise.
3811 (analyze_increments): Use wide-int interfaces.
3812 (ncd_with_phi): Take a widest_int instead of a double_int.
3813 (ncd_of_cand_and_phis): Likewise.
3814 (nearest_common_dominator_for_cands): Likewise.
3815 (insert_initializers): Use wide-int interfaces.
3816 (all_phi_incrs_profitable): Likewise.
3817 (replace_one_candidate): Likewise.
3818 (replace_profitable_candidates): Likewise.
3819 * godump.c: Include wide-int-print.h.
3820 (go_output_typedef): Use wide-int interfaces.
3821 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
3822 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
3823 (build_loop_iteration_domains): Likewise.
3824 * hooks.h: Include wide-int.h rather than double-int.h.
3825 (hook_bool_dint_dint_uint_bool_true): Delete.
3826 (hook_bool_wint_wint_uint_bool_true): Declare.
3827 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
3828 (hook_bool_wint_wint_uint_bool_true): New.
3829 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
3831 (ubsan_expand_si_overflow_mul_check): Likewise.
3832 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
3833 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
3834 (get_ancestor_addr_info): Likewise.
3835 (ipa_modify_call_arguments): Likewise.
3836 * loop-doloop.c (doloop_modify): Likewise.
3837 (doloop_optimize): Likewise.
3838 * loop-iv.c (iv_number_of_iterations): Likewise.
3839 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
3840 (unroll_loop_constant_iterations): Likewise.
3841 (decide_unroll_runtime_iterations): Likewise.
3842 (unroll_loop_runtime_iterations): Likewise.
3843 (decide_peel_simple): Likewise.
3844 (decide_unroll_stupid): Likewise.
3845 * lto-streamer-in.c (streamer_read_wi): Add.
3846 (input_cfg): Use wide-int interfaces.
3847 (lto_input_tree_1): Likewise.
3848 * lto-streamer-out.c (streamer_write_wi): Add.
3849 (hash_tree): Use wide-int interfaces.
3850 (output_cfg): Likewise.
3851 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
3852 (GTFILES): Add wide-int.h and signop.h.
3853 (TAGS): Look for .cc files too.
3854 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
3855 * optabs.c (expand_subword_shift): Likewise.
3856 (expand_doubleword_shift): Likewise.
3857 (expand_absneg_bit): Likewise.
3858 (expand_copysign_absneg): Likewise.
3859 (expand_copysign_bit): Likewise.
3860 * postreload.c (reload_cse_simplify_set): Likewise.
3861 * predict.c (predict_iv_comparison): Likewise.
3862 * pretty-print.h: Include wide-int-print.h.
3864 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
3865 * print-tree.c: Include wide-int-print.h.
3866 (print_node_brief): Use wide-int interfaces.
3867 (print_node): Likewise.
3868 * read-rtl.c (validate_const_wide_int): New.
3869 (read_rtx_code): Add CONST_WIDE_INT case.
3870 * real.c: Include wide-int.h.
3871 (real_to_integer2): Delete.
3872 (real_to_integer): New function, returning a wide_int.
3873 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3874 (ten_to_ptwo): Update call to real_from_integer.
3875 (real_digit): Likewise.
3876 * real.h: Include signop.h, wide-int.h and insn-modes.h.
3877 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
3878 (REAL_VALUE_TO_INT): Delete.
3879 (real_to_integer): Declare a wide-int form.
3880 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3881 * recog.c (const_int_operand): Improve comment.
3882 (const_scalar_int_operand): New.
3883 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
3884 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
3885 (split_double): Likewise.
3886 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
3887 (rtx_size): Likewise.
3888 (rtx_alloc_stat_v): New.
3889 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
3890 (cwi_output_hex): New.
3891 (iterative_hash_rtx): Handle CONST_WIDE_INT.
3892 (cwi_check_failed_bounds): New.
3893 * rtl.def (CONST_WIDE_INT): New.
3894 * rtl.h: Include <utility> and wide-int.h.
3895 (struct hwivec_def): New.
3896 (CWI_GET_NUM_ELEM): New.
3897 (CWI_PUT_NUM_ELEM): New.
3898 (struct rtx_def): Add num_elem and hwiv.
3899 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
3900 (CASE_CONST_UNIQUE): Likewise.
3901 (CASE_CONST_ANY): Likewise.
3902 (CONST_SCALAR_INT_P): Likewise.
3903 (CONST_WIDE_INT_P): New.
3905 (HWIVEC_CHECK): New.
3906 (cwi_check_failed_bounds): New.
3908 (HWIVEC_CHECK): New.
3909 (CONST_WIDE_INT_VEC) New.
3910 (CONST_WIDE_INT_NUNITS) New.
3911 (CONST_WIDE_INT_ELT) New.
3912 (rtx_mode_t): New type.
3913 (wi::int_traits <rtx_mode_t>): New.
3915 (wi::min_value): New.
3916 (wi::max_value): New.
3918 (const_wide_int_alloc): New.
3919 (immed_wide_int_const): New.
3920 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
3921 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
3922 * signop.h: New file.
3923 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
3924 (simplify_const_unary_operation): Use wide-int interfaces.
3925 (simplify_binary_operation_1): Likewise.
3926 (simplify_const_binary_operation): Likewise.
3927 (simplify_const_relational_operation): Likewise.
3928 (simplify_immed_subreg): Likewise.
3929 * stmt.c (expand_case): Likewise.
3930 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
3931 signop rather than a bool.
3932 * stor-layout.c (layout_type): Use wide-int interfaces.
3933 (initialize_sizetypes): Update calls to
3934 set_min_and_max_values_for_integral_type.
3935 (set_min_and_max_values_for_integral_type): Take a signop rather
3936 than a bool. Use wide-int interfaces.
3937 (fixup_signed_type): Update accordingly. Remove
3938 HOST_BITS_PER_DOUBLE_INT limit.
3939 (fixup_unsigned_type): Likewise.
3940 * system.h (STATIC_CONSTANT_P): New.
3941 (STATIC_ASSERT): New.
3942 * target.def (can_use_doloop_p): Take widest_ints rather than
3944 * target.h: Include wide-int.h rather than double-int.h.
3945 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
3947 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
3948 rather than INT_CST_LT_UNSIGNED.
3949 (can_use_doloop_if_innermost): Take widest_ints rather than
3951 * tree-affine.c: Include wide-int-print.h.
3952 (double_int_ext_for_comb): Delete.
3953 (wide_int_ext_for_comb): New.
3954 (aff_combination_zero): Use wide-int interfaces.
3955 (aff_combination_const): Take a widest_int instead of a double_int.
3956 (aff_combination_elt): Use wide-int interfaces.
3957 (aff_combination_scale): Take a widest_int instead of a double_int.
3958 (aff_combination_add_elt): Likewise.
3959 (aff_combination_add_cst): Likewise.
3960 (aff_combination_add): Use wide-int interfaces.
3961 (aff_combination_convert): Likewise.
3962 (tree_to_aff_combination): Likewise.
3963 (add_elt_to_tree): Take a widest_int instead of a double_int.
3964 (aff_combination_to_tree): Use wide-int interfaces.
3965 (aff_combination_remove_elt): Likewise.
3966 (aff_combination_add_product): Take a widest_int instead of
3968 (aff_combination_mult): Use wide-int interfaces.
3969 (aff_combination_expand): Likewise.
3970 (double_int_constant_multiple_p): Delete.
3971 (wide_int_constant_multiple_p): New.
3972 (aff_combination_constant_multiple_p): Take a widest_int pointer
3973 instead of a double_int pointer.
3974 (print_aff): Use wide-int interfaces.
3975 (get_inner_reference_aff): Take a widest_int pointer
3976 instead of a double_int pointer.
3977 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
3978 * tree-affine.h: Include wide-int.h.
3979 (struct aff_comb_elt): Change type of coef to widest_int.
3980 (struct affine_tree_combination): Change type of offset to widest_int.
3981 (double_int_ext_for_comb): Delete.
3982 (wide_int_ext_for_comb): New.
3983 (aff_combination_const): Use widest_int instead of double_int.
3984 (aff_combination_scale): Likewise.
3985 (aff_combination_add_elt): Likewise.
3986 (aff_combination_constant_multiple_p): Likewise.
3987 (get_inner_reference_aff): Likewise.
3988 (aff_comb_cannot_overlap_p): Likewise.
3989 (aff_combination_zero_p): Use wide-int interfaces.
3990 * tree.c: Include tree.h.
3991 (init_ttree): Use make_int_cst.
3992 (tree_code_size): Removed code for INTEGER_CST case.
3993 (tree_size): Add INTEGER_CST case.
3994 (make_node_stat): Update comment.
3995 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
3996 (build_int_cst_type): Use wide-int interfaces.
3997 (double_int_to_tree): Likewise.
3998 (double_int_fits_to_tree_p): Delete.
3999 (force_fit_type_double): Delete.
4000 (force_fit_type): New.
4001 (int_cst_hash_hash): Use wide-int interfaces.
4002 (int_cst_hash_eq): Likewise.
4003 (build_int_cst_wide): Delete.
4004 (wide_int_to_tree): New.
4005 (cache_integer_cst): Use wide-int interfaces.
4006 (build_low_bits_mask): Likewise.
4007 (cst_and_fits_in_hwi): Likewise.
4008 (real_value_from_int_cst): Likewise.
4009 (make_int_cst_stat): New.
4010 (integer_zerop): Use wide_int interfaces.
4011 (integer_onep): Likewise.
4012 (integer_all_onesp): Likewise.
4013 (integer_pow2p): Likewise.
4014 (integer_nonzerop): Likewise.
4015 (tree_log2): Likewise.
4016 (tree_floor_log2): Likewise.
4017 (tree_ctz): Likewise.
4018 (int_size_in_bytes): Likewise.
4019 (mem_ref_offset): Return an offset_int rather than a double_int.
4020 (build_type_attribute_qual_variant): Use wide_int interfaces.
4021 (type_hash_eq): Likewise
4022 (tree_int_cst_equal): Likewise.
4023 (tree_int_cst_lt): Delete.
4024 (tree_int_cst_compare): Likewise.
4025 (tree_fits_shwi_p): Use wide_int interfaces.
4026 (tree_fits_uhwi_p): Likewise.
4027 (tree_int_cst_sign_bit): Likewise.
4028 (tree_int_cst_sgn): Likewise.
4029 (tree_int_cst_min_precision): Take a signop rather than a bool.
4030 (simple_cst_equal): Use wide_int interfaces.
4031 (compare_tree_int): Likewise.
4032 (iterative_hash_expr): Likewise.
4033 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
4035 (get_type_static_bounds): Use wide_int interfaces.
4036 (tree_int_cst_elt_check_failed): New.
4037 (build_common_tree_nodes): Reordered to set prec before filling in
4039 (int_cst_value): Check cst_and_fits_in_hwi.
4040 (widest_int_cst_value): Use wide_int interfaces.
4041 (upper_bound_in_type): Likewise.
4042 (lower_bound_in_type): Likewise.
4043 (num_ending_zeros): Likewise.
4044 (drop_tree_overflow): Likewise.
4045 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
4046 (gen_conditions_for_pow_cst_base): Likewise.
4047 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
4048 (group_case_labels_stmt): Use wide-int interfaces.
4049 (verify_gimple_assign_binary): Likewise.
4050 (print_loop): Likewise.
4051 * tree-chrec.c (tree_fold_binomial): Likewise.
4052 * tree-core.h (struct tree_base): Add int_length.
4053 (struct tree_int_cst): Change rep of value.
4054 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
4055 (dr_may_alias_p): Likewise.
4056 (max_stmt_executions_tree): Likewise.
4057 * tree.def (INTEGER_CST): Update comment.
4058 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
4059 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
4060 * tree-dump.c: Include wide-int.h and wide-int-print.h.
4061 (dequeue_and_dump): Use wide-int interfaces.
4062 * tree.h: Include wide-int.h.
4063 (NULL_TREE): Moved to earlier loc in file.
4064 (TREE_INT_CST_ELT_CHECK): New.
4065 (tree_int_cst_elt_check_failed): New.
4067 (TREE_INT_CST): Delete.
4068 (TREE_INT_CST_LOW): Use wide-int interfaces.
4069 (TREE_INT_CST_HIGH): Delete.
4070 (TREE_INT_CST_NUNITS): New.
4071 (TREE_INT_CST_EXT_NUNITS): Likewise.
4072 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
4073 (TREE_INT_CST_ELT): Likewise.
4074 (INT_CST_LT): Delete.
4075 (tree_int_cst_elt_check): New (two forms).
4076 (type_code_size): Update comment.
4077 (make_int_cst_stat, make_int_cst): New.
4078 (tree_to_double_int): Delete.
4079 (double_int_fits_to_tree_p): Delete.
4080 (force_fit_type_double): Delete.
4081 (build_int_cstu): Replace with out-of-line function.
4082 (build_int_cst_wide): Delete.
4083 (tree_int_cst_lt): Define inline.
4084 (tree_int_cst_le): New.
4085 (tree_int_cst_compare): Define inline.
4086 (tree_int_cst_min_precision): Take a signop rather than a bool.
4087 (wi::int_traits <const_tree>): New.
4088 (wi::int_traits <tree>): New.
4089 (wi::extended_tree): New.
4090 (wi::int_traits <wi::extended_tree>): New.
4091 (wi::to_widest): New.
4092 (wi::to_offset): New.
4093 (wi::fits_to_tree_p): New.
4094 (wi::min_value): New.
4095 (wi::max_value): New.
4096 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
4097 (copy_tree_body_r): Likewise.
4098 * tree-object-size.c (compute_object_offset): Likewise.
4099 (addr_object_size): Likewise.
4100 * tree-predcom.c: Include wide-int-print.h.
4101 (struct dref_d): Change type of offset to widest_int.
4102 (dump_dref): Call wide-int printer.
4103 (aff_combination_dr_offset): Use wide-int interfaces.
4104 (determine_offset): Take a widest_int pointer rather than a
4106 (split_data_refs_to_components): Use wide-int interfaces.
4107 (suitable_component_p): Likewise.
4108 (order_drefs): Likewise.
4109 (add_ref_to_chain): Likewise.
4110 (valid_initializer_p): Likewise.
4111 (determine_roots_comp): Likewise.
4112 * tree-pretty-print.c: Include wide-int-print.h.
4113 (dump_generic_node): Use wide-int interfaces.
4114 * tree-sra.c (sra_ipa_modify_expr): Likewise.
4115 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
4116 (move_fixed_address_to_symbol): Likewise.
4117 (move_hint_to_base): Likewise.
4118 (move_pointer_to_base): Likewise.
4119 (move_variant_to_index): Likewise.
4120 (most_expensive_mult_to_index): Likewise.
4121 (addr_to_parts): Likewise.
4122 (copy_ref_info): Likewise.
4123 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
4124 (indirect_refs_may_alias_p): Likewise.
4125 (stmt_kills_ref_p_1): Likewise.
4126 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
4127 * tree-ssa-ccp.c: Update comment at top of file. Include
4129 (struct prop_value_d): Change type of mask to widest_int.
4130 (extend_mask): New function.
4131 (dump_lattice_value): Use wide-int interfaces.
4132 (get_default_value): Likewise.
4133 (set_constant_value): Likewise.
4134 (set_value_varying): Likewise.
4135 (valid_lattice_transition): Likewise.
4136 (set_lattice_value): Likewise.
4137 (value_to_double_int): Delete.
4138 (value_to_wide_int): New.
4139 (get_value_from_alignment): Use wide-int interfaces.
4140 (get_value_for_expr): Likewise.
4141 (do_dbg_cnt): Likewise.
4142 (ccp_finalize): Likewise.
4143 (ccp_lattice_meet): Likewise.
4144 (bit_value_unop_1): Use widest_ints rather than double_ints.
4145 (bit_value_binop_1): Likewise.
4146 (bit_value_unop): Use wide-int interfaces.
4147 (bit_value_binop): Likewise.
4148 (bit_value_assume_aligned): Likewise.
4149 (evaluate_stmt): Likewise.
4150 (ccp_fold_stmt): Likewise.
4151 (visit_cond_stmt): Likewise.
4152 (ccp_visit_stmt): Likewise.
4153 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
4154 (constant_pointer_difference): Likewise.
4155 (associate_pointerplus): Likewise.
4156 (combine_conversions): Likewise.
4157 * tree-ssa-loop.h: Include wide-int.h.
4158 (struct tree_niter_desc): Change type of max to widest_int.
4159 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
4160 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
4161 (remove_redundant_iv_tests): Likewise.
4162 (canonicalize_loop_induction_variables): Likewise.
4163 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
4164 (constant_multiple_of): Take a widest_int pointer instead of
4165 a double_int pointer.
4166 (get_computation_aff): Use wide-int interfaces.
4167 (ptr_difference_cost): Likewise.
4168 (difference_cost): Likewise.
4169 (get_loop_invariant_expr_id): Likewise.
4170 (get_computation_cost_at): Likewise.
4171 (iv_elimination_compare_lt): Likewise.
4172 (may_eliminate_iv): Likewise.
4173 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
4174 instead of double_int.
4175 (max_loop_iterations): Likewise.
4176 (max_stmt_executions): Likewise.
4177 (estimated_stmt_executions): Likewise.
4178 * tree-ssa-loop-niter.c: Include wide-int-print.h.
4179 (split_to_var_and_offset): Use wide-int interfaces.
4180 (determine_value_range): Likewise.
4181 (bound_difference_of_offsetted_base): Likewise.
4182 (bounds_add): Take a widest_int instead of a double_int.
4183 (number_of_iterations_ne_max): Use wide-int interfaces.
4184 (number_of_iterations_ne): Likewise.
4185 (number_of_iterations_lt_to_ne): Likewise.
4186 (assert_loop_rolls_lt): Likewise.
4187 (number_of_iterations_lt): Likewise.
4188 (number_of_iterations_le): Likewise.
4189 (number_of_iterations_cond): Likewise.
4190 (number_of_iterations_exit): Likewise.
4191 (finite_loop_p): Likewise.
4192 (derive_constant_upper_bound_assign): Likewise.
4193 (derive_constant_upper_bound): Return a widest_int.
4194 (derive_constant_upper_bound_ops): Likewise.
4195 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
4196 (record_estimate): Take a widest_int rather than a double_int.
4197 (record_nonwrapping_iv): Use wide-int interfaces.
4198 (double_int_cmp): Delete.
4199 (wide_int_cmp): New.
4200 (bound_index): Take a widest_int rather than a double_int.
4201 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
4202 (maybe_lower_iteration_bound): Likewise.
4203 (estimate_numbers_of_iterations_loop): Likewise.
4204 (estimated_loop_iterations): Take a widest_int pointer than than
4205 a double_int pointer.
4206 (estimated_loop_iterations_int): Use wide-int interfaces.
4207 (max_loop_iterations): Take a widest_int pointer than than
4208 a double_int pointer.
4209 (max_loop_iterations_int): Use wide-int interfaces.
4210 (max_stmt_executions): Take a widest_int pointer than than
4211 a double_int pointer.
4212 (estimated_stmt_executions): Likewise.
4213 (n_of_executions_at_most): Use wide-int interfaces.
4214 (scev_probably_wraps_p): Likewise.
4215 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
4217 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
4219 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
4220 double_ints. Adjust for trailing_wide_ints <3> representation.
4221 (set_nonzero_bits): Likewise.
4222 (get_range_info): Return wide_ints rather than double_ints.
4223 Adjust for trailing_wide_ints <3> representation.
4224 (get_nonzero_bits): Likewise.
4225 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
4227 * tree-ssanames.h (struct range_info_def): Replace min, max and
4228 nonzero_bits with a trailing_wide_ints <3>.
4229 (set_range_info): Use wide_int_refs rather than double_ints.
4230 (set_nonzero_bits): Likewise.
4231 (get_range_info): Return wide_ints rather than double_ints.
4232 (get_nonzero_bits): Likewise.
4233 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
4234 * tree-ssa-pre.c (phi_translate_1): Likewise.
4235 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
4236 (acceptable_pow_call): Likewise.
4237 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
4239 (vn_reference_fold_indirect): Likewise.
4240 (vn_reference_maybe_forwprop_address): Likewise.
4241 (valueize_refs_1): Likewise.
4242 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
4243 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
4244 tree_int_cst_lt and tree_int_cst_le.
4245 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
4247 (streamer_alloc_tree): Likewise.
4248 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
4249 (streamer_write_tree_header): Likewise.
4250 (streamer_write_integer_cst): Likewise.
4251 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
4252 (build_constructors): Likewise.
4253 (array_value_type): Likewise.
4254 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
4255 (vect_check_gather): Likewise.
4256 * tree-vect-generic.c (build_replicated_const): Likewise.
4257 (expand_vector_divmod): Likewise.
4258 * tree-vect-loop.c (vect_transform_loop): Likewise.
4259 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
4260 (vect_do_peeling_for_alignment): Likewise.
4261 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
4262 * tree-vrp.c: Include wide-int.h.
4263 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
4264 (extract_range_from_assert): Use wide-int interfaces.
4265 (vrp_int_const_binop): Likewise.
4266 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
4267 double_int pointers.
4268 (ranges_from_anti_range): Use wide-int interfaces.
4269 (quad_int_cmp): Delete.
4270 (quad_int_pair_sort): Likewise.
4271 (extract_range_from_binary_expr_1): Use wide-int interfaces.
4272 (extract_range_from_unary_expr_1): Likewise.
4273 (adjust_range_with_scev): Likewise.
4274 (masked_increment): Take and return wide_ints rather than double_ints.
4275 (register_edge_assert_for_2): Use wide-int interfaces.
4276 (check_array_ref): Likewise.
4277 (search_for_addr_array): Likewise.
4278 (maybe_set_nonzero_bits): Likewise.
4279 (union_ranges): Pass an integer of the correct type instead of
4280 using integer_one_node.
4281 (intersect_ranges): Likewise.
4282 (simplify_truth_ops_using_ranges): Likewise.
4283 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
4284 (range_fits_type_p): Likewise.
4285 (simplify_cond_using_ranges): Likewise. Take a signop rather than
4287 (simplify_conversion_using_ranges): Use wide-int interfaces.
4288 (simplify_float_conversion_using_ranges): Likewise.
4289 (vrp_finalize): Likewise.
4290 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
4291 (gimple_stringops_transform): Likewise.
4292 * varasm.c (decode_addr_const): Likewise.
4293 (const_hash_1): Likewise.
4294 (const_rtx_hash_1): Likewise
4295 (output_constant): Likewise.
4296 (array_size_for_constructor): Likewise.
4297 (output_constructor_regular_field): Likewise.
4298 (output_constructor_bitfield): Likewise.
4299 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
4300 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
4302 * gencheck.c: Define BITS_PER_UNIT.
4305 * wide-int-print.cc: New.
4306 * wide-int-print.h: New.
4308 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4310 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
4312 2014-05-06 Richard Biener <rguenther@suse.de>
4314 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
4315 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
4316 (TODO_verify_all): Adjust.
4317 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
4318 TODO_verify_stmts and TODO_verify_rtl_sharing.
4319 * bb-reorder.c: Likewise.
4320 * cfgexpand.c: Likewise.
4321 * cprop.c: Likewise.
4323 * function.c: Likewise.
4324 * fwprop.c: Likewise.
4326 * gimple-ssa-isolate-paths.c: Likewise.
4327 * gimple-ssa-strength-reduction.c: Likewise.
4328 * ipa-split.c: Likewise.
4329 * loop-init.c: Likewise.
4330 * loop-unroll.c: Likewise.
4331 * lower-subreg.c: Likewise.
4332 * modulo-sched.c: Likewise.
4333 * postreload-gcse.c: Likewise.
4334 * predict.c: Likewise.
4335 * recog.c: Likewise.
4336 * sched-rgn.c: Likewise.
4337 * store-motion.c: Likewise.
4338 * tracer.c: Likewise.
4339 * trans-mem.c: Likewise.
4340 * tree-call-cdce.c: Likewise.
4341 * tree-cfg.c: Likewise.
4342 * tree-cfgcleanup.c: Likewise.
4343 * tree-complex.c: Likewise.
4344 * tree-eh.c: Likewise.
4345 * tree-emutls.c: Likewise.
4346 * tree-if-conv.c: Likewise.
4347 * tree-into-ssa.c: Likewise.
4348 * tree-loop-distribution.c: Likewise.
4349 * tree-object-size.c: Likewise.
4350 * tree-parloops.c: Likewise.
4351 * tree-pass.h: Likewise.
4352 * tree-sra.c: Likewise.
4353 * tree-ssa-ccp.c: Likewise.
4354 * tree-ssa-copy.c: Likewise.
4355 * tree-ssa-copyrename.c: Likewise.
4356 * tree-ssa-dce.c: Likewise.
4357 * tree-ssa-dom.c: Likewise.
4358 * tree-ssa-dse.c: Likewise.
4359 * tree-ssa-forwprop.c: Likewise.
4360 * tree-ssa-ifcombine.c: Likewise.
4361 * tree-ssa-loop-ch.c: Likewise.
4362 * tree-ssa-loop-ivcanon.c: Likewise.
4363 * tree-ssa-loop.c: Likewise.
4364 * tree-ssa-math-opts.c: Likewise.
4365 * tree-ssa-phiopt.c: Likewise.
4366 * tree-ssa-phiprop.c: Likewise.
4367 * tree-ssa-pre.c: Likewise.
4368 * tree-ssa-reassoc.c: Likewise.
4369 * tree-ssa-sink.c: Likewise.
4370 * tree-ssa-strlen.c: Likewise.
4371 * tree-ssa-tail-merge.c: Likewise.
4372 * tree-ssa-uncprop.c: Likewise.
4373 * tree-switch-conversion.c: Likewise.
4374 * tree-tailcall.c: Likewise.
4375 * tree-vect-generic.c: Likewise.
4376 * tree-vectorizer.c: Likewise.
4377 * tree-vrp.c: Likewise.
4379 * var-tracking.c: Likewise.
4380 * bt-load.c: Likewise.
4381 * cfgcleanup.c: Likewise.
4382 * combine-stack-adj.c: Likewise.
4383 * combine.c: Likewise.
4384 * compare-elim.c: Likewise.
4385 * config/epiphany/resolve-sw-modes.c: Likewise.
4386 * config/i386/i386.c: Likewise.
4387 * config/mips/mips.c: Likewise.
4388 * config/s390/s390.c: Likewise.
4389 * config/sh/sh_treg_combine.cc: Likewise.
4390 * config/sparc/sparc.c: Likewise.
4393 * final.c: Likewise.
4394 * ifcvt.c: Likewise.
4395 * mode-switching.c: Likewise.
4396 * passes.c: Likewise.
4397 * postreload.c: Likewise.
4399 * reg-stack.c: Likewise.
4400 * regcprop.c: Likewise.
4401 * regrename.c: Likewise.
4404 2014-05-06 Richard Biener <rguenther@suse.de>
4407 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
4408 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
4410 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
4413 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
4415 2014-05-05 Radovan Obradovic <robradovic@mips.com>
4416 Tom de Vries <tom@codesourcery.com>
4418 * target.def (call_fusage_contains_non_callee_clobbers): New
4420 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
4422 (@node Miscellaneous Register Hooks): New node.
4423 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
4424 * doc/tm.texi: Regenerate.
4426 2014-05-05 Marek Polacek <polacek@redhat.com>
4429 * opts.c (common_handle_option): Call error_at instead of warning_at.
4431 2014-05-05 Richard Biener <rguenther@suse.de>
4433 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
4434 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
4435 under the TODO_verify_il umbrella.
4437 2014-05-05 Richard Biener <rguenther@suse.de>
4439 * passes.c (execute_function_todo): Move TODO_verify_flow under
4440 the TODO_verify_ul umbrella.
4442 2014-05-05 Richard Biener <rguenther@suse.de>
4445 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
4446 X & CST away from a CST that is the mask of a mode.
4448 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4450 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
4451 int argument to enum machine_mode.
4452 (picochip_class_max_nregs): Ditto.
4453 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
4454 (picochip_class_max_nregs): Ditto.
4456 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4458 * target.def: Add new target hook.
4459 * doc/tm.texi: Regenerate.
4460 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
4461 * targhooks.c (default_keep_leaf_when_profiled): New function.
4463 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
4464 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
4466 2014-05-05 Bin Cheng <bin.cheng@arm.com>
4468 PR tree-optimization/60363
4469 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
4470 (copy_phi_args): New parameters. Call get_value_locus_in_path.
4471 (update_destination_phis): New parameter.
4472 (create_edge_and_update_destination_phis): Ditto.
4473 (ssa_fix_duplicate_block_edges): Pass new arguments.
4474 (thread_single_edge): Ditto.
4476 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
4478 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
4479 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
4480 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
4481 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
4482 Use RS6000_BTM_HARD_FLOAT.
4483 (BU_MISC_2): Likewise.
4484 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
4485 RS6000_BTM_HARD_FLOAT.
4486 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
4488 (rs6000_invalid_builtin): Add hard floating builtin support.
4489 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
4490 hard float builtins.
4491 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
4493 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4495 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
4496 Add missing function* argument.
4498 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4500 * lra-constraints.c (valid_address_p): Move earlier in file.
4501 Add a constraint argument to the address_info version.
4502 (satisfies_memory_constraint_p): New function.
4503 (satisfies_address_constraint_p): Likewise.
4504 (process_alt_operands, curr_insn_transform): Use them.
4505 (process_address): Pass the constraint to valid_address_p when
4506 checking address operands.
4508 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4510 * config/mips/mips.c (mips_isa_rev): New variable.
4511 (mips_set_architecture): Set it.
4512 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
4514 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
4515 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
4516 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
4517 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
4518 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
4519 conditions in terms of mips_isa_rev.
4520 (mips_isa_rev): Declare.
4522 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4524 * config/sh/sh-mem.cc: Use tabs instead of spaces.
4525 (prob_unlikely, prob_likely): Make variables const.
4527 2014-05-03 Denis Chertykov <chertykov@gmail.com>
4529 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
4531 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4533 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
4535 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4537 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
4538 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
4539 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
4541 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
4542 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
4544 Replace usage of ROUND_REG with sh_round_reg.
4545 Use CEIL instead of ROUND_ADVANCE.
4547 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4550 * config/sh/sh.c: Include stdlib headers before everything else.
4552 2014-05-02 Jakub Jelinek <jakub@redhat.com>
4554 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
4555 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
4556 (gimplify_adjust_omp_clauses): Simd region is never
4557 directly nested in combined parallel. Instead, for linear
4558 with copyin/copyout, if in combined for simd loop, make decl
4559 firstprivate/lastprivate on OMP_FOR.
4560 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
4561 expand_omp_for_static_chunk): When setting endvar, also set
4562 fd->loop.v to the same value.
4564 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4566 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
4568 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
4570 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
4573 2014-05-02 Marek Polacek <polacek@redhat.com>
4575 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
4577 2014-05-02 Kito Cheng <kito@0xlab.org>
4579 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
4580 to a C expression marco.
4581 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
4582 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
4583 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
4584 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
4585 HONOR_REG_ALLOC_ORDER.
4586 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
4588 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4590 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
4592 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4594 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
4596 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
4598 * tree-if-conv.c (is_cond_scalar_reduction): New function.
4599 (convert_scalar_cond_reduction): Likewise.
4600 (predicate_scalar_phi): Add recognition and transformation
4601 of simple conditioanl reduction to be vectorizable.
4603 2014-05-01 Marek Polacek <polacek@redhat.com>
4606 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
4608 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
4610 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
4611 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
4612 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
4613 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
4614 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
4615 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
4616 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
4617 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
4619 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
4621 * config/arc/arc.opt (mlra): Move comment above option name
4622 to avoid mis-parsing as language options.
4624 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4626 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
4627 * config/sol2.h: ... here.
4628 * config/sol2-10.h: Remove.
4630 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
4631 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
4632 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
4633 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
4634 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
4635 * config/sol2.h: ... here.
4636 (SECTION_NAME_FORMAT): Don't redefine.
4637 (STARTFILE_ARCH32_SPEC): Rename to ...
4638 (STARTFILE_ARCH_SPEC): ... this.
4639 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
4640 * config/sparc/sol2.h: ... here.
4641 (SECTION_NAME_FORMAT): Don't undef.
4642 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
4643 (SUBTARGET_EXTRA_SPECS): Remove.
4644 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
4646 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
4647 (MD_STARTFILE_PREFIX): Remove.
4648 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
4649 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
4650 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
4651 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
4652 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
4653 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
4654 * config/i386/sol2.h: ... here.
4655 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
4656 * config/i386/sol2-bi.h: Remove.
4657 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
4658 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
4660 * config/i386/t-sol2-64: Rename to ...
4661 * config/i386/t-sol2: ... this.
4662 * config/sparc/t-sol2-64: Rename to ...
4663 * config/sparc/t-sol2: ... this.
4665 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
4666 sol2_tm_file_head, sol2_tm_file_tail.
4667 Include ${cpu_type}/sol2.h before sol2.h.
4669 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
4670 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
4671 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
4672 Reflect i386/t-sol2-64 renaming.
4673 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
4674 Reflect sparc/t-sol2-64 renaming.
4676 2014-04-30 Richard Biener <rguenther@suse.de>
4678 * passes.c (execute_function_todo): Move TODO_verify_stmts
4679 and TODO_verify_ssa under the TODO_verify_il umbrella.
4680 * tree-ssa.h (verify_ssa): Adjust prototype.
4681 * tree-ssa.c (verify_ssa): Add parameter to tell whether
4682 we should verify SSA operands.
4683 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
4684 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
4685 whether we should verify whether not throwing stmts have EH info.
4686 * graphite-scop-detection.c (create_sese_edges): Adjust.
4687 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
4688 * tree-eh.c (lower_try_finally_switch): Do not add the
4689 default case label twice.
4691 2014-04-30 Marek Polacek <polacek@redhat.com>
4693 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
4694 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
4695 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
4696 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
4698 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
4700 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
4701 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
4702 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
4703 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
4704 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
4705 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
4706 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
4707 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
4709 2014-04-29 David Malcolm <dmalcolm@redhat.com>
4711 * tree-cfg.c (dump_function_to_file): Dump the return type of
4712 functions, in a line to itself before the function body, mimicking
4713 the layout of a C function.
4715 2014-04-29 Jakub Jelinek <jakub@redhat.com>
4717 PR tree-optimization/60971
4718 * tree-tailcall.c (process_assignment): Reject conversions which
4721 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4723 * calls.c (initialize_argument_information): Always treat
4724 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4725 (expand_call): Likewise.
4726 (emit_library_call_calue_1): Likewise.
4727 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4728 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4731 2014-04-29 Nick Clifton <nickc@redhat.com>
4733 * config/msp430/msp430.md (umulsidi): Fix typo.
4734 (mulhisi3): Enable even inside interrupt handlers.
4735 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4736 bigger return address pushed in large mode.
4738 2014-04-29 Nick Clifton <nickc@redhat.com>
4740 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4741 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4743 * config/m32r/m32r.c (init_reg_tables): Likewise.
4744 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4745 enum to hold the modes.
4747 2014-04-29 Richard Biener <rguenther@suse.de>
4749 * dominance.c (free_dominance_info): Add overload with
4751 (dom_info_state): Likewise.
4752 (dom_info_available_p): Likewise.
4753 * basic-block.h (free_dominance_info, dom_info_state,
4754 dom_info_available_p): Declare overloads.
4755 * passes.c (execute_function_todo): Verify that verifiers
4756 don't change dominator info state. Drop dominator info
4757 for IPA pass invocations.
4758 * cgraph.c (release_function_body): Restore asserts that
4759 dominator information is released.
4761 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4763 * doc/invoke.texi: Fix typo.
4764 * tree-vrp.c: Fix typos.
4765 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4767 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4769 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4771 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4773 * config/aarch64/aarch64-builtins.c
4774 (aarch64_types_storestruct_lane_qualifiers): New.
4775 (TYPES_STORESTRUCT_LANE): Likewise.
4776 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4777 (st3_lane): Likewise.
4778 (st4_lane): Likewise.
4779 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4780 (vec_store_lanesci_lane<mode>): Likewise.
4781 (vec_store_lanesxi_lane<mode>): Likewise.
4782 (aarch64_st2_lane<VQ:mode>): Likewise.
4783 (aarch64_st3_lane<VQ:mode>): Likewise.
4784 (aarch64_st4_lane<VQ:mode>): Likewise.
4785 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
4786 * config/aarch64/arm_neon.h
4787 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
4788 use new macro arguments.
4789 (__ST3_LANE_FUNC): Likewise.
4790 (__ST4_LANE_FUNC): Likewise.
4791 * config/aarch64/iterators.md (V_TWO_ELEM): New.
4792 (V_THREE_ELEM): Likewise.
4793 (V_FOUR_ELEM): Likewise.
4795 2014-04-28 David Malcolm <dmalcolm@redhat.com>
4797 * doc/gimple.texi: Replace the description of the now-defunct
4798 union gimple_statement_d with a diagram showing the
4799 gimple_statement_base class hierarchy and its relationships to
4800 the GSS_ and GIMPLE_ enums.
4802 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4804 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4805 * config/aarch64/aarch64.c
4806 (aarch64_cannot_change_mode_class): Weaken conditions.
4807 (aarch64_modes_tieable_p): New.
4808 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
4810 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
4812 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4813 (loadsync_<mode>): Change mode.
4814 (load_quadpti, store_quadpti): New.
4815 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4816 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4818 2014-04-28 Martin Jambor <mjambor@suse.cz>
4820 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
4821 same alias type as the original statement.
4822 (subreplacement_assignment_data): New type.
4823 (handle_unscalarized_data_in_subtree): New type of parameter,
4824 generate new memory accesses with same alias type as the original
4826 (load_assign_lhs_subreplacements): Likewise.
4827 (sra_modify_constructor_assign): Generate new memory accesses with
4828 same alias type as the original statement.
4830 2014-04-28 Richard Biener <rguenther@suse.de>
4832 * tree-pass.h (TODO_verify_il): Define.
4833 (TODO_verify_all): Complete properly.
4834 * passes.c (execute_function_todo): Move existing loop-closed
4835 SSA verification under TODO_verify_il.
4836 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
4837 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
4838 Fix tree sharing issue.
4840 2014-04-28 Richard Biener <rguenther@suse.de>
4843 * builtins.def (DEF_C11_BUILTIN): Add.
4844 (BUILT_IN_ALIGNED_ALLOC): Likewise.
4845 * coretypes.h (enum function_class): Add function_c11_misc.
4846 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
4847 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
4848 (call_may_clobber_ref_p_1): Likewise.
4849 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
4850 (mark_all_reaching_defs_necessary_1): Likewise.
4851 (propagate_necessity): Likewise.
4852 (eliminate_unnecessary_stmts): Likewise.
4853 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
4855 2014-04-28 Richard Biener <rguenther@suse.de>
4857 * tree-vrp.c (vrp_var_may_overflow): Remove.
4858 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
4859 with overflow immediately bump to one before that value and
4860 let iteration figure out overflow status.
4862 2014-04-28 Richard Biener <rguenther@suse.de>
4864 * configure.ac: Do valgrind header checks unconditionally.
4865 Add --enable-valgrind-annotations.
4866 * system.h: Guard valgrind header inclusion with
4867 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
4868 * alloc-pool.c (pool_alloc, pool_free): Use
4869 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
4870 to guard possibly dead code.
4871 * config.in: Regenerated.
4872 * configure: Likewise.
4874 2014-04-28 Jeff Law <law@redhat.com>
4876 PR tree-optimization/60902
4877 * tree-ssa-threadedge.c
4878 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
4879 over real defs when invalidating outputs from statements that do not
4880 produce useful outputs for threading.
4882 2014-04-28 Richard Biener <rguenther@suse.de>
4884 PR tree-optimization/60979
4885 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
4886 SCOPs that end in a block with a successor with abnormal
4889 2014-04-28 Richard Biener <rguenther@suse.de>
4891 * tree-pass.h (execute_pass_list): Adjust prototype.
4892 * passes.c (pass_manager::execute_early_local_passes): Adjust.
4893 (do_per_function): Change callback signature, push all actual
4894 work to the callbals.
4895 (do_per_function_toporder): Likewise.
4896 (execute_function_dump): Adjust.
4897 (execute_function_todo): Likewise.
4898 (clear_last_verified): Likewise.
4899 (verify_curr_properties): Likewise.
4900 (update_properties_after_pass): Likewise.
4901 (execute_pass_list_1): Split out from ...
4902 (execute_pass_list): ... here. Adjust.
4903 (execute_ipa_pass_list): Likewise.
4904 * cgraphunit.c (cgraph_add_new_function): Adjust.
4905 (analyze_function): Likewise.
4906 (expand_function): Likewise.
4907 * cgraph.c (release_function_body): Free dominance info
4908 here instead of asserting it was magically freed elsewhere.
4910 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
4912 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
4913 * configure: Regenerate.
4914 * config/sparc/sparc.opt (muser-mode): New option.
4915 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
4917 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
4918 * doc/invoke.texi (SPARC options): Document -muser-mode.
4920 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
4922 * cselib.c (find_slot_memmode): Delete.
4923 (cselib_hasher): Change compare_type to a struct.
4924 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
4926 (preserve_constants_and_equivs): Adjust for new compare_type.
4927 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
4928 (wrap_constant): Delete.
4929 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
4931 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
4933 * doc/install.texi (Building with profile feedback): Remove
4936 2014-04-26 Tom de Vries <tom@codesourcery.com>
4938 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
4941 2014-04-25 Cary Coutant <ccoutant@google.com>
4944 * dwarf2out.c (should_move_die_to_comdat): A type definition
4945 can contain a subprogram definition, but don't move it to a
4947 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
4948 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
4950 (clone_tree_hash): Rename to...
4951 (clone_tree_partial): ...this; change callers. Copy
4952 DW_TAG_subprogram DIEs as declarations.
4953 (copy_decls_walk): Don't copy children of a declaration into a
4956 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
4959 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
4962 2014-04-25 Jiong Wang <jiong.wang@arm.com>
4964 * config/arm/predicates.md (call_insn_operand): Add long_call check.
4965 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
4967 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
4970 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4972 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
4974 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4976 PR tree-optimization/60930
4977 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
4978 creating a multiply candidate by folding two constant
4979 multiplicands when the result overflows.
4981 2014-04-25 Jakub Jelinek <jakub@redhat.com>
4983 PR tree-optimization/60960
4984 * tree-vect-generic.c (expand_vector_operation): Only call
4985 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
4987 2014-04-25 Tom de Vries <tom@codesourcery.com>
4989 * expr.c (clobber_reg_mode): New function.
4990 * expr.h (clobber_reg): New function.
4992 2014-04-25 Tom de Vries <tom@codesourcery.com>
4994 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
4997 2014-04-25 Radovan Obradovic <robradovic@mips.com>
4998 Tom de Vries <tom@codesourcery.com>
5000 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
5002 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
5003 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
5004 new argument to find_all_hard_reg_sets call.
5006 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5008 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
5009 Use HOST_WIDE_INT_C for mask literal.
5010 (aarch_rev16_shleft_mask_imm_p): Likewise.
5012 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
5015 * config/sparc/sparc.md (ashlsi3_extend): Delete.
5017 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
5019 PR preprocessor/56540
5020 * config/i386/i386-c.c (ix86_target_macros): Define
5021 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
5023 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5025 * configure.ac (tga_func): Remove.
5026 (LIB_TLS_SPEC): Remove.
5027 * configure: Regenerate.
5028 * config.in: Regenerate.
5029 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
5031 2014-04-25 Richard Biener <rguenther@suse.de>
5034 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
5035 call stmt use/clobber sets during stmt walk instead of
5036 walking the possibly incomplete set of caller edges.
5038 2014-04-25 Richard Biener <rguenther@suse.de>
5041 * passes.c (apply_ipa_transforms): Inline into only caller ...
5042 (execute_one_pass): ... here. Properly bring in function
5043 bodies for nodes we want to apply IPA transforms to.
5045 2014-04-24 Cong Hou <congh@google.com>
5047 PR tree-optimization/60896
5048 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
5049 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
5050 (vect_mark_pattern_stmts): Set the def type of all statements in
5051 PATTERN_DEF_SEQ as vect_internal_def.
5053 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
5055 * doc/extend.texi (PowerPC Built-in Functions): Document new
5056 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
5057 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
5059 * config/rs6000/predicates.md (const_0_to_3_operand): New
5060 predicate to match 0..3 integer constants.
5062 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
5063 to support adding miscellaneous builtin functions.
5064 (BU_DFP_MISC_2): Likewise.
5065 (BU_P7_MISC_1): Likewise.
5066 (BU_P7_MISC_2): Likewise.
5067 (BU_P8V_MISC_3): Likewise.
5068 (BU_MISC_1): Likewise.
5069 (BU_MISC_2): Likewise.
5070 (DIVWE): Add extended divide builtin functions.
5073 (DIVWEUO): Likewise.
5077 (DIVDEUO): Likewise.
5078 (DXEX): Add decimal floating-point builtin functions.
5081 (DDEDPDQ): Likewise.
5083 (DENBCDQ): Likewise.
5090 (CDTBCD): Add new BCD builtin functions.
5094 (BCDADD_LT): Likewise.
5095 (BCDADD_EQ): Likewise.
5096 (BCDADD_GT): Likewise.
5097 (BCDADD_OV): Likewise.
5099 (BCDSUB_LT): Likewise.
5100 (BCDSUB_EQ): Likewise.
5101 (BCDSUB_GT): Likewise.
5102 (BCDSUB_OV): Likewise.
5103 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
5104 (UNPACK_TD): Likewise.
5105 (PACK_TF): Likewise.
5106 (UNPACK_TF): Likewise.
5107 (UNPACK_TF_0): Likewise.
5108 (UNPACK_TF_1): Likewise.
5109 (PACK_V1TI): Likewise.
5110 (UNPACK_V1TI): Likewise.
5112 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
5113 support for decimal floating point builtin functions.
5114 (rs6000_expand_ternop_builtin): Add checks for the new builtin
5115 functions that take constant arguments.
5116 (rs6000_invalid_builtin): Add decimal floating point builtin support.
5117 (rs6000_init_builtins): Setup long double, _Decimal64, and
5118 _Decimal128 types for new builtin functions.
5119 (builtin_function_type): Set the unsigned flags appropriately for
5120 the new builtin functions.
5121 (rs6000_opt_masks): Add support for decimal floating point builtin
5124 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
5125 floating point builtin functions.
5126 (RS6000_BTM_COMMON): Likewise.
5127 (RS6000_BTI_long_double): Likewise.
5128 (RS6000_BTI_dfloat64): Likewise.
5129 (RS6000_BTI_dfloat128): Likewise.
5130 (long_double_type_internal_node): Likewise.
5131 (dfloat64_type_internal_node): Likewise.
5132 (dfloat128_type_internal_node): Likewise.
5134 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
5135 2.07 bcd arithmetic instructions.
5136 (UNSPEC_BCDSUB): Likewise.
5137 (UNSPEC_BCD_OVERFLOW): Likewise.
5138 (UNSPEC_BCD_ADD_SUB): Likewise.
5139 (bcd_add_sub): Likewise.
5140 (BCD_TEST): Likewise.
5141 (bcd<bcd_add_sub>): Likewise.
5142 (bcd<bcd_add_sub>_test): Likewise.
5143 (bcd<bcd_add_sub>_test2): Likewise.
5144 (bcd<bcd_add_sub>_<code>): Likewise.
5145 (peephole2 for combined bcd ops): Likewise.
5147 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
5148 decimal floating point builtin functions.
5149 (UNSPEC_DENBCD): Likewise.
5150 (UNSPEC_DXEX): Likewise.
5151 (UNSPEC_DIEX): Likewise.
5152 (UNSPEC_DSCLI): Likewise.
5153 (UNSPEC_DSCRI): Likewise.
5154 (D64_D128): Likewise.
5155 (dfp_suffix): Likewise.
5156 (dfp_ddedpd_<mode>): Likewise.
5157 (dfp_denbcd_<mode>): Likewise.
5158 (dfp_dxex_<mode>): Likewise.
5159 (dfp_diex_<mode>): Likewise.
5160 (dfp_dscli_<mode>): Likewise.
5161 (dfp_dscri_<mode>): Likewise.
5163 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
5165 (UNSPEC_CDTBCD): Likewise.
5166 (UNSPEC_CBCDTD): Likewise.
5167 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
5168 (UNSPEC_DIVEO): Likewise.
5169 (UNSPEC_DIVEU): Likewise.
5170 (UNSPEC_DIVEUO): Likewise.
5171 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
5172 pack/unpack 128-bit types.
5173 (UNSPEC_PACK_128BIT): Likewise.
5174 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
5175 (udiv<mode>3): Use idiv_ldiv mode attribute.
5176 (div<mode>3): Likewise.
5177 (addg6s): Add new BCD builtin functions.
5180 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
5181 (div_extend): Likewise.
5182 (div<div_extend>_<mode>"): Likewise.
5183 (FP128_64): Add support for new builtin functions to pack/unpack
5185 (unpack<mode>): Likewise.
5186 (unpacktf_0): Likewise.
5187 (unpacktf_1): Likewise.
5188 (unpack<mode>_dm): Likewise.
5189 (unpack<mode>_nodm): Likewise.
5190 (pack<mode>): Likewise.
5191 (unpackv1ti): Likewise.
5192 (packv1ti): Likewise.
5194 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
5196 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
5199 2014-04-24 Jakub Jelinek <jakub@redhat.com>
5201 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
5202 * gimplify.c (omp_is_private): Change last argument's type to int.
5203 Only diagnose lastprivate if the simd argument is 1, only diagnose
5204 linear if the simd argument is 2.
5205 (gimplify_omp_for): Adjust omp_is_private callers. When adding
5206 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
5207 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
5208 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
5209 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
5210 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5211 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
5212 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5213 * tree-nested.c (convert_nonlocal_omp_clauses,
5214 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
5216 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
5219 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
5222 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
5224 * flag-types.h (enum ivar_visibility): Add.
5226 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
5228 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
5229 function * argument.
5231 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
5233 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
5235 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5236 Tom de Vries <tom@codesourcery.com>
5238 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
5239 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
5241 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
5242 * emit-rtl.c (try_split): Same.
5244 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5245 Tom de Vries <tom@codesourcery.com>
5247 * common.opt (fuse-caller-save): New option.
5249 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
5251 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
5252 elements for big-endian.
5254 2014-04-24 Richard Biener <rguenther@suse.de>
5256 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
5257 during TER and instead use the sepops interface for expanding
5258 non-GIMPLE_SINGLE_RHS.
5260 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5262 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
5263 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
5265 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5267 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
5268 assembler 64-bit option.
5269 * configure: Regenerate.
5271 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5273 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
5274 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
5275 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
5276 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
5277 (TARGET_CRYPTO): Take TARGET_SIMD into account.
5279 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5281 * config/aarch64/aarch64-builtins.c
5282 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
5283 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
5284 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
5285 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
5287 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
5288 (Vrevsuff): New mode attribute.
5290 2014-04-24 Terry Guo <terry.guo@arm.com>
5292 * config/arm/arm.h (machine_function): Define variable
5293 after_arm_reorg here.
5294 * config/arm/arm.c (after_arm_reorg): Remove the definition.
5295 (arm_split_constant): Update the way to access variable
5298 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
5300 2014-04-23 Tom de Vries <tom@codesourcery.com>
5302 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
5304 2014-04-23 David Malcolm <dmalcolm@redhat.com>
5306 * is-a.h: Update comments to reflect the following changes to the
5307 "pointerness" of the API, making the template parameter match the
5308 return type, allowing use of is-a.h with typedefs of pointers.
5309 (is_a_helper::cast): Return a T rather then a pointer to a T, so
5310 that the return type matches the parameter to the is_a_helper.
5312 (dyn_cast): Likewise.
5314 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
5315 pointer from the is-a.h API.
5317 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
5318 (is_a_helper <cgraph_node *>::test): ...this, matching change to
5320 (is_a_helper <varpool_node>::test): Likewise, convert to...
5321 (is_a_helper <varpool_node *>::test): ...this.
5323 (varpool_first_variable): Update for removal of implicit pointer
5324 from the is-a.h API.
5325 (varpool_next_variable): Likewise.
5326 (varpool_first_static_initializer): Likewise.
5327 (varpool_next_static_initializer): Likewise.
5328 (varpool_first_defined_variable): Likewise.
5329 (varpool_next_defined_variable): Likewise.
5330 (cgraph_first_defined_function): Likewise.
5331 (cgraph_next_defined_function): Likewise.
5332 (cgraph_first_function): Likewise.
5333 (cgraph_next_function): Likewise.
5334 (cgraph_first_function_with_gimple_body): Likewise.
5335 (cgraph_next_function_with_gimple_body): Likewise.
5336 (cgraph_alias_target): Likewise.
5337 (varpool_alias_target): Likewise.
5338 (cgraph_function_or_thunk_node): Likewise.
5339 (varpool_variable_node): Likewise.
5340 (symtab_real_symbol_p): Likewise.
5341 * cgraphunit.c (referred_to_p): Likewise.
5342 (analyze_functions): Likewise.
5343 (handle_alias_pairs): Likewise.
5344 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
5345 * gimple-ssa.h (gimple_vuse_op): Likewise.
5346 (gimple_vdef_op): Likewise.
5347 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
5348 * gimple.c (gimple_build_asm_1): Likewise.
5349 (gimple_build_try): Likewise.
5350 (gimple_build_resx): Likewise.
5351 (gimple_build_eh_dispatch): Likewise.
5352 (gimple_build_omp_for): Likewise.
5353 (gimple_omp_for_set_clauses): Likewise.
5355 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
5356 (is_a_helper <gimple_statement_asm *>::test): ...this.
5357 (is_a_helper <gimple_statement_bind>::test): Convert to...
5358 (is_a_helper <gimple_statement_bind *>::test): ...this.
5359 (is_a_helper <gimple_statement_call>::test): Convert to...
5360 (is_a_helper <gimple_statement_call *>::test): ...this.
5361 (is_a_helper <gimple_statement_catch>::test): Convert to...
5362 (is_a_helper <gimple_statement_catch *>::test): ...this.
5363 (is_a_helper <gimple_statement_resx>::test): Convert to...
5364 (is_a_helper <gimple_statement_resx *>::test): ...this.
5365 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
5366 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
5367 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
5368 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
5369 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
5370 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
5371 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
5372 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
5373 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
5374 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
5375 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
5376 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
5377 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
5378 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
5379 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
5380 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
5381 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
5382 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
5383 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
5384 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
5385 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
5386 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
5387 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
5388 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
5389 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
5390 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
5391 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
5392 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
5393 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
5394 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
5395 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
5396 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
5397 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
5398 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
5399 (is_a_helper <gimple_statement_phi>::test): Convert to...
5400 (is_a_helper <gimple_statement_phi *>::test): ...this.
5401 (is_a_helper <gimple_statement_transaction>::test): Convert to...
5402 (is_a_helper <gimple_statement_transaction *>::test): ...this.
5403 (is_a_helper <gimple_statement_try>::test): Convert to...
5404 (is_a_helper <gimple_statement_try *>::test): ...this.
5405 (is_a_helper <gimple_statement_wce>::test): Convert to...
5406 (is_a_helper <gimple_statement_wce *>::test): ...this.
5407 (is_a_helper <const gimple_statement_asm>::test): Convert to...
5408 (is_a_helper <const gimple_statement_asm *>::test): ...this.
5409 (is_a_helper <const gimple_statement_bind>::test): Convert to...
5410 (is_a_helper <const gimple_statement_bind *>::test): ...this.
5411 (is_a_helper <const gimple_statement_call>::test): Convert to...
5412 (is_a_helper <const gimple_statement_call *>::test): ...this.
5413 (is_a_helper <const gimple_statement_catch>::test): Convert to...
5414 (is_a_helper <const gimple_statement_catch *>::test): ...this.
5415 (is_a_helper <const gimple_statement_resx>::test): Convert to...
5416 (is_a_helper <const gimple_statement_resx *>::test): ...this.
5417 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
5418 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
5419 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
5420 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
5421 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
5423 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
5425 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
5427 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
5429 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
5430 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
5431 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
5433 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
5434 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
5436 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
5437 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
5438 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
5439 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
5440 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
5441 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
5443 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
5444 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
5445 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
5446 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
5448 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
5449 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
5450 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
5451 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
5452 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
5453 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
5454 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
5455 (is_a_helper <const gimple_statement_phi>::test): Convert to...
5456 (is_a_helper <const gimple_statement_phi *>::test): ...this.
5457 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
5458 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
5459 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
5460 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
5461 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
5462 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
5463 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
5465 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
5467 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
5468 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
5470 (gimple_use_ops): Update for removal of implicit pointer from the
5472 (gimple_set_use_ops): Likewise.
5473 (gimple_vuse): Likewise.
5474 (gimple_vdef): Likewise.
5475 (gimple_vuse_ptr): Likewise.
5476 (gimple_vdef_ptr): Likewise.
5477 (gimple_set_vuse): Likewise.
5478 (gimple_set_vdef): Likewise.
5479 (gimple_omp_return_set_lhs): Likewise.
5480 (gimple_omp_return_lhs): Likewise.
5481 (gimple_omp_return_lhs_ptr): Likewise.
5482 (gimple_call_fntype): Likewise.
5483 (gimple_call_set_fntype): Likewise.
5484 (gimple_call_set_internal_fn): Likewise.
5485 (gimple_call_use_set): Likewise.
5486 (gimple_call_clobber_set): Likewise.
5487 (gimple_bind_vars): Likewise.
5488 (gimple_bind_set_vars): Likewise.
5489 (gimple_bind_body_ptr): Likewise.
5490 (gimple_bind_set_body): Likewise.
5491 (gimple_bind_add_stmt): Likewise.
5492 (gimple_bind_block): Likewise.
5493 (gimple_bind_set_block): Likewise.
5494 (gimple_asm_ninputs): Likewise.
5495 (gimple_asm_noutputs): Likewise.
5496 (gimple_asm_nclobbers): Likewise.
5497 (gimple_asm_nlabels): Likewise.
5498 (gimple_asm_input_op): Likewise.
5499 (gimple_asm_input_op_ptr): Likewise.
5500 (gimple_asm_output_op): Likewise.
5501 (gimple_asm_output_op_ptr): Likewise.
5502 (gimple_asm_set_output_op): Likewise.
5503 (gimple_asm_clobber_op): Likewise.
5504 (gimple_asm_set_clobber_op): Likewise.
5505 (gimple_asm_label_op): Likewise.
5506 (gimple_asm_set_label_op): Likewise.
5507 (gimple_asm_string): Likewise.
5508 (gimple_catch_types): Likewise.
5509 (gimple_catch_types_ptr): Likewise.
5510 (gimple_catch_handler_ptr): Likewise.
5511 (gimple_catch_set_types): Likewise.
5512 (gimple_catch_set_handler): Likewise.
5513 (gimple_eh_filter_types): Likewise.
5514 (gimple_eh_filter_types_ptr): Likewise.
5515 (gimple_eh_filter_failure_ptr): Likewise.
5516 (gimple_eh_filter_set_types): Likewise.
5517 (gimple_eh_filter_set_failure): Likewise.
5518 (gimple_eh_must_not_throw_fndecl): Likewise.
5519 (gimple_eh_must_not_throw_set_fndecl): Likewise.
5520 (gimple_eh_else_n_body_ptr): Likewise.
5521 (gimple_eh_else_e_body_ptr): Likewise.
5522 (gimple_eh_else_set_n_body): Likewise.
5523 (gimple_eh_else_set_e_body): Likewise.
5524 (gimple_try_eval_ptr): Likewise.
5525 (gimple_try_cleanup_ptr): Likewise.
5526 (gimple_try_set_eval): Likewise.
5527 (gimple_try_set_cleanup): Likewise.
5528 (gimple_wce_cleanup_ptr): Likewise.
5529 (gimple_wce_set_cleanup): Likewise.
5530 (gimple_phi_capacity): Likewise.
5531 (gimple_phi_num_args): Likewise.
5532 (gimple_phi_result): Likewise.
5533 (gimple_phi_result_ptr): Likewise.
5534 (gimple_phi_set_result): Likewise.
5535 (gimple_phi_arg): Likewise.
5536 (gimple_phi_set_arg): Likewise.
5537 (gimple_resx_region): Likewise.
5538 (gimple_resx_set_region): Likewise.
5539 (gimple_eh_dispatch_region): Likewise.
5540 (gimple_eh_dispatch_set_region): Likewise.
5541 (gimple_omp_critical_name): Likewise.
5542 (gimple_omp_critical_name_ptr): Likewise.
5543 (gimple_omp_critical_set_name): Likewise.
5544 (gimple_omp_for_clauses): Likewise.
5545 (gimple_omp_for_clauses_ptr): Likewise.
5546 (gimple_omp_for_set_clauses): Likewise.
5547 (gimple_omp_for_collapse): Likewise.
5548 (gimple_omp_for_index): Likewise.
5549 (gimple_omp_for_index_ptr): Likewise.
5550 (gimple_omp_for_set_index): Likewise.
5551 (gimple_omp_for_initial): Likewise.
5552 (gimple_omp_for_initial_ptr): Likewise.
5553 (gimple_omp_for_set_initial): Likewise.
5554 (gimple_omp_for_final): Likewise.
5555 (gimple_omp_for_final_ptr): Likewise.
5556 (gimple_omp_for_set_final): Likewise.
5557 (gimple_omp_for_incr): Likewise.
5558 (gimple_omp_for_incr_ptr): Likewise.
5559 (gimple_omp_for_set_incr): Likewise.
5560 (gimple_omp_for_pre_body_ptr): Likewise.
5561 (gimple_omp_for_set_pre_body): Likewise.
5562 (gimple_omp_parallel_clauses): Likewise.
5563 (gimple_omp_parallel_clauses_ptr): Likewise.
5564 (gimple_omp_parallel_set_clauses): Likewise.
5565 (gimple_omp_parallel_child_fn): Likewise.
5566 (gimple_omp_parallel_child_fn_ptr): Likewise.
5567 (gimple_omp_parallel_set_child_fn): Likewise.
5568 (gimple_omp_parallel_data_arg): Likewise.
5569 (gimple_omp_parallel_data_arg_ptr): Likewise.
5570 (gimple_omp_parallel_set_data_arg): Likewise.
5571 (gimple_omp_task_clauses): Likewise.
5572 (gimple_omp_task_clauses_ptr): Likewise.
5573 (gimple_omp_task_set_clauses): Likewise.
5574 (gimple_omp_task_child_fn): Likewise.
5575 (gimple_omp_task_child_fn_ptr): Likewise.
5576 (gimple_omp_task_set_child_fn): Likewise.
5577 (gimple_omp_task_data_arg): Likewise.
5578 (gimple_omp_task_data_arg_ptr): Likewise.
5579 (gimple_omp_task_set_data_arg): Likewise.
5580 (gimple_omp_taskreg_clauses): Likewise.
5581 (gimple_omp_taskreg_clauses_ptr): Likewise.
5582 (gimple_omp_taskreg_set_clauses): Likewise.
5583 (gimple_omp_taskreg_child_fn): Likewise.
5584 (gimple_omp_taskreg_child_fn_ptr): Likewise.
5585 (gimple_omp_taskreg_set_child_fn): Likewise.
5586 (gimple_omp_taskreg_data_arg): Likewise.
5587 (gimple_omp_taskreg_data_arg_ptr): Likewise.
5588 (gimple_omp_taskreg_set_data_arg): Likewise.
5589 (gimple_omp_task_copy_fn): Likewise.
5590 (gimple_omp_task_copy_fn_ptr): Likewise.
5591 (gimple_omp_task_set_copy_fn): Likewise.
5592 (gimple_omp_task_arg_size): Likewise.
5593 (gimple_omp_task_arg_size_ptr): Likewise.
5594 (gimple_omp_task_set_arg_size): Likewise.
5595 (gimple_omp_task_arg_align): Likewise.
5596 (gimple_omp_task_arg_align_ptr): Likewise.
5597 (gimple_omp_task_set_arg_align): Likewise.
5598 (gimple_omp_single_clauses): Likewise.
5599 (gimple_omp_single_clauses_ptr): Likewise.
5600 (gimple_omp_single_set_clauses): Likewise.
5601 (gimple_omp_target_clauses): Likewise.
5602 (gimple_omp_target_clauses_ptr): Likewise.
5603 (gimple_omp_target_set_clauses): Likewise.
5604 (gimple_omp_target_child_fn): Likewise.
5605 (gimple_omp_target_child_fn_ptr): Likewise.
5606 (gimple_omp_target_set_child_fn): Likewise.
5607 (gimple_omp_target_data_arg): Likewise.
5608 (gimple_omp_target_data_arg_ptr): Likewise.
5609 (gimple_omp_target_set_data_arg): Likewise.
5610 (gimple_omp_teams_clauses): Likewise.
5611 (gimple_omp_teams_clauses_ptr): Likewise.
5612 (gimple_omp_teams_set_clauses): Likewise.
5613 (gimple_omp_sections_clauses): Likewise.
5614 (gimple_omp_sections_clauses_ptr): Likewise.
5615 (gimple_omp_sections_set_clauses): Likewise.
5616 (gimple_omp_sections_control): Likewise.
5617 (gimple_omp_sections_control_ptr): Likewise.
5618 (gimple_omp_sections_set_control): Likewise.
5619 (gimple_omp_for_set_cond): Likewise.
5620 (gimple_omp_for_cond): Likewise.
5621 (gimple_omp_atomic_store_set_val): Likewise.
5622 (gimple_omp_atomic_store_val): Likewise.
5623 (gimple_omp_atomic_store_val_ptr): Likewise.
5624 (gimple_omp_atomic_load_set_lhs): Likewise.
5625 (gimple_omp_atomic_load_lhs): Likewise.
5626 (gimple_omp_atomic_load_lhs_ptr): Likewise.
5627 (gimple_omp_atomic_load_set_rhs): Likewise.
5628 (gimple_omp_atomic_load_rhs): Likewise.
5629 (gimple_omp_atomic_load_rhs_ptr): Likewise.
5630 (gimple_omp_continue_control_def): Likewise.
5631 (gimple_omp_continue_control_def_ptr): Likewise.
5632 (gimple_omp_continue_set_control_def): Likewise.
5633 (gimple_omp_continue_control_use): Likewise.
5634 (gimple_omp_continue_control_use_ptr): Likewise.
5635 (gimple_omp_continue_set_control_use): Likewise.
5636 (gimple_transaction_body_ptr): Likewise.
5637 (gimple_transaction_label): Likewise.
5638 (gimple_transaction_label_ptr): Likewise.
5639 (gimple_transaction_set_body): Likewise.
5640 (gimple_transaction_set_label): Likewise.
5642 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
5643 * ipa-inline-analysis.c (inline_write_summary): Likewise.
5644 * ipa-ref.c (ipa_record_reference): Likewise.
5645 * ipa-reference.c (analyze_function): Likewise.
5646 (ipa_reference_write_optimization_summary): Likewise.
5647 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
5648 (address_taken_from_non_vtable_p): Likewise.
5649 (comdat_can_be_unshared_p_1): Likewise.
5650 * lto-cgraph.c (lto_output_ref): Likewise.
5651 (add_references): Likewise.
5652 (compute_ltrans_boundary): Likewise.
5653 (output_symtab): Likewise.
5654 (input_ref): Likewise.
5655 (input_cgraph_1): Likewise.
5656 (output_cgraph_opt_summary): Likewise.
5657 * lto-streamer-out.c (lto_output): Likewise.
5658 (output_symbol_p): Likewise.
5659 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
5660 (lsei_start_function_in_partition): Likewise.
5661 (lsei_next_variable_in_partition): Likewise.
5662 (lsei_start_variable_in_partition): Likewise.
5663 * symtab.c (insert_to_assembler_name_hash): Likewise.
5664 (unlink_from_assembler_name_hash): Likewise.
5665 (symtab_unregister_node): Likewise.
5666 (symtab_remove_node): Likewise.
5667 (dump_symtab_node): Likewise.
5668 (verify_symtab_base): Likewise.
5669 (verify_symtab_node): Likewise.
5670 (symtab_make_decl_local): Likewise.
5671 (symtab_alias_ultimate_target): Likewise.
5672 (symtab_resolve_alias): Likewise.
5673 (symtab_get_symbol_partitioning_class): Likewise.
5674 * tree-phinodes.c (allocate_phi_node): Likewise.
5675 (reserve_phi_args_for_new_edge): Likewise.
5676 (remove_phi_args): Likewise.
5677 * varpool.c (varpool_node_for_asm): Likewise.
5678 (varpool_remove_unreferenced_decls): Likewise.
5680 2014-04-23 Jeff Law <law@redhat.com>
5682 PR tree-optimization/60902
5683 * tree-ssa-threadedge.c
5684 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
5685 invalidate outputs from statements that do not produce useful
5686 outputs for threading.
5688 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5690 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
5691 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
5692 machine descriptions for Stack Smashing Protector.
5694 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
5696 * aarch64.md (<optab>_rol<mode>3): New pattern.
5697 (<optab>_rolsi3_uxtw): Likewise.
5698 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
5700 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
5702 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
5703 (arm_cortex_a12_tune): Likewise.
5705 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5707 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
5709 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5711 * config/arm/arm.md (arm_rev16si2): New pattern.
5712 (arm_rev16si2_alt): Likewise.
5713 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
5715 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5717 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
5718 (rev16<mode>2_alt): Likewise.
5719 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
5720 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
5721 (aarch_rev16_shleft_mask_imm_p): Likewise.
5722 (aarch_rev16_p_1): Likewise.
5723 (aarch_rev16_p): Likewise.
5724 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5725 (aarch_rev16_shright_mask_imm_p): Likewise.
5726 (aarch_rev16_shleft_mask_imm_p): Likewise.
5728 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5730 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5731 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5733 (cortex_a53_extra_costs): Likewise.
5734 (cortex_a57_extra_costs): Likewise.
5735 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5736 (cortexa7_extra_costs): Likewise.
5737 (cortexa8_extra_costs): Likewise.
5738 (cortexa12_extra_costs): Likewise.
5739 (cortexa15_extra_costs): Likewise.
5740 (v7m_extra_costs): Likewise.
5741 (arm_new_rtx_costs): Handle BSWAP.
5743 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5745 * config/arm/arm.c (cortexa8_extra_costs): New table.
5746 (arm_cortex_a8_tune): New tuning struct.
5747 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5749 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5751 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5753 2014-04-23 Richard Biener <rguenther@suse.de>
5755 * Makefile.in (OBJS): Remove loop-unswitch.o.
5756 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5757 * passes.def (pass_rtl_unswitch): Likewise.
5758 * loop-init.c (gate_rtl_unswitch): Likewise.
5759 (rtl_unswitch): Likewise.
5760 (pass_data_rtl_unswitch): Likewise.
5761 (pass_rtl_unswitch): Likewise.
5762 (make_pass_rtl_unswitch): Likewise.
5763 * rtl.h (reversed_condition): Likewise.
5764 (compare_and_jump_seq): Likewise.
5765 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5767 * loop-unroll.c (compare_and_jump_seq): Likewise.
5769 2014-04-23 Richard Biener <rguenther@suse.de>
5771 PR tree-optimization/60903
5772 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5773 commented code block.
5774 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5775 loop flags to newly created BBs and edges.
5777 2014-04-23 Nick Clifton <nickc@redhat.com>
5779 * config/msp430/msp430.c (msp430_handle_option): Move function
5781 (msp430_option_override): Simplify mcu and mcpu option handling.
5782 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
5783 support for -mhwmult command line option.
5784 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
5785 -mhwmult command line option.
5786 (msp430_hwmult_enabled): Delete.
5787 (msp43o_output_labelref): Add support for -mhwmult command line option.
5788 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
5789 (umulsidi3): Likewise.
5790 * config/msp430/msp430.opt (mmcu): Add Report attribute.
5791 (mcpu, mlarge, msmall): Likewise.
5792 (mhwmult): New option.
5793 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
5795 (msp430_is_f5_mcu): Remove prototype.
5796 (msp430_use_f5_series_hwmult): Add prototype.
5797 * config/msp430/msp430-opts.h: New file.
5798 * common/config/msp430: New directory.
5799 * common/config/msp430/msp430-common.c: New file.
5800 * config.gcc (msp430): Remove target_has_targetm_common.
5801 * doc/invoke.texi: Document -mhwmult command line option.
5803 2014-04-23 Nick Clifton <nickc@redhat.com>
5805 * config/i386/cygwin.h (ENDFILE_SPEC): Include
5806 default-manifest.o if it can be found in the search path.
5807 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
5809 2014-04-23 Terry Guo <terry.guo@arm.com>
5811 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
5813 2014-04-23 Richard Biener <rguenther@suse.de>
5816 * tree-inline.c (declare_return_variable): Use mark_addressable.
5818 2014-04-23 Richard Biener <rguenther@suse.de>
5821 * loop-init.c (loop_optimizer_init): Make sure to apply
5822 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
5824 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5827 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
5829 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
5830 if flag_sanitize_undefined_trap_on_error.
5831 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
5832 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
5833 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
5834 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
5835 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
5836 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
5837 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
5838 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
5839 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
5840 * ubsan.c (ubsan_instrument_unreachable): Return
5841 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
5842 (ubsan_expand_null_ifn): Emit __builtin_trap ()
5843 if flag_sanitize_undefined_trap_on_error and
5844 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
5845 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
5846 instrument_bool_enum_load): Emit __builtin_trap () if
5847 flag_sanitize_undefined_trap_on_error and
5848 __builtin_handle_*_abort () if !flag_sanitize_recover.
5849 * doc/invoke.texi (-fsanitize-recover,
5850 -fsanitize-undefined-trap-on-error): Document.
5852 2014-04-22 Christian Bruel <christian.bruel@st.com>
5854 * config/sh/sh.md (mov<mode>): Replace movQIHI.
5855 Force immediates to SImode.
5857 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
5859 * config/nios2/nios2.md (UNSPEC_ROUND): New.
5861 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
5862 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
5863 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
5864 (nios2_fpu_insn): Add entry for round.
5865 (N2FPU_NO_ERRNO_P): Define.
5866 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
5868 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
5870 2014-04-22 Richard Henderson <rth@redhat.com>
5872 * config/aarch64/aarch64 (addti3, subti3): New expanders.
5873 (add<GPI>3_compare0): Remove leading * from name.
5874 (add<GPI>3_carryin): Likewise.
5875 (sub<GPI>3_compare0): Likewise.
5876 (sub<GPI>3_carryin): Likewise.
5877 (<su_optab>mulditi3): New expander.
5878 (multi3): New expander.
5879 (madd<GPI>): Remove leading * from name.
5881 2014-04-22 Martin Jambor <mjambor@suse.cz>
5883 * cgraphclones.c (cgraph_function_versioning): Copy
5884 ipa_transforms_to_apply instead of asserting it is empty.
5886 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
5889 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
5890 on count_exp to get mode.
5892 2014-04-22 Andrew Pinski <apinski@cavium.com>
5894 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5895 Handle TLS for ILP32.
5896 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
5897 (tlsie_small_<mode>): this and handle PTR.
5898 (tlsie_small_sidi): New pattern.
5899 (tlsle_small): Change to an expand to handle ILP32.
5900 (tlsle_small_<mode>): New pattern.
5901 (tlsdesc_small): Rename to ...
5902 (tlsdesc_small_<mode>): this and handle PTR.
5904 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5906 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
5908 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5910 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5911 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
5912 (aarch64_types_signed_poly_qualifiers): Likewise.
5913 (aarch64_types_unsigned_signed_qualifiers): Likewise.
5914 (aarch64_types_poly_signed_qualifiers): Likewise.
5915 (TYPES_REINTERP_SS): Type macro added.
5916 (TYPES_REINTERP_SU): Likewise.
5917 (TYPES_REINTERP_SP): Likewise.
5918 (TYPES_REINTERP_US): Likewise.
5919 (TYPES_REINTERP_PS): Likewise.
5920 (aarch64_fold_builtin): New expression folding added.
5921 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
5922 Declarations removed.
5923 (REINTERP_SS): Declarations added.
5924 (REINTERP_US): Likewise.
5925 (REINTERP_PS): Likewise.
5926 (REINTERP_SU): Likewise.
5927 (REINTERP_SP): Likewise.
5928 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
5929 (vreinterpretq_p8_f64): Likewise.
5930 (vreinterpret_p16_f64): Likewise.
5931 (vreinterpretq_p16_f64): Likewise.
5932 (vreinterpret_f32_f64): Likewise.
5933 (vreinterpretq_f32_f64): Likewise.
5934 (vreinterpret_f64_f32): Likewise.
5935 (vreinterpret_f64_p8): Likewise.
5936 (vreinterpret_f64_p16): Likewise.
5937 (vreinterpret_f64_s8): Likewise.
5938 (vreinterpret_f64_s16): Likewise.
5939 (vreinterpret_f64_s32): Likewise.
5940 (vreinterpret_f64_s64): Likewise.
5941 (vreinterpret_f64_u8): Likewise.
5942 (vreinterpret_f64_u16): Likewise.
5943 (vreinterpret_f64_u32): Likewise.
5944 (vreinterpret_f64_u64): Likewise.
5945 (vreinterpretq_f64_f32): Likewise.
5946 (vreinterpretq_f64_p8): Likewise.
5947 (vreinterpretq_f64_p16): Likewise.
5948 (vreinterpretq_f64_s8): Likewise.
5949 (vreinterpretq_f64_s16): Likewise.
5950 (vreinterpretq_f64_s32): Likewise.
5951 (vreinterpretq_f64_s64): Likewise.
5952 (vreinterpretq_f64_u8): Likewise.
5953 (vreinterpretq_f64_u16): Likewise.
5954 (vreinterpretq_f64_u32): Likewise.
5955 (vreinterpretq_f64_u64): Likewise.
5956 (vreinterpret_s64_f64): Likewise.
5957 (vreinterpretq_s64_f64): Likewise.
5958 (vreinterpret_u64_f64): Likewise.
5959 (vreinterpretq_u64_f64): Likewise.
5960 (vreinterpret_s8_f64): Likewise.
5961 (vreinterpretq_s8_f64): Likewise.
5962 (vreinterpret_s16_f64): Likewise.
5963 (vreinterpretq_s16_f64): Likewise.
5964 (vreinterpret_s32_f64): Likewise.
5965 (vreinterpretq_s32_f64): Likewise.
5966 (vreinterpret_u8_f64): Likewise.
5967 (vreinterpretq_u8_f64): Likewise.
5968 (vreinterpret_u16_f64): Likewise.
5969 (vreinterpretq_u16_f64): Likewise.
5970 (vreinterpret_u32_f64): Likewise.
5971 (vreinterpretq_u32_f64): Likewise.
5973 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5975 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5976 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
5977 (vreinterpret_p8_s8): Likewise.
5978 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
5979 (vreinterpret_p8_s16): Likewise.
5980 (vreinterpret_p8_s32): Likewise.
5981 (vreinterpret_p8_s64): Likewise.
5982 (vreinterpret_p8_f32): Likewise.
5983 (vreinterpret_p8_u8): Likewise.
5984 (vreinterpret_p8_u16): Likewise.
5985 (vreinterpret_p8_u32): Likewise.
5986 (vreinterpret_p8_u64): Likewise.
5987 (vreinterpret_p8_p16): Likewise.
5988 (vreinterpretq_p8_s8): Likewise.
5989 (vreinterpretq_p8_s16): Likewise.
5990 (vreinterpretq_p8_s32): Likewise.
5991 (vreinterpretq_p8_s64): Likewise.
5992 (vreinterpretq_p8_f32): Likewise.
5993 (vreinterpretq_p8_u8): Likewise.
5994 (vreinterpretq_p8_u16): Likewise.
5995 (vreinterpretq_p8_u32): Likewise.
5996 (vreinterpretq_p8_u64): Likewise.
5997 (vreinterpretq_p8_p16): Likewise.
5998 (vreinterpret_p16_s8): Likewise.
5999 (vreinterpret_p16_s16): Likewise.
6000 (vreinterpret_p16_s32): Likewise.
6001 (vreinterpret_p16_s64): Likewise.
6002 (vreinterpret_p16_f32): Likewise.
6003 (vreinterpret_p16_u8): Likewise.
6004 (vreinterpret_p16_u16): Likewise.
6005 (vreinterpret_p16_u32): Likewise.
6006 (vreinterpret_p16_u64): Likewise.
6007 (vreinterpret_p16_p8): Likewise.
6008 (vreinterpretq_p16_s8): Likewise.
6009 (vreinterpretq_p16_s16): Likewise.
6010 (vreinterpretq_p16_s32): Likewise.
6011 (vreinterpretq_p16_s64): Likewise.
6012 (vreinterpretq_p16_f32): Likewise.
6013 (vreinterpretq_p16_u8): Likewise.
6014 (vreinterpretq_p16_u16): Likewise.
6015 (vreinterpretq_p16_u32): Likewise.
6016 (vreinterpretq_p16_u64): Likewise.
6017 (vreinterpretq_p16_p8): Likewise.
6018 (vreinterpret_f32_s8): Likewise.
6019 (vreinterpret_f32_s16): Likewise.
6020 (vreinterpret_f32_s32): Likewise.
6021 (vreinterpret_f32_s64): Likewise.
6022 (vreinterpret_f32_u8): Likewise.
6023 (vreinterpret_f32_u16): Likewise.
6024 (vreinterpret_f32_u32): Likewise.
6025 (vreinterpret_f32_u64): Likewise.
6026 (vreinterpret_f32_p8): Likewise.
6027 (vreinterpret_f32_p16): Likewise.
6028 (vreinterpretq_f32_s8): Likewise.
6029 (vreinterpretq_f32_s16): Likewise.
6030 (vreinterpretq_f32_s32): Likewise.
6031 (vreinterpretq_f32_s64): Likewise.
6032 (vreinterpretq_f32_u8): Likewise.
6033 (vreinterpretq_f32_u16): Likewise.
6034 (vreinterpretq_f32_u32): Likewise.
6035 (vreinterpretq_f32_u64): Likewise.
6036 (vreinterpretq_f32_p8): Likewise.
6037 (vreinterpretq_f32_p16): Likewise.
6038 (vreinterpret_s64_s8): Likewise.
6039 (vreinterpret_s64_s16): Likewise.
6040 (vreinterpret_s64_s32): Likewise.
6041 (vreinterpret_s64_f32): Likewise.
6042 (vreinterpret_s64_u8): Likewise.
6043 (vreinterpret_s64_u16): Likewise.
6044 (vreinterpret_s64_u32): Likewise.
6045 (vreinterpret_s64_u64): Likewise.
6046 (vreinterpret_s64_p8): Likewise.
6047 (vreinterpret_s64_p16): Likewise.
6048 (vreinterpretq_s64_s8): Likewise.
6049 (vreinterpretq_s64_s16): Likewise.
6050 (vreinterpretq_s64_s32): Likewise.
6051 (vreinterpretq_s64_f32): Likewise.
6052 (vreinterpretq_s64_u8): Likewise.
6053 (vreinterpretq_s64_u16): Likewise.
6054 (vreinterpretq_s64_u32): Likewise.
6055 (vreinterpretq_s64_u64): Likewise.
6056 (vreinterpretq_s64_p8): Likewise.
6057 (vreinterpretq_s64_p16): Likewise.
6058 (vreinterpret_u64_s8): Likewise.
6059 (vreinterpret_u64_s16): Likewise.
6060 (vreinterpret_u64_s32): Likewise.
6061 (vreinterpret_u64_s64): Likewise.
6062 (vreinterpret_u64_f32): Likewise.
6063 (vreinterpret_u64_u8): Likewise.
6064 (vreinterpret_u64_u16): Likewise.
6065 (vreinterpret_u64_u32): Likewise.
6066 (vreinterpret_u64_p8): Likewise.
6067 (vreinterpret_u64_p16): Likewise.
6068 (vreinterpretq_u64_s8): Likewise.
6069 (vreinterpretq_u64_s16): Likewise.
6070 (vreinterpretq_u64_s32): Likewise.
6071 (vreinterpretq_u64_s64): Likewise.
6072 (vreinterpretq_u64_f32): Likewise.
6073 (vreinterpretq_u64_u8): Likewise.
6074 (vreinterpretq_u64_u16): Likewise.
6075 (vreinterpretq_u64_u32): Likewise.
6076 (vreinterpretq_u64_p8): Likewise.
6077 (vreinterpretq_u64_p16): Likewise.
6078 (vreinterpret_s8_s16): Likewise.
6079 (vreinterpret_s8_s32): Likewise.
6080 (vreinterpret_s8_s64): Likewise.
6081 (vreinterpret_s8_f32): Likewise.
6082 (vreinterpret_s8_u8): Likewise.
6083 (vreinterpret_s8_u16): Likewise.
6084 (vreinterpret_s8_u32): Likewise.
6085 (vreinterpret_s8_u64): Likewise.
6086 (vreinterpret_s8_p8): Likewise.
6087 (vreinterpret_s8_p16): Likewise.
6088 (vreinterpretq_s8_s16): Likewise.
6089 (vreinterpretq_s8_s32): Likewise.
6090 (vreinterpretq_s8_s64): Likewise.
6091 (vreinterpretq_s8_f32): Likewise.
6092 (vreinterpretq_s8_u8): Likewise.
6093 (vreinterpretq_s8_u16): Likewise.
6094 (vreinterpretq_s8_u32): Likewise.
6095 (vreinterpretq_s8_u64): Likewise.
6096 (vreinterpretq_s8_p8): Likewise.
6097 (vreinterpretq_s8_p16): Likewise.
6098 (vreinterpret_s16_s8): Likewise.
6099 (vreinterpret_s16_s32): Likewise.
6100 (vreinterpret_s16_s64): Likewise.
6101 (vreinterpret_s16_f32): Likewise.
6102 (vreinterpret_s16_u8): Likewise.
6103 (vreinterpret_s16_u16): Likewise.
6104 (vreinterpret_s16_u32): Likewise.
6105 (vreinterpret_s16_u64): Likewise.
6106 (vreinterpret_s16_p8): Likewise.
6107 (vreinterpret_s16_p16): Likewise.
6108 (vreinterpretq_s16_s8): Likewise.
6109 (vreinterpretq_s16_s32): Likewise.
6110 (vreinterpretq_s16_s64): Likewise.
6111 (vreinterpretq_s16_f32): Likewise.
6112 (vreinterpretq_s16_u8): Likewise.
6113 (vreinterpretq_s16_u16): Likewise.
6114 (vreinterpretq_s16_u32): Likewise.
6115 (vreinterpretq_s16_u64): Likewise.
6116 (vreinterpretq_s16_p8): Likewise.
6117 (vreinterpretq_s16_p16): Likewise.
6118 (vreinterpret_s32_s8): Likewise.
6119 (vreinterpret_s32_s16): Likewise.
6120 (vreinterpret_s32_s64): Likewise.
6121 (vreinterpret_s32_f32): Likewise.
6122 (vreinterpret_s32_u8): Likewise.
6123 (vreinterpret_s32_u16): Likewise.
6124 (vreinterpret_s32_u32): Likewise.
6125 (vreinterpret_s32_u64): Likewise.
6126 (vreinterpret_s32_p8): Likewise.
6127 (vreinterpret_s32_p16): Likewise.
6128 (vreinterpretq_s32_s8): Likewise.
6129 (vreinterpretq_s32_s16): Likewise.
6130 (vreinterpretq_s32_s64): Likewise.
6131 (vreinterpretq_s32_f32): Likewise.
6132 (vreinterpretq_s32_u8): Likewise.
6133 (vreinterpretq_s32_u16): Likewise.
6134 (vreinterpretq_s32_u32): Likewise.
6135 (vreinterpretq_s32_u64): Likewise.
6136 (vreinterpretq_s32_p8): Likewise.
6137 (vreinterpretq_s32_p16): Likewise.
6138 (vreinterpret_u8_s8): Likewise.
6139 (vreinterpret_u8_s16): Likewise.
6140 (vreinterpret_u8_s32): Likewise.
6141 (vreinterpret_u8_s64): Likewise.
6142 (vreinterpret_u8_f32): Likewise.
6143 (vreinterpret_u8_u16): Likewise.
6144 (vreinterpret_u8_u32): Likewise.
6145 (vreinterpret_u8_u64): Likewise.
6146 (vreinterpret_u8_p8): Likewise.
6147 (vreinterpret_u8_p16): Likewise.
6148 (vreinterpretq_u8_s8): Likewise.
6149 (vreinterpretq_u8_s16): Likewise.
6150 (vreinterpretq_u8_s32): Likewise.
6151 (vreinterpretq_u8_s64): Likewise.
6152 (vreinterpretq_u8_f32): Likewise.
6153 (vreinterpretq_u8_u16): Likewise.
6154 (vreinterpretq_u8_u32): Likewise.
6155 (vreinterpretq_u8_u64): Likewise.
6156 (vreinterpretq_u8_p8): Likewise.
6157 (vreinterpretq_u8_p16): Likewise.
6158 (vreinterpret_u16_s8): Likewise.
6159 (vreinterpret_u16_s16): Likewise.
6160 (vreinterpret_u16_s32): Likewise.
6161 (vreinterpret_u16_s64): Likewise.
6162 (vreinterpret_u16_f32): Likewise.
6163 (vreinterpret_u16_u8): Likewise.
6164 (vreinterpret_u16_u32): Likewise.
6165 (vreinterpret_u16_u64): Likewise.
6166 (vreinterpret_u16_p8): Likewise.
6167 (vreinterpret_u16_p16): Likewise.
6168 (vreinterpretq_u16_s8): Likewise.
6169 (vreinterpretq_u16_s16): Likewise.
6170 (vreinterpretq_u16_s32): Likewise.
6171 (vreinterpretq_u16_s64): Likewise.
6172 (vreinterpretq_u16_f32): Likewise.
6173 (vreinterpretq_u16_u8): Likewise.
6174 (vreinterpretq_u16_u32): Likewise.
6175 (vreinterpretq_u16_u64): Likewise.
6176 (vreinterpretq_u16_p8): Likewise.
6177 (vreinterpretq_u16_p16): Likewise.
6178 (vreinterpret_u32_s8): Likewise.
6179 (vreinterpret_u32_s16): Likewise.
6180 (vreinterpret_u32_s32): Likewise.
6181 (vreinterpret_u32_s64): Likewise.
6182 (vreinterpret_u32_f32): Likewise.
6183 (vreinterpret_u32_u8): Likewise.
6184 (vreinterpret_u32_u16): Likewise.
6185 (vreinterpret_u32_u64): Likewise.
6186 (vreinterpret_u32_p8): Likewise.
6187 (vreinterpret_u32_p16): Likewise.
6188 (vreinterpretq_u32_s8): Likewise.
6189 (vreinterpretq_u32_s16): Likewise.
6190 (vreinterpretq_u32_s32): Likewise.
6191 (vreinterpretq_u32_s64): Likewise.
6192 (vreinterpretq_u32_f32): Likewise.
6193 (vreinterpretq_u32_u8): Likewise.
6194 (vreinterpretq_u32_u16): Likewise.
6195 (vreinterpretq_u32_u64): Likewise.
6196 (vreinterpretq_u32_p8): Likewise.
6197 (vreinterpretq_u32_p16): Likewise.
6199 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6201 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
6203 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
6205 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
6206 (vqnegd_s64): Likewise.
6207 (vqabs_s64): Likewise.
6208 (vqabsd_s64): Likewise.
6210 2014-04-22 Richard Henderson <rth@redhat.com>
6212 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
6213 computation to the top of the loop.
6215 2014-04-22 Renlin <renlin.li@arm.com>
6216 Jiong Wang <jiong.wang@arm.com>
6218 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
6219 * config/aarch64/aarch64.c (aarch64_layout_frame)
6220 (aarch64_initial_elimination_offset): Likewise.
6222 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
6224 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
6227 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
6229 * machmode.h (bitwise_mode_for_mode): Declare.
6230 * stor-layout.h (bitwise_type_for_mode): Likewise.
6231 * stor-layout.c (bitwise_mode_for_mode): New function.
6232 (bitwise_type_for_mode): Likewise.
6233 * builtins.c (fold_builtin_memory_op): Use it instead of
6234 int_mode_for_mode and build_nonstandard_integer_type.
6236 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6238 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
6239 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
6240 (*-*-solaris2*): Simplify.
6241 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
6242 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
6243 *-*-solaris2.9* handling.
6245 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
6247 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
6248 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
6250 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
6251 * configure: Regenerate.
6253 * config/i386/sol2-9.h: Remove.
6255 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
6256 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
6257 Remove Solaris 9 references.
6259 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
6261 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
6262 (floatuns<GPI:mode><GPF:mode>2): Remove.
6263 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
6264 and floatuns conversions.
6265 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
6266 and floatuns conversions.
6267 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
6268 (w1,w2): New mode attributes for inequal width conversions.
6270 2014-04-22 Renlin Li <Renlin.Li@arm.com>
6272 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
6273 the output asm format.
6275 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
6277 * config/aarch64/aarch64-simd.md
6278 (aarch64_cm<optab>di): Always split.
6279 (*aarch64_cm<optab>di): New.
6280 (aarch64_cmtstdi): Always split.
6281 (*aarch64_cmtstdi): New.
6283 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6285 PR tree-optimization/60823
6286 * omp-low.c (ipa_simd_modify_function_body): Go through
6287 all SSA_NAMEs and for those refering to vector arguments
6288 which are going to be replaced adjust SSA_NAME_VAR and,
6289 if it is a default definition, change it into a non-default
6290 definition assigned at the beginning of function from new_decl.
6291 (ipa_simd_modify_stmt_ops): Rewritten.
6292 * tree-dfa.c (set_ssa_default_def): When removing default def,
6293 check for NULL loc instead of NULL *loc.
6295 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6297 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
6298 restrictions on core registers for DImode values in Thumb2.
6300 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6302 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
6303 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
6305 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6307 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
6308 (*iordi_notzesidi_di): Likewise.
6309 (*iordi_notsesidi_di): Likewise.
6311 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6313 * config/arm/arm-protos.h (tune_params): New struct members.
6314 * config/arm/arm.c: Initialise tune_params per processor.
6315 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
6316 for speed, based on new tune_params.
6318 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6320 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
6321 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
6322 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
6323 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
6324 * config/aarch64/arm_neon.h (vrnd_f64): Added.
6325 (vrnda_f64): Likewise.
6326 (vrndi_f64): Likewise.
6327 (vrndm_f64): Likewise.
6328 (vrndn_f64): Likewise.
6329 (vrndp_f64): Likewise.
6330 (vrndx_f64): Likewise.
6332 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6334 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
6335 GET_MODE_SIZE argument is enum machine_mode.
6337 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6340 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
6341 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
6343 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
6346 * asan.c (asan_emit_stack_protection): Force the base to align to
6347 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
6348 appropriate bits if STRICT_ALIGNMENT.
6349 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
6351 (expand_used_vars): Leave a space in the stack frame for alignment
6352 if STRICT_ALIGNMENT.
6354 2014-04-21 David Malcolm <dmalcolm@redhat.com>
6356 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
6358 (gimple_store_p): Likewise.
6359 (gimple_assign_load_p): Likewise.
6360 (gimple_assign_cast_p): Likewise.
6361 (gimple_clobber_p): Likewise.
6363 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
6364 rather than a gimple.
6365 (gimple_assign_cast_p): Likewise.
6367 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
6370 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
6371 If mode is DDmode and TARGET_E500_DOUBLE allow move.
6373 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
6374 more debug information for E500 if -mdebug=reg.
6376 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
6379 * config/i386/i386.c (ix86_expand_builtin)
6380 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
6381 register for target RTX.
6382 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
6384 2014-04-18 Cong Hou <congh@google.com>
6386 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
6387 the widen-mult pattern by handling two operands with different sizes,
6388 and operands whose size is smaller than half of the result type.
6390 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6392 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
6393 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
6394 (do_estimate_edge_time): Compute it.
6395 * ipa-inline.c (want_inline_small_function_p): Bypass
6396 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
6398 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6400 * ipa-inline.c (spec_rem): New static variable.
6401 (dump_overall_stats): New function.
6402 (dump_inline_stats): New function.
6404 2014-04-18 Richard Henderson <rth@redhat.com>
6406 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
6407 to GET_MODE_SIZE, not a reg_class_t.
6409 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6411 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
6412 (vsx_xxmrglw_<mode>): Likewise.
6414 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
6417 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
6418 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
6419 (rs6000_init_hard_regno_mode_ok): Likewise.
6421 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
6423 * ipa-inline.c (inline_small_functions): Account only non-cold
6425 * doc/invoke.texi (inline-unit-growth): Update documentation.
6427 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
6429 * config/rs6000/rs6000.md (addti3, subti3): New.
6431 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
6434 * config/i386/i386.c (ix86_expand_clear): Remove outdated
6435 comment. Check optimize_insn_for_size_p instead of
6436 optimize_insn_for_speed_p.
6438 2014-04-17 Martin Jambor <mjambor@suse.cz>
6440 * gimple-iterator.c (gsi_start_edge): New function.
6441 * gimple-iterator.h (gsi_start_edge): Declare.
6442 * tree-sra.c (single_non_eh_succ): New function.
6443 (disqualify_ops_if_throwing_stmt): Renamed to
6444 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
6445 having one non-EH successor BB.
6446 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
6447 generate loads into replacements.
6448 (sra_modify_assign): Likewise and and also use the simple path for
6450 (sra_modify_function_body): Commit statements on edges.
6452 2014-04-17 Richard Biener <rguenther@suse.de>
6455 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
6456 comparison results and add clarifying comment.
6458 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6460 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
6461 (blank_mode): Initialize it.
6462 (emit_mode_size_inline, emit_mode_nunits_inline,
6463 emit_mode_inner_inline): New functions.
6464 (emit_insn_modes_h): Call them and surround their output with
6465 #if GCC_VERSION >= 4001 ... #endif.
6466 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
6467 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
6468 mode_* arrays if the argument is __builtin_constant_p.
6469 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
6470 is enum machine_mode.
6472 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6474 * passes.c (opt_pass::execute): Adjust.
6475 (pass_manager::execute_pass_mode_switching): Likewise.
6476 (early_local_passes::execute): Likewise.
6477 (execute_one_pass): Pass cfun to the pass's execute method.
6478 * tree-pass.h (opt_pass::execute): Add function * argument.
6479 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6480 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6481 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6482 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6483 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6484 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
6485 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
6486 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6487 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6488 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
6489 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
6490 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
6491 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
6492 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
6493 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6494 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6495 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6496 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
6497 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
6498 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6499 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6500 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6501 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6502 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6503 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6504 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6505 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6506 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6507 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6510 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6512 * passes.c (opt_pass::gate): Take function * argument.
6513 (gate_all_early_local_passes): Merge into
6514 (early_local_passes::gate): this.
6515 (gate_all_early_optimizations): Merge into
6516 (all_early_optimizations::gate): this.
6517 (gate_all_optimizations): Mege into
6518 (all_optimizations::gate): this.
6519 (gate_all_optimizations_g): Merge into
6520 (all_optimizations_g::gate): this.
6521 (gate_rest_of_compilation): Mege into
6522 (rest_of_compilation::gate): this.
6523 (gate_postreload): Merge into
6524 (postreload::gate): this.
6525 (dump_one_pass): Pass cfun to the pass's gate method.
6526 (execute_ipa_summary_passes): Likewise.
6527 (execute_one_pass): Likewise.
6528 (ipa_write_summaries_2): Likewise.
6529 (ipa_write_optimization_summaries_1): Likewise.
6530 (ipa_read_summaries_1): Likewise.
6531 (ipa_read_optimization_summaries_1): Likewise.
6532 (execute_ipa_stmt_fixups): Likewise.
6533 * tree-pass.h (opt_pass::gate): Add function * argument.
6534 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
6535 combine-stack-adj.c, combine.c, compare-elim.c,
6536 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6537 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
6538 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
6539 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
6540 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6541 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6542 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6543 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
6544 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
6545 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
6546 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6547 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6548 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
6549 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6550 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6551 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6552 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6553 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6554 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6555 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6556 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6557 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6558 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6559 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
6560 var-tracking.c, vtable-verify.c, web.c: Adjust.
6562 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6564 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
6565 * configure: Regenerate.
6567 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6569 * passes.c (dump_one_pass): don't check pass->has_gate.
6570 (execute_ipa_summary_passes): Likewise.
6571 (execute_one_pass): Likewise.
6572 (ipa_write_summaries_2): Likewise.
6573 (ipa_write_optimization_summaries_1): Likewise.
6574 (ipa_read_optimization_summaries_1): Likewise.
6575 (execute_ipa_stmt_fixups): Likewise.
6576 * tree-pass.h (pass_data::has_gate): Remove.
6577 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6578 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6579 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6580 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6581 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6582 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
6583 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
6584 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
6585 gimple-low.c, gimple-ssa-isolate-paths.c,
6586 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
6587 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6588 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6589 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
6590 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
6591 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
6592 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
6593 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6594 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
6595 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6596 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6597 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6598 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6599 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6600 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6601 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6602 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6603 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6604 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6605 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6606 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6607 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6610 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6612 * pass_manager.h (pass_manager::register_dump_files_1): Remove
6614 * passes.c (pass_manager::register_dump_files_1): Merge into
6615 (pass_manager::register_dump_files): this, and remove its handling of
6616 properties since the pass always has the properties anyway.
6617 (pass_manager::pass_manager): Adjust.
6619 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6621 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
6622 * passes.c (pass_manager::register_dump_files_1): Remove dead code
6623 dealing with properties.
6624 (pass_manager::register_dump_files): Adjust.
6626 2014-03-20 Mark Wielaard <mjw@redhat.com>
6628 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
6629 then represent the bound as normal constant value.
6631 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6634 Forward port from 4.8 branch
6635 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
6637 * config/i386/bmiintrin.h (_blsi_u32): New.
6641 (_blsmsk_u32): Ditto.
6642 (_blsmsk_u64): Ditto.
6643 (_tzcnt_u32): Ditto.
6644 (_tzcnt_u64): Ditto.
6646 2014-04-17 Kito Cheng <kito@0xlab.org>
6648 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
6650 2014-04-17 Richard Biener <rguenther@suse.de>
6653 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
6654 boolean results for comparisons.
6656 2014-04-17 Richard Biener <rguenther@suse.de>
6658 PR tree-optimization/60836
6659 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
6660 initial PHI args to be gimple values.
6662 2014-04-17 Richard Biener <rguenther@suse.de>
6664 PR tree-optimization/60841
6665 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
6666 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
6667 of stmts to SLP build.
6668 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
6669 (vect_analyze_slp): Likewise.
6670 (vect_analyze_slp_instance): Likewise.
6671 (vect_build_slp_tree): Limit overall SLP tree growth.
6672 * tree-vectorizer.h (vect_analyze_data_refs,
6673 vect_analyze_slp): Adjust prototypes.
6675 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6677 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
6680 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6682 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
6683 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
6684 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
6685 for TARGET_SLOW_PSHUFB
6687 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6689 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
6690 * config/i386/i386.c (intel_cost): Ditto.
6692 2014-04-17 Joey Ye <joey.ye@arm.com>
6694 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
6696 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6698 * opts.c (common_handle_option): Disable -fipa-reference coorectly
6701 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6703 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
6704 (type_all_derivations_known_p): New predicate.
6705 (type_all_ctors_visible_p): New predicate.
6706 (type_possibly_instantiated_p): New predicate.
6707 (get_odr_type): Compute all_derivations_known.
6708 (dump_odr_type): Dump the flag.
6709 (maybe_record_type): Cleanup.
6710 (record_target_from_binfo): Add bases_to_consider array;
6711 record bases for types w/o instances and skip CXX destructor.
6712 (possible_polymorphic_call_targets_1): Add bases_to_consider
6713 and consider_construction parameters; check if type may have instance.
6714 (get_polymorphic_call_info): Set maybe_in_construction to true
6715 when we know nothing.
6716 (record_targets_from_bases): Skip CXX destructors; they are
6717 never called for types in construction.
6718 (possible_polymorphic_call_targets): Do not record target when
6719 type may not have instance.
6721 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6724 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6725 external aliases alive, too.
6727 2014-04-16 Andrew Pinski <apinski@cavium.com>
6729 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6732 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6734 * final.c (compute_alignments): Do not apply loop alignment to a block
6735 falling through to the exit.
6737 2014-04-16 Catherine Moore <clm@codesourcery.com>
6739 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6740 Adjust constraints for microMIPS store patterns.
6742 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6744 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6746 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6748 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6749 (append_use): Run at -O0.
6750 (append_vdef): Likewise.
6751 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6752 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6754 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6756 PR tree-optimization/60844
6757 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6758 (propagate_op_to_single_use, remove_visited_stmt_chain,
6759 linearize_expr, repropagate_negates, reassociate_bb): Use it
6760 instead of gsi_remove.
6762 2014-04-16 Martin Jambor <mjambor@suse.cz>
6764 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6765 ipa_transforms_to_apply.
6766 (cgraph_function_versioning): Assert that old_node has empty
6767 ipa_transforms_to_apply.
6768 * trans-mem.c (ipa_tm_create_version): Likewise.
6769 * tree-inline.c (tree_function_versioning): Do not duplicate
6770 ipa_transforms_to_apply.
6772 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6775 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6777 Pass necessary as flags on 64-bit Solaris/x86.
6778 Use lowercase relocs for x86_64-*-*.
6779 * configure: Regenerate.
6781 2014-04-15 Jan Hubicka <jh@suse.cz>
6783 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
6784 (maybe_record_node, likely_target_p): Use it.
6786 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6789 Revert following patch
6791 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6794 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6795 software floating point or no floating point registers, do not
6796 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6797 in GPRs that occurs after we tested for GPRs that would never be
6800 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6801 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6802 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6803 specifically allow DDmode, since that does not use the SPE SIMD
6806 2014-03-21 Mark Wielaard <mjw@redhat.com>
6808 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
6809 as unsigned or int depending on type and value used.
6811 2014-04-15 Richard Biener <rguenther@suse.de>
6813 PR rtl-optimization/56965
6814 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
6815 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
6817 * alias.c (true_dependence_1): Do not call
6818 nonoverlapping_component_refs_p.
6819 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
6820 nonoverlapping_component_refs_p.
6821 (indirect_refs_may_alias_p): Likewise.
6823 2014-04-15 Teresa Johnson <tejohnson@google.com>
6825 * cfg.c (dump_bb_info): Fix flags check.
6826 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
6828 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6830 PR rtl-optimization/60663
6831 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
6834 2014-04-15 Richard Biener <rguenther@suse.de>
6836 * lto-streamer.h (LTO_major_version): Bump to 4.
6838 2014-04-15 Richard Biener <rguenther@suse.de>
6840 * common.opt (lto_partition_model): New enum.
6841 (flto-partition=): Merge separate options with a single with argument,
6842 add -flto-partition=one support.
6843 * flag-types.h (enum lto_partition_model): Declare.
6844 * opts.c (finish_options): Remove duplicate -flto-partition=
6846 * lto-wrapper.c (run_gcc): Adjust.
6848 2014-04-15 Richard Biener <rguenther@suse.de>
6850 * alias.c (ncr_compar): New function.
6851 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
6853 2014-04-15 Richard Biener <rguenther@suse.de>
6855 * alias.c (record_component_aliases): Do not walk BINFOs.
6857 2014-04-15 Richard Biener <rguenther@suse.de>
6859 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6860 Add struct function argument and adjust.
6861 (find_func_aliases_for_call): Likewise.
6862 (find_func_aliases): Likewise.
6863 (find_func_clobbers): Likewise.
6864 (intra_create_variable_infos): Likewise.
6865 (compute_points_to_sets): Likewise.
6866 (ipa_pta_execute): Adjust. Do not push/pop cfun.
6868 2014-04-15 Richard Biener <rguenther@suse.de>
6870 * tree.c (iterative_hash_expr): Use enum tree_code_class
6871 to store TREE_CODE_CLASS.
6872 (tree_block): Likewise.
6873 (tree_set_block): Likewise.
6874 * tree.h (fold_build_pointer_plus_loc): Use
6875 convert_to_ptrofftype_loc.
6877 2014-04-15 Jakub Jelinek <jakub@redhat.com>
6880 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
6883 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
6885 * cfgloop.h (struct loop): Move force_vectorize down.
6886 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
6887 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
6888 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
6889 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
6890 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
6891 * tree-core.h (enum annot_expr_kind): Add new kind values.
6892 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
6893 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
6895 * tree.def (ANNOTATE_EXPR): Tweak comment.
6897 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6899 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
6902 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
6904 * tree.h (TYPE_IDENTIFIER): Declare.
6905 * tree.c (subrange_type_for_debug_p): Use it.
6906 * godump.c (go_format_type): Likewise.
6907 * dwarf2out.c (is_cxx_auto, modified_type_die,
6908 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
6909 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
6911 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6914 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
6916 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
6918 * config/i386/i386.c (examine_argument): Return bool. Return true if
6919 parameter should be passed in memory.
6920 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
6921 (construct_container): Update calls to examine_argument.
6922 (function_arg_advance_64): Ditto.
6923 (return_in_memory_32): Merge with ix86_return_in_memory.
6924 (return_in_memory_64): Ditto.
6925 (return_in_memory_ms_64): Ditto.
6927 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6929 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
6930 * coverage.c (coverage_compute_profile_id): Handle externally visible
6933 2014-04-14 Martin Jambor <mjambor@suse.cz>
6935 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
6936 DECL_DISREGARD_INLINE_LIMITS functions.
6938 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6941 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
6943 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6946 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
6947 optimize_insn_for_speed_p instead of
6948 optimize_function_for_speed_p.
6950 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
6952 * doc/invoke.texi (free): Document AArch64.
6954 2014-04-14 Richard Biener <rguenther@suse.de>
6956 PR tree-optimization/60042
6957 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
6958 (insert_into_preds_of_block): Do not prevent PHI insertion
6959 for REFERENCE exprs here ...
6960 (eliminate_dom_walker::before_dom_children): ... but prevent
6961 their use here under similar conditions when applied to the
6962 IL after PRE optimizations.
6964 2014-04-14 Richard Biener <rguenther@suse.de>
6966 * passes.def: Move early points-to after early SRA.
6968 2014-04-14 Richard Biener <rguenther@suse.de>
6970 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
6971 check for which sign-changes we allow when forwarding
6972 a converted value into a switch.
6974 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6976 * stor-layout.c (place_field): Finalize non-constant offset for the
6979 2014-04-14 Richard Biener <rguenther@suse.de>
6981 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
6983 (expand_switch_using_bit_tests_p): Likewise.
6984 (process_switch): Compute and pass on speed_p based on the
6986 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
6987 optimize_bb_for_speed_p.
6989 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6991 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
6992 * function.h (struct function): Rename has_force_vect_loops into
6993 has_force_vectorize_loops.
6994 * lto-streamer-in.c (input_cfg): Adjust for renaming.
6995 (input_struct_function_base): Likewise.
6996 * lto-streamer-out.c (output_cfg): Likewise.
6997 (output_struct_function_base): Likewise.
6998 * omp-low.c (expand_omp_simd): Likewise.
6999 * tree-cfg.c (move_sese_region_to_fn): Likewise.
7000 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
7001 (version_loop_for_if_conversion): Likewise.
7002 (tree_if_conversion): Likewise.
7003 (main_tree_if_conversion): Likewise.
7004 (gate_tree_if_conversion): Likewise.
7005 * tree-inline.c (copy_loops): Likewise.
7006 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
7007 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
7008 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
7009 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
7010 * tree-vectorizer.c (vectorize_loops): Likewise.
7011 * tree-vectorizer.h (unlimited_cost_model): Likewise.
7013 2014-04-14 Richard Biener <rguenther@suse.de>
7016 * lto-streamer-out.c (wrap_refs): New function.
7017 (lto_output): Wrap symbol references in global initializes in
7018 type-preserving MEM_REFs.
7020 2014-04-14 Christian Bruel <christian.bruel@st.com>
7022 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
7024 2014-04-14 Christian Bruel <christian.bruel@st.com>
7026 * config/sh/sh.md (setmemqi): New expand pattern.
7027 * config/sh/sh.h (CLEAR_RATIO): Define.
7028 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
7029 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
7031 2014-04-14 Richard Biener <rguenther@suse.de>
7034 * fold-const.c (negate_expr_p): Don't negate directional rounding
7036 (fold_negate_expr): Likewise.
7038 2014-04-14 Richard Biener <rguenther@suse.de>
7040 PR tree-optimization/59817
7041 PR tree-optimization/60453
7042 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
7043 recursion to catch all CHRECs in the scalar evolution and restrict
7044 the predicate for the remains appropriately.
7046 2014-04-12 Catherine Moore <clm@codesourcery.com>
7048 * config/mips/constraints.md: Add new register constraint "kb".
7049 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
7050 (*movhi_internal): Likewise.
7051 (*movqi_internal): Likewise.
7052 * config/mips/mips.h (M16_STORE_REGS): New register class.
7053 (REG_CLASS_NAMES): Add M16_STORE_REGS.
7054 (REG_CLASS_CONTENTS): Likewise.
7055 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
7057 2014-04-11 Tobias Burnus <burnus@net-b.de>
7060 * doc/invoke.texi (-Wformat-signedness): Document it.
7061 (Wformat=2): Mention that this enables -Wformat-signedness.
7063 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7065 * common/config/epiphany/epiphany-common.c
7066 (epiphany_option_optimization_table): Enable section anchors by
7067 default at -O1 or higher.
7068 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
7069 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
7070 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
7071 carries no extra cost.
7072 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
7073 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
7074 * config/epiphany/predicates.md (memclob_operand): New predicate.
7075 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
7076 Use memclob_operand predicate and X constraint for operand 3.
7078 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7080 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
7081 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
7084 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7086 PR rtl-optimization/60651
7087 * mode-switching.c (optimize_mode_switching): Make sure to emit
7088 sets of a lower numbered entity before sets of a higher numbered
7089 entity to a mode of the same or lower priority.
7090 When creating a seginfo for a basic block that starts with a code
7091 label, move the insertion point past the code label.
7092 (new_seginfo): Document and enforce requirement that
7093 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
7094 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
7095 * doc/tm.texi: Regenerate.
7097 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
7100 * config/arc/arc.c (arc_save_restore): Fix assert typo.
7102 2013-04-11 Jakub Jelinek <jakub@redhat.com>
7104 * BASE-VER: Set to 4.10.0.
7106 2014-04-11 Tobias Burnus <burnus@net-b.de>
7109 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
7110 * doc/gcc.texi (Service): Update description in the @menu
7111 * doc/invoke.texi (Option Summary): Remove misplaced and
7114 2014-04-11 Steve Ellcey <sellcey@mips.com>
7115 Jakub Jelinek <jakub@redhat.com>
7118 * expr.c (convert_move): Use emit_store_flag_force instead of
7119 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
7122 2014-04-11 Richard Biener <rguenther@suse.de>
7125 * varasm.c (assemble_alias): Avoid endless error reporting
7126 recursion by setting TREE_ASM_WRITTEN.
7128 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7130 * config/s390/s390.md: Add a splitter for NOT rtx.
7132 2014-04-11 Jakub Jelinek <jakub@redhat.com>
7134 PR rtl-optimization/60663
7135 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
7137 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
7138 Jakub Jelinek <jakub@redhat.com>
7141 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
7142 flag from decl_node to node.
7144 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7147 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
7148 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
7149 ameliorating the cases where it can be.
7151 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
7154 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7156 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7157 (loadsync_<mode>): Change mode.
7158 (load_quadpti, store_quadpti): New.
7159 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7160 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7161 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
7163 2014-04-09 Cong Hou <congh@google.com>
7166 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
7169 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7171 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
7172 instead of vnor to exploit possible fusion opportunity in the
7174 (altivec_expand_vec_perm_const_le): Likewise.
7176 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7178 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7179 (loadsync_<mode>): Change mode.
7180 (load_quadpti, store_quadpti): New.
7181 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7182 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7184 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
7187 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
7188 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
7189 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
7191 2014-04-08 Richard Biener <rguenther@suse.de>
7194 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
7195 a 64bit widest int print double-int similar to on HWI64 hosts.
7197 2014-04-08 Richard Biener <rguenther@suse.de>
7199 PR tree-optimization/60785
7200 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
7201 default defs properly.
7203 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
7205 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
7206 (Weffc++): Likewise.
7208 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
7210 * ipa-devirt.c (maybe_record_node): When node is not recorded,
7211 set completep to false rather than true.
7213 2014-04-07 Douglas B Rupp <rupp@adacore.com>
7216 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
7217 ARM_TARGET2_DWARF_FORMAT.
7219 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
7222 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
7223 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
7226 2014-04-07 Richard Biener <rguenther@suse.de>
7228 PR tree-optimization/60766
7229 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
7230 (may_eliminate_iv): Convert cand_value_at result to desired type.
7232 2014-04-07 Jason Merrill <jason@redhat.com>
7235 * common.opt (-fno-gnu-unique): Add.
7236 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
7238 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7240 * haifa-sched.c: Fix outdated function reference and minor
7241 grammar errors in introductory comment.
7243 2014-04-07 Richard Biener <rguenther@suse.de>
7246 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
7248 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
7250 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
7253 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
7254 size accounting for thunks.
7255 (pa_asm_output_mi_thunk): Use final_start_function() and
7256 final_end_function() to output function start and end directives.
7258 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
7260 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
7261 device specific ISA/ feature information. Remove short_sp and
7262 errata_skip ds. Add avr_device_specific_features enum to have device
7264 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
7265 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
7266 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
7267 updated device specific info.
7268 * config/avr/avr-mcus.def: Merge device specific details to
7269 dev_attribute field.
7270 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
7272 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
7273 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
7274 assembler if RMW isa supported by current device.
7275 * config/avr/genmultilib.awk: Update as device info structure changed.
7276 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
7278 2014-04-04 Cong Hou <congh@google.com>
7280 PR tree-optimization/60656
7281 * tree-vect-stmts.c (supportable_widening_operation):
7282 Fix a bug that elements in a vector with vect_used_by_reduction
7283 property are incorrectly reordered when the operation on it is not
7284 consistant with the one in reduction operation.
7286 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
7288 PR rtl-optimization/60155
7289 * gcse.c (record_set_data): New function.
7290 (single_set_gcse): New function.
7291 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
7292 (hoist_code): Likewise.
7293 (get_pressure_class_and_nregs): Likewise.
7295 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
7297 * explow.c (probe_stack_range): Emit a final optimization blockage.
7299 2014-04-04 Anthony Green <green@moxielogic.com>
7301 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
7304 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
7307 * lto-cgraph.c (input_overwrite_node): Check that partitioning
7308 flags are set only during streaming.
7309 * ipa.c (process_references, walk_polymorphic_call_targets,
7310 symtab_remove_unreachable_nodes): Drop bodies of always inline
7311 after early inlining.
7312 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
7314 2014-04-04 Jakub Jelinek <jakub@redhat.com>
7315 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7318 * dwarf2out.c (const_ok_for_output_1): Reject expressions
7321 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7324 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
7326 (cortex_a53_fdivd): Likewise.
7328 2014-04-04 Martin Jambor <mjambor@suse.cz>
7331 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
7333 * cgraph.c (clone_of_p): Also return true if thunks match.
7334 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
7335 cgraph_function_or_thunk_node and an obsolete comment.
7336 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
7338 (build_function_decl_skip_args): Likewise.
7339 (set_new_clone_decl_and_node_flags): New function.
7340 (duplicate_thunk_for_node): Likewise.
7341 (redirect_edge_duplicating_thunks): Likewise.
7342 (cgraph_clone_node): New parameter args_to_skip, pass it to
7343 redirect_edge_duplicating_thunks which is called instead of
7344 cgraph_redirect_edge_callee.
7345 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
7346 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
7348 2014-04-04 Jeff Law <law@redhat.com>
7351 * config/arm/predicates.md (const_int_I_operand): New predicate.
7352 (const_int_M_operand): Similarly.
7353 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
7355 (insv_t2, extv_reg, extzv_t2): Likewise.
7356 (load_multiple_with_writeback): Similarly for const_int_I_operand.
7357 (pop_multiple_with_writeback_and_return): Likewise.
7358 (vfp_pop_multiple_with_writeback): Likewise
7360 2014-04-04 Richard Biener <rguenther@suse.de>
7363 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
7364 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
7366 * gimplify.h (gimple_add_tmp_var_fn): Declare.
7367 * gimplify.c (gimple_add_tmp_var_fn): New function.
7368 * gimple-expr.h (create_tmp_reg_fn): Declare.
7369 * gimple-expr.c (create_tmp_reg_fn): New function.
7370 * gimple-low.c (record_vars_into): Don't change cfun.
7371 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
7372 code generation without cfun.
7374 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
7377 * Makefile.in (install-driver): Fix shell scripting.
7379 2014-04-03 Cong Hou <congh@google.com>
7381 PR tree-optimization/60505
7382 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
7383 threshold of number of iterations below which no vectorization
7385 * tree-vect-loop.c (new_loop_vec_info):
7386 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
7387 * tree-vect-loop.c (vect_analyze_loop_operations):
7388 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
7389 * tree-vect-loop.c (vect_transform_loop):
7390 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
7391 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
7392 of iterations of the loop and see if we should build the epilogue.
7394 2014-04-03 Richard Biener <rguenther@suse.de>
7396 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
7397 (streamer_tree_cache_create): Adjust.
7398 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
7399 to allow optional nodes array.
7400 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
7401 (streamer_tree_cache_append): Likewise.
7402 (streamer_tree_cache_create): Create nodes array optionally
7403 as specified by parameter.
7404 * lto-streamer-out.c (create_output_block): Avoid maintaining
7405 the node array in the writer cache.
7406 (DFS_write_tree): Remove assertion.
7407 (produce_asm_for_decls): Free the out decl state hash table early.
7408 * lto-streamer-in.c (lto_data_in_create): Adjust for
7409 streamer_tree_cache_create prototype change.
7411 2014-04-03 Richard Biener <rguenther@suse.de>
7413 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
7414 set TREE_CHAIN to NULL_TREE.
7416 2014-04-03 Richard Biener <rguenther@suse.de>
7418 PR tree-optimization/60740
7419 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
7420 over all GIMPLE_COND operands.
7422 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
7424 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
7425 (Weffc++): Remove Scott's numbering, merge lists and reference
7428 2014-04-03 Nick Clifton <nickc@redhat.com>
7430 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
7433 2014-04-03 Martin Jambor <mjambor@suse.cz>
7435 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
7436 mention gcc_unreachable before failing.
7437 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
7440 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
7443 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
7444 inconsistent code and instead mark the context inconsistent.
7445 (possible_polymorphic_call_targets): For inconsistent contexts
7446 return empty complete list.
7448 2014-04-02 Anthony Green <green@moxielogic.com>
7450 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
7451 (extendqisi2, extendhisi2): Define.
7452 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
7453 (WCHAR_TYPE): Change to unsigned int.
7455 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7457 PR tree-optimization/60733
7458 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
7459 insertion point for PHI candidates to be the end of the feeding
7460 block for the PHI argument.
7462 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
7464 PR rtl-optimization/60650
7465 * lra-constraints.c (process_alt_operands): Decrease reject for
7466 earlyclobber matching.
7468 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7470 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
7472 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7474 * config/spu/spu.c (pad_bb): Do not crash when the last
7475 insn is CODE_FOR_blockage.
7477 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7479 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
7480 lies outside the target mode.
7482 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7485 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7486 software floating point or no floating point registers, do not
7487 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7488 in GPRs that occurs after we tested for GPRs that would never be
7491 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7492 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7493 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7494 specifically allow DDmode, since that does not use the SPE SIMD
7497 2014-04-02 Richard Biener <rguenther@suse.de>
7500 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
7501 MODE_INTs. Properly use negv_optab.
7502 (expand_abs): Likewise.
7504 2014-04-02 Richard Biener <rguenther@suse.de>
7507 * Makefile.in (install-driver): Guard extra installs with special
7510 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
7512 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7515 2014-04-01 Richard Henderson <rth@redhat.com>
7518 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
7519 alternative enabled before register allocation.
7521 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
7523 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
7524 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
7526 (nios2_large_got_address): Remove unneeded 'sym' parameter.
7527 (nios2_got_address): Update nios2_large_got_address call site.
7528 (nios2_delegitimize_address): New function.
7529 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
7530 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
7531 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
7533 2014-04-01 Martin Husemann <martin@duskware.de>
7535 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
7538 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
7540 PR rtl-optimization/60604
7541 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
7542 check from register_operand.
7543 (register_operand): Redefine in terms of general_operand.
7544 (nonmemory_operand): Use register_operand for the non-constant cases.
7546 2014-04-01 Richard Biener <rguenther@suse.de>
7548 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
7550 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
7552 * doc/invoke.texi (mapp-regs): Clarify.
7554 2014-03-31 Ulrich Drepper <drepper@gmail.com>
7556 * config/i386/avx512fintrin.h (__v32hi): Define type.
7557 (__v64qi): Likewise.
7558 (_mm512_set1_epi8): Define.
7559 (_mm512_set1_epi16): Define.
7560 (_mm512_set4_epi32): Define.
7561 (_mm512_set4_epi64): Define.
7562 (_mm512_set4_pd): Define.
7563 (_mm512_set4_ps): Define.
7564 (_mm512_setr4_epi64): Define.
7565 (_mm512_setr4_epi32): Define.
7566 (_mm512_setr4_pd): Define.
7567 (_mm512_setr4_ps): Define.
7568 (_mm512_setzero_epi32): Define.
7570 2014-03-31 Martin Jambor <mjambor@suse.cz>
7573 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
7574 callsite_arguments_match_p. Updated all callers. Also check types of
7575 corresponding formal parameters and actual arguments.
7576 (not_all_callers_have_enough_arguments_p) Renamed to
7577 some_callers_have_mismatched_arguments_p.
7579 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
7581 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
7583 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
7586 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
7589 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
7591 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
7592 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
7594 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
7595 Use FMAMODE_NOVF512 mode iterator.
7596 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
7597 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
7598 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
7599 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
7601 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
7602 Use VF_128_256 mode iterator.
7603 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
7606 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7608 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
7609 static chain if needed.
7611 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7614 * lra-constraints.c (index_part_to_reg): New.
7615 (process_address): Use it.
7617 2014-03-27 Jeff Law <law@redhat.com>
7618 Jakub Jelinek <jakub@redhat.com>
7621 * expr.c (do_tablejump): Use simplify_gen_binary rather than
7622 gen_rtx_{PLUS,MULT} to build up the address expression.
7624 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
7625 creating non-canonical RTL.
7627 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7630 * ipa-inline.c (want_inline_small_function_p): Short circuit large
7631 functions; reorganize to make cheap checks first.
7632 (inline_small_functions): Do not estimate growth when dumping;
7634 * ipa-inline.h (inline_summary): Add min_size.
7635 (growth_likely_positive): New function.
7636 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
7637 (set_cond_stmt_execution_predicate): Cleanup.
7638 (estimate_edge_size_and_time): Compute min_size.
7639 (estimate_calls_size_and_time): Likewise.
7640 (estimate_node_size_and_time): Likewise.
7641 (inline_update_overall_summary): Update min_size.
7642 (do_estimate_edge_time): Likewise.
7643 (do_estimate_edge_size): Update.
7644 (do_estimate_edge_hints): Update.
7645 (growth_likely_positive): New function.
7647 2014-03-28 Jakub Jelinek <jakub@redhat.com>
7650 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
7651 also if addr has VOIDmode.
7653 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7655 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
7656 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
7658 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
7659 instructions as well as AdvancedSIMD loads.
7661 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7663 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
7664 Use crypto_aese type.
7665 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
7666 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
7667 crypto_aese, crypto_aesmc. Move to types.md.
7668 * config/arm/types.md (crypto_aes): Split into crypto_aese,
7670 * config/arm/iterators.md (crypto_type): Likewise.
7672 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7674 * cgraph.c: Include expr.h and tree-dfa.h.
7675 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
7678 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7681 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
7682 regs from checking multi-reg pseudos.
7684 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7686 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
7688 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7690 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
7691 if it would clobber the stack pointer, even temporarily.
7693 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
7695 * mode-switching.c: Make small adjustments to the top comment.
7697 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7699 * config/rs6000/constraints.md (wD constraint): New constraint to
7700 match the constant integer to get the top DImode/DFmode out of a
7701 vector in a VSX register.
7703 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
7704 match the constant integer to get the top DImode/DFmode out of a
7705 vector in a VSX register.
7707 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
7710 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7713 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
7714 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
7716 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
7717 Optimize vec_extract of 64-bit values, where the value being
7718 extracted is in the top word, where we can use scalar
7719 instructions. Add direct move and store support. Combine the big
7720 endian/little endian vector select load support into a single insn.
7721 (vsx_extract_<mode>_internal1): Likewise.
7722 (vsx_extract_<mode>_internal2): Likewise.
7723 (vsx_extract_<mode>_load): Likewise.
7724 (vsx_extract_<mode>_store): Likewise.
7725 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7726 combined into vsx_extract_<mode>_load.
7727 (vsx_extract_<mode>_one_le): Likewise.
7729 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7730 define the top 64-bit vector element.
7732 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7735 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7736 Document vec_vbpermq builtin.
7739 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7740 enable use of xxsldwi and xxpermdi builtin functions.
7741 (vec_xxpermdi): Likewise.
7743 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7744 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7746 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7748 PR rtl-optimization/60650
7749 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7751 (find_spills_for): New.
7752 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7753 Spill all pseudos on the second iteration.
7755 2014-03-27 Marek Polacek <polacek@redhat.com>
7758 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7761 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7763 * config/s390/s390.c (s390_can_use_return_insn): Check for
7764 call-saved FPRs on 31 bit.
7766 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7769 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7770 if they need regimplification, just drop them instead of
7771 calling gimple_regimplify_operands on them.
7773 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7776 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7777 (aarch64_frame_pointer_required): Adjust logic.
7778 (aarch64_can_eliminate): Adjust logic.
7779 (aarch64_override_options_after_change): Adjust logic.
7781 2014-03-27 Dehao Chen <dehao@google.com>
7783 * ipa-inline.c (early_inliner): Update node's inline info.
7785 2014-03-26 Dehao Chen <dehao@google.com>
7787 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
7788 compiler inserted conditional jumps for NAN float check.
7790 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7792 * ubsan.h (ubsan_create_data): Change second argument's type
7793 to const location_t *.
7794 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
7796 (ubsan_create_data): Change second argument to const location_t *PLOC.
7797 Create Loc field whenever PLOC is non-NULL.
7798 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
7799 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
7803 * real.c (real_to_integer2): Change type of low to UHWI.
7805 2014-03-26 Tobias Burnus <burnus@net-b.de>
7807 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
7808 (CILK_SELF_SPECS): New define.
7809 (driver_self_specs): Use it.
7811 2014-03-26 Richard Biener <rguenther@suse.de>
7813 * tree-pretty-print.c (percent_K_format): Implement special
7814 case for LTO and its stripped down BLOCK tree.
7816 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7819 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
7821 * tree-vrp.c (simplify_internal_call_using_ranges): If only
7822 one range is range_int_cst_p, but not both, at least optimize
7823 addition/subtraction of 0 and multiplication by 0 or 1.
7824 * gimple-fold.c (gimple_fold_call): Fold
7825 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
7826 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
7827 INTEGER_CSTs, try to fold at least x * 0 and y - y.
7829 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
7831 PR rtl-optimization/60452
7832 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
7833 <case REG>: Return 1 for invalid offsets from the frame pointer.
7835 2014-03-26 Marek Polacek <polacek@redhat.com>
7838 * doc/extend.texi (C Extensions): Mention variable-length arrays in
7841 2014-03-26 Marek Polacek <polacek@redhat.com>
7844 * doc/extend.texi (Designated Inits): Describe what happens to omitted
7847 2014-03-26 Marek Polacek <polacek@redhat.com>
7850 * ira-color.c (update_conflict_hard_regno_costs): Perform the
7851 multiplication in unsigned type.
7853 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7855 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
7857 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7859 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
7861 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7864 * cif-code.def (UNREACHABLE) New code.
7865 * ipa-inline.c (inline_small_functions): Skip edges to
7866 __builtlin_unreachable.
7867 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
7868 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
7869 predicate to __bulitin_unreachable.
7870 (set_cond_stmt_execution_predicate): Fix issue when
7871 invert_tree_comparison returns ERROR_MARK.
7872 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
7873 propagate to inline clones.
7874 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
7876 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
7877 * cgraphclones.c (cgraph_clone_node): If call destination is already
7878 ureachable, do not redirect it back.
7879 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
7882 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7884 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
7885 Do not modify inline clones.
7887 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7889 * config/i386/i386.md (general_sext_operand): New mode attr.
7890 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
7891 don't generate (sign_extend (const_int)).
7892 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
7893 operands[2]. Use We constraint instead of <i> and
7894 <general_sext_operand> predicate instead of <general_operand>.
7895 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
7896 * config/i386/constraints.md (We): New constraint.
7897 * config/i386/predicates.md (x86_64_sext_operand,
7898 sext_operand): New predicates.
7900 2014-03-25 Martin Jambor <mjambor@suse.cz>
7903 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
7904 inconsistent devirtualizations to __builtin_unreachable.
7906 2014-03-25 Marek Polacek <polacek@redhat.com>
7909 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
7911 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
7913 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
7914 order of elements for big-endian.
7916 2014-03-25 Richard Biener <rguenther@suse.de>
7919 * gimplify-me.c (gimple_regimplify_operands): Update the
7922 2014-03-25 Martin Jambor <mjambor@suse.cz>
7925 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
7926 (lto_output_varpool_node): Likewise.
7927 (input_overwrite_node): Likewise.
7928 (input_varpool_node): Likewise.
7930 2014-03-25 Richard Biener <rguenther@suse.de>
7932 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
7933 (run_gcc): Likewise.
7935 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7937 * combine.c (simplify_compare_const): Add MODE argument.
7938 Handle mode_width 0 as very large mode_width.
7939 (try_combine, simplify_comparison): Adjust callers.
7941 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
7942 type to avoid signed integer overflow.
7943 * explow.c (plus_constant): Likewise.
7945 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
7947 * doc/generic.texi: Correct typos.
7949 2014-03-24 Tobias Burnus <burnus@net-b.de>
7951 * doc/invoke.texi (-flto): Expand section about
7952 using static libraries with LTO.
7954 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7956 PR rtl-optimization/60501
7957 * optabs.def (addptr3_optab): New optab.
7958 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
7959 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
7960 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
7962 * lra.c (emit_add3_insn): Use the addptr pattern if available.
7964 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
7966 2014-03-24 Ulrich Drepper <drepper@gmail.com>
7968 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
7971 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
7972 (_mm256_undefined_ps): Define.
7973 (_mm256_undefined_pd): Define.
7974 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
7975 (_mm_undefined_pd): Define.
7976 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
7977 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
7978 (_mm512_undefined_ps): Define.
7979 (_mm512_undefined_pd): Define.
7980 Use _mm*_undefined_*.
7981 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
7983 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
7985 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
7986 (lshr_simd): DI mode added.
7987 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
7988 (aarch64_ushr_simddi): Likewise.
7989 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
7990 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
7991 (vshrd_n_u64): Likewise.
7993 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7995 * Makefile.in (s-macro_list): Depend on cc1.
7997 2014-03-23 Teresa Johnson <tejohnson@google.com>
7999 * ipa-utils.c (ipa_print_order): Use specified dump file.
8001 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
8003 PR rtl-optimization/60601
8004 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
8006 * gcc.c (eval_spec_function): Initialize save_growing_value.
8008 2014-03-22 Jakub Jelinek <jakub@redhat.com>
8011 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
8012 code == MINUS_EXPR, never swap op0 with op1.
8014 * toplev.c (init_local_tick): Avoid signed integer multiplication
8016 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
8017 shift by first operand's bitsize.
8019 2014-03-21 Jakub Jelinek <jakub@redhat.com>
8022 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
8024 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
8025 TARGET_ISA_64BIT_P(x).
8027 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8029 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
8030 pattern for vector nor instead of subtract from splat(-1).
8031 (altivec_expand_vec_perm_const_le): Likewise.
8033 2014-03-21 Richard Henderson <rth@twiddle.net>
8036 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
8037 related insns after epilogue_completed.
8039 2014-03-21 Martin Jambor <mjambor@suse.cz>
8042 * cgraph.h (symtab_node): New flag body_removed.
8043 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
8044 when removing bodies.
8045 * symtab.c (dump_symtab_base): Dump body_removed flag.
8046 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
8047 had their bodies removed.
8049 2014-03-21 Martin Jambor <mjambor@suse.cz>
8052 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
8055 2014-03-21 Richard Biener <rguenther@suse.de>
8057 PR tree-optimization/60577
8058 * tree-core.h (struct tree_base): Document nothrow_flag use
8059 in VAR_DECL_NONALIASED.
8060 * tree.h (VAR_DECL_NONALIASED): New.
8061 (may_be_aliased): Adjust.
8062 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
8064 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8066 * expr.c (expand_expr_real_1): Remove outdated comment.
8068 2014-03-20 Jakub Jelinek <jakub@redhat.com>
8071 * ira.c (adjust_cleared_regs): Call copy_rtx on
8072 *reg_equiv[REGNO (loc)].src_p before passing it to
8073 simplify_replace_fn_rtx.
8076 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
8077 into CONST, put pic register as first operand of PLUS. Use
8078 gen_const_mem for both 32-bit and 64-bit PIC got loads.
8080 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8082 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
8084 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8086 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
8087 around for store forwarding issue in the FPU on the UT699.
8088 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
8089 loads and operations if -mfix-ut699 is specified.
8090 (divtf3_hq): Tweak attribute.
8091 (sqrttf2_hq): Likewise.
8093 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8095 * calls.c (store_one_arg): Remove incorrect const qualification on the
8096 type of the temporary.
8097 * cfgexpand.c (expand_return): Likewise.
8098 * expr.c (expand_constructor): Likewise.
8099 (expand_expr_real_1): Likewise.
8101 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8103 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
8106 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
8109 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
8111 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
8113 * config/arm/aarch-common-protos.h
8114 (alu_cost_table): Fix spelling of "extend".
8115 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
8117 2014-03-19 Richard Biener <rguenther@suse.de>
8120 * tree-core.h (tree_type_common): Re-order pointer members
8121 to reduce recursion depth during GC walks.
8123 2014-03-19 Marek Polacek <polacek@redhat.com>
8126 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
8127 before accessing it.
8129 2014-03-19 Richard Biener <rguenther@suse.de>
8132 * lto-streamer-in.c (input_function): In WPA stage do not drop
8135 2014-03-19 Jakub Jelinek <jakub@redhat.com>
8137 PR tree-optimization/60559
8138 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
8139 with build_zero_cst assignment.
8141 2014-03-18 Kai Tietz <ktietz@redhat.com>
8143 PR rtl-optimization/56356
8144 * sdbout.c (sdbout_parms): Verify that parms'
8145 incoming argument is valid.
8146 (sdbout_reg_parms): Likewise.
8148 2014-03-18 Richard Henderson <rth@redhat.com>
8151 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
8152 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
8153 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
8155 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
8157 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
8158 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
8159 Italicize plugin event names in description. Explain that
8160 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
8161 Remind that no GCC functions should be called after PLUGIN_FINISH.
8162 Explain what pragmas with expansion are.
8164 2014-03-18 Martin Liska <mliska@suse.cz>
8166 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
8167 gimple call statement is update.
8168 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
8169 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
8171 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8174 * ubsan.c (ubsan_instrument_unreachable): Call
8175 initialize_sanitizer_builtins.
8176 (ubsan_pass): Likewise.
8179 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
8180 varpool_finalize_decl instead of rest_of_decl_compilation.
8182 2014-03-18 Richard Biener <rguenther@suse.de>
8184 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
8185 by using bitmap_and_compl instead of bitmap_and_compl_into.
8186 (df_rd_transfer_function): Likewise.
8188 2014-03-18 Richard Biener <rguenther@suse.de>
8190 * doc/lto.texi (fresolution): Fix typo.
8192 2014-03-18 Richard Biener <rguenther@suse.de>
8194 * doc/invoke.texi (flto): Update for changes in 4.9.
8196 2014-03-18 Richard Biener <rguenther@suse.de>
8198 * doc/loop.texi: Remove section on the removed lambda framework.
8199 Update loop docs with recent changes in preserving loop structure.
8201 2014-03-18 Richard Biener <rguenther@suse.de>
8203 * doc/lto.texi (-fresolution): Document.
8205 2014-03-18 Richard Biener <rguenther@suse.de>
8207 * doc/contrib.texi: Adjust my name.
8209 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8212 * internal-fn.c: Include diagnostic-core.h.
8213 (expand_BUILTIN_EXPECT): New function.
8214 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
8215 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
8216 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
8217 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
8219 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
8220 Revert 3 argument __builtin_expect code.
8221 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
8222 * gimple-fold.c (gimple_fold_call): Likewise.
8223 * tree.h (fold_builtin_expect): New prototype.
8224 * builtins.c (build_builtin_expect_predicate): Add predictor
8225 argument, if non-NULL, create 3 argument __builtin_expect.
8226 (fold_builtin_expect): No longer static. Add ARG2 argument,
8227 pass it through to build_builtin_expect_predicate.
8228 (fold_builtin_2): Adjust caller.
8229 (fold_builtin_3): Handle BUILT_IN_EXPECT.
8230 * internal-fn.def (BUILTIN_EXPECT): New.
8232 2014-03-18 Tobias Burnus <burnus@net-b.de>
8235 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
8236 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
8237 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
8239 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
8242 * predict.c (combine_predictions_for_bb): Fix up formatting.
8243 (expr_expected_value_1, expr_expected_value): Add predictor argument,
8244 fill what it points to if non-NULL.
8245 (tree_predict_by_opcode): Adjust caller, use the predictor.
8246 * predict.def (PRED_COMPARE_AND_SWAP): Add.
8248 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
8250 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
8251 proper constant for the store mode.
8253 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
8255 * symtab.c (change_decl_assembler_name): Fix transparent alias
8258 2014-03-16 Renlin Li <Renlin.Li@arm.com>
8260 * config/aarch64/aarch64.c: Correct the comments about the
8261 aarch64 stack layout.
8263 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
8265 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
8266 check for GF_OMP_FOR_KIND_FOR.
8268 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
8270 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
8271 ymm and zmm register names.
8273 2014-03-17 Jakub Jelinek <jakub@redhat.com>
8276 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
8277 note creation for the 2010-08-31 changes.
8279 2014-03-17 Marek Polacek <polacek@redhat.com>
8282 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
8283 as -fno-tree-loop-vectorize.
8284 (expand_omp_simd): Likewise.
8286 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
8288 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
8289 (eligible_for_call_delay): New prototype.
8290 * config/sparc/sparc.c (tls_call_delay): Rename into...
8291 (eligible_for_call_delay): ...this. Return false if the instruction
8292 cannot be put in the delay slot of a branch.
8293 (eligible_for_restore_insn): Simplify.
8294 (eligible_for_return_delay): Return false if the instruction cannot be
8295 put in the delay slot of a branch and simplify.
8296 (eligible_for_sibcall_delay): Return false if the instruction cannot be
8297 put in the delay slot of a branch.
8298 * config/sparc/sparc.md (fix_ut699): New attribute.
8299 (tls_call_delay): Delete.
8300 (in_call_delay): Reimplement.
8301 (eligible_for_sibcall_delay): Rename into...
8302 (in_sibcall_delay): ...this.
8303 (eligible_for_return_delay): Rename into...
8304 (in_return_delay): ...this.
8305 (in_branch_delay): Reimplement.
8306 (in_uncond_branch_delay): Delete.
8307 (in_annul_branch_delay): Delete.
8309 2014-03-14 Richard Henderson <rth@redhat.com>
8312 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
8313 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
8314 (*floathi<X87MODEF>2_i387_with_temp): Remove.
8315 (floathi splitters): Remove.
8316 (float<SWI48x>xf2): New pattern.
8317 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
8318 code that tried to handle DImode for 32-bit, but which was excluded
8319 by the pattern's condition. Drop allocation of stack temporary.
8320 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
8321 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
8322 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
8323 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
8324 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
8325 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
8326 (*float<SWI48><MODEF>2_sse_interunit): Remove.
8327 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
8328 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
8329 (*float<SWI48x><X87MODEF>2_i387): Remove.
8330 (all float _with_temp splitters): Remove.
8331 (*float<SWI48x><MODEF>2_i387): New pattern.
8332 (*float<SWI48><MODEF>2_sse): New pattern.
8333 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
8334 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
8336 2014-03-14 Jakub Jelinek <jakub@redhat.com>
8337 Marek Polacek <polacek@redhat.com>
8340 * common.opt (dump_base_name_prefixed): New Variable.
8341 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
8342 if x_dump_base_name_prefixed is already set, set it at the end.
8344 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
8346 PR rtl-optimization/60508
8347 * lra-constraints.c (get_reload_reg): Add new parameter
8349 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
8350 Pass the new parameter values.
8352 2014-03-14 Richard Biener <rguenther@suse.de>
8354 * common.opt: Revert unintented changes from r205065.
8357 2014-03-14 Richard Biener <rguenther@suse.de>
8360 * cfghooks.c (split_block): Properly adjust all loops the
8361 block was a latch of.
8363 2014-03-14 Martin Jambor <mjambor@suse.cz>
8366 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
8369 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
8372 * config/avr/avr.c (avr_set_current_function): Pass function name
8373 through default_strip_name_encoding before sanity checking instead
8374 of skipping the first char of the assembler name.
8376 2014-03-13 Richard Henderson <rth@redhat.com>
8379 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
8380 (ix86_force_to_memory, ix86_free_from_memory): Remove.
8381 * config/i386/i386-protos.h: Likewise.
8382 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
8383 in the expander instead of a splitter.
8384 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
8385 any possibility of requiring a memory.
8386 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
8387 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
8388 (fp branch splitters): Update for ix86_split_fp_branch.
8389 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
8390 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
8391 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
8392 (*fop_<MODEF>_2_i387): Remove f/r alternative.
8393 (*fop_<MODEF>_3_i387): Likewise.
8394 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
8395 (splitters for the fop_* register patterns): Remove.
8396 (fscalexf4_i387): Rename from *fscalexf4_i387.
8397 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
8399 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8401 PR tree-optimization/59779
8402 * tree-dfa.c (get_ref_base_and_extent): Use double_int
8403 type for bitsize and maxsize instead of HOST_WIDE_INT.
8405 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
8407 PR rtl-optimization/57320
8408 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
8409 the CFG after thread_prologue_and_epilogue_insns.
8411 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
8413 PR rtl-optimization/57189
8414 * lra-constraints.c (process_alt_operands): Disfavor spilling
8417 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
8419 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
8421 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8423 PR tree-optimization/59025
8425 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
8426 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
8428 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
8431 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
8432 calls of avr_out_plus_1.
8434 2014-03-13 Bin Cheng <bin.cheng@arm.com>
8436 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
8437 BB's single pred and update the father loop's latch info later.
8439 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8441 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
8445 (VEC_base): Likewise.
8446 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
8447 registers, we need to swap double words in little endian mode.
8449 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
8450 to be a container mode for 128-bit integer operations added in ISA
8451 2.07. Unlike TImode and PTImode, the preferred register set is
8452 the Altivec/VMX registers for the 128-bit operations.
8454 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
8456 (rs6000_split_128bit_ok_p): Likewise.
8458 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
8459 macros for creating ISA 2.07 normal and overloaded builtin
8460 functions with 3 arguments.
8461 (BU_P8V_OVERLOAD_3): Likewise.
8462 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
8463 for use as overloaded functions.
8464 (VPERM_1TI_UNS): Likewise.
8465 (VSEL_1TI): Likewise.
8466 (VSEL_1TI_UNS): Likewise.
8467 (ST_INTERNAL_1ti): Likewise.
8468 (LD_INTERNAL_1ti): Likewise.
8469 (XXSEL_1TI): Likewise.
8470 (XXSEL_1TI_UNS): Likewise.
8471 (VPERM_1TI): Likewise.
8472 (VPERM_1TI_UNS): Likewise.
8473 (XXPERMDI_1TI): Likewise.
8474 (SET_1TI): Likewise.
8475 (LXVD2X_V1TI): Likewise.
8476 (STXVD2X_V1TI): Likewise.
8477 (VEC_INIT_V1TI): Likewise.
8478 (VEC_SET_V1TI): Likewise.
8479 (VEC_EXT_V1TI): Likewise.
8480 (EQV_V1TI): Likewise.
8481 (NAND_V1TI): Likewise.
8482 (ORC_V1TI): Likewise.
8483 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
8484 added in ISA 2.07. Add both normal 'altivec' builtins, and the
8486 (VADDUQM): Likewise.
8487 (VSUBCUQ): Likewise.
8488 (VADDEUQM): Likewise.
8489 (VADDECUQ): Likewise.
8490 (VSUBEUQM): Likewise.
8491 (VSUBECUQ): Likewise.
8493 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
8494 __int128_t and __uint128_t types.
8495 (__uint128_type): Likewise.
8496 (altivec_categorize_keyword): Add support for vector __int128_t,
8497 vector __uint128_t, vector __int128, and vector unsigned __int128
8498 as a container type for TImode operations that need to be done in
8499 VSX/Altivec registers.
8500 (rs6000_macro_to_expand): Likewise.
8501 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
8502 to support 128-bit integer instructions vaddcuq, vadduqm,
8503 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
8504 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
8506 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
8507 for V1TImode, and set up preferences to use VSX/Altivec registers.
8508 Setup VSX reload handlers.
8509 (rs6000_debug_reg_global): Likewise.
8510 (rs6000_init_hard_regno_mode_ok): Likewise.
8511 (rs6000_preferred_simd_mode): Likewise.
8512 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
8513 (easy_altivec_constant): Likewise.
8514 (output_vec_const_move): Likewise.
8515 (rs6000_expand_vector_set): Convert V1TImode set and extract to
8517 (rs6000_expand_vector_extract): Likewise.
8518 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
8520 (rs6000_const_vec): Add support for V1TImode.
8521 (rs6000_emit_le_vsx_load): Swap double words when loading or
8522 storing TImode/V1TImode.
8523 (rs6000_emit_le_vsx_store): Likewise.
8524 (rs6000_emit_le_vsx_move): Likewise.
8525 (rs6000_emit_move): Add support for V1TImode.
8526 (altivec_expand_ld_builtin): Likewise.
8527 (altivec_expand_st_builtin): Likewise.
8528 (altivec_expand_vec_init_builtin): Likewise.
8529 (altivec_expand_builtin): Likewise.
8530 (rs6000_init_builtins): Add support for V1TImode type. Add
8531 support for ISA 2.07 128-bit integer builtins. Define type names
8532 for the VSX/Altivec vector types.
8533 (altivec_init_builtins): Add support for overloaded vector
8534 functions with V1TImode type.
8535 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
8536 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
8538 (rs6000_split_128bit_ok_p): Likewise.
8539 (rs6000_handle_altivec_attribute): Create V1TImode from vector
8540 __int128_t and vector __uint128_t.
8542 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
8543 and mode attributes.
8550 (VS_scalar): Likewise.
8551 (VS_double): Likewise.
8552 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
8554 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
8555 we support the ISA 2.07 128-bit integer arithmetic instructions.
8556 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
8557 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
8558 and TImode types for use with the builtin functions.
8559 (V1TI_type_node): Likewise.
8560 (unsigned_V1TI_type_node): Likewise.
8561 (intTI_type_internal_node): Likewise.
8562 (uintTI_type_internal_node): Likewise.
8564 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
8565 128-bit builtin functions.
8566 (UNSPEC_VADDEUQM): Likewise.
8567 (UNSPEC_VADDECUQ): Likewise.
8568 (UNSPEC_VSUBCUQ): Likewise.
8569 (UNSPEC_VSUBEUQM): Likewise.
8570 (UNSPEC_VSUBECUQ): Likewise.
8571 (VM): Add V1TImode to vector mode iterators.
8573 (VI_unit): Likewise.
8574 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
8575 (altivec_vaddcuq): Likewise.
8576 (altivec_vsubuqm): Likewise.
8577 (altivec_vsubcuq): Likewise.
8578 (altivec_vaddeuqm): Likewise.
8579 (altivec_vaddecuq): Likewise.
8580 (altivec_vsubeuqm): Likewise.
8581 (altivec_vsubecuq): Likewise.
8583 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
8585 (BOOL_128): Likewise.
8586 (BOOL_REGS_OUTPUT): Likewise.
8587 (BOOL_REGS_OP1): Likewise.
8588 (BOOL_REGS_OP2): Likewise.
8589 (BOOL_REGS_UNARY): Likewise.
8590 (BOOL_REGS_AND_CR0): Likewise.
8592 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
8593 128-bit integer builtin support.
8594 (vec_vadduqm): Likewise.
8595 (vec_vaddecuq): Likewise.
8596 (vec_vaddeuqm): Likewise.
8597 (vec_vsubecuq): Likewise.
8598 (vec_vsubeuqm): Likewise.
8599 (vec_vsubcuq): Likewise.
8600 (vec_vsubuqm): Likewise.
8602 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8603 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
8604 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
8605 128-bit integer add/subtract to ISA 2.07.
8607 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
8609 * config/arc/arc.c (arc_predicate_delay_insns):
8610 Fix third argument passed to conditionalize_nonjump.
8612 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
8614 * config/aarch64/aarch64-builtins.c
8615 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
8616 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
8617 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
8618 instead of __builtin_lfloor.
8619 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
8621 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8623 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
8624 (tree_ssa_ifcombine_bb_1): New function.
8625 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
8626 is an empty forwarder block to then_bb or vice versa and then_bb
8627 and else_bb are effectively swapped.
8629 2014-03-12 Christian Bruel <christian.bruel@st.com>
8632 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
8633 REG_CFA_DEF_CFA note.
8634 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
8635 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
8637 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
8639 PR tree-optimization/60454
8640 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
8642 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8644 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
8645 Do not define target_cpu_default2 to generic.
8646 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
8647 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
8648 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
8650 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8651 Marc Glisse <marc.glisse@inria.fr>
8653 PR tree-optimization/60502
8654 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
8655 instead of build_low_bits_mask.
8657 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8660 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
8661 if there are multiple uses, but op doesn't live on E edge.
8662 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
8663 clobber stmts before __builtin_unreachable.
8665 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
8667 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
8668 hard_frame_pointer_rtx.
8669 * cse.c (cse_insn): Remove volatile check.
8670 * cselib.c (cselib_process_insn): Likewise.
8671 * dse.c (scan_insn): Likewise.
8673 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8675 * config/arc/arc.c (conditionalize_nonjump): New function,
8677 (arc_ifcvt): ... this.
8678 (arc_predicate_delay_insns): Use it.
8680 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8682 * config/arc/predicates.md (extend_operand): During/after reload,
8683 allow const_int_operand.
8684 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
8685 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
8686 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
8688 (umulsi3_highpart_i): Likewise.
8690 2014-03-11 Richard Biener <rguenther@suse.de>
8692 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
8693 Add asserts to guard possible wrong-code bugs.
8695 2014-03-11 Richard Biener <rguenther@suse.de>
8697 PR tree-optimization/60429
8698 PR tree-optimization/60485
8699 * tree-ssa-structalias.c (set_union_with_increment): Properly
8700 take into account all fields that overlap the shifted vars.
8701 (do_sd_constraint): Likewise.
8702 (do_ds_constraint): Likewise.
8703 (get_constraint_for_ptr_offset): Likewise.
8705 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
8707 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
8708 (nios2_compute_frame_layout):
8709 Add calculation of cfun->machine->fp_save_offset.
8710 (nios2_expand_prologue): Correct setting of frame pointer register
8712 (nios2_expand_epilogue): Update recovery of stack pointer from
8713 frame pointer accordingly.
8714 (nios2_initial_elimination_offset): Update calculation of offset
8715 for eliminating to HARD_FRAME_POINTER_REGNUM.
8717 2014-03-10 Jakub Jelinek <jakub@redhat.com>
8720 * ipa.c (symtab_remove_unreachable_nodes): Don't call
8721 cgraph_get_create_node on VAR_DECLs.
8723 2014-03-10 Richard Biener <rguenther@suse.de>
8726 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8728 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8730 * config/vms/vms.opt (vms_float_format): New variable.
8732 2014-03-08 Tobias Burnus <burnus@net-b.de>
8734 * doc/invoke.texi (-fcilkplus): Update implementation status.
8736 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8737 Richard Biener <rguenther@suse.de>
8739 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8740 consistently accross all TUs.
8741 (run_gcc): Enable -fshort-double automatically at link at link-time
8742 and disallow override.
8744 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8747 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8748 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8749 if they can't be used.
8751 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8753 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8754 for Solaris 11/x86 ld.
8755 * configure: Regenerate.
8757 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8759 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8760 (LIB_TLS_SPEC): Save as ld_tls_libs.
8761 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8762 (HAVE_AS_IX86_TLSLDM): New test.
8763 * configure, config.in: Regenerate.
8764 * config/i386/i386.c (legitimize_tls_address): Fall back to
8765 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8766 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8767 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8768 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8770 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8772 * common.opt (fira-loop-pressure): Mark as optimization.
8774 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8776 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8777 an OpenMP mappable type.
8779 2014-03-06 Matthias Klose <doko@ubuntu.com>
8781 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
8782 MULTILIB_OSDIRNAMES is not defined.
8784 2014-03-06 Jakub Jelinek <jakub@redhat.com>
8785 Meador Inge <meadori@codesourcery.com>
8788 * config/arm/arm.c (arm_tls_symbol_p): Remove.
8789 (arm_legitimize_address): Call legitimize_tls_address for any
8790 arm_tls_referenced_p expression, handle constant addend. Call it
8791 before testing for !TARGET_ARM.
8792 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
8794 2014-03-06 Richard Biener <rguenther@suse.de>
8800 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8802 * tree-streamer.c (record_common_node): Assert we don't record
8803 nodes with type double.
8804 (preload_common_node): Skip type double, complex double and double
8805 pointer since it is now frontend dependent due to fshort-double option.
8807 2014-03-06 Richard Biener <rguenther@suse.de>
8809 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
8810 or -fno-lto is specified and the linker has full plugin support.
8811 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
8812 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
8813 * lto-wrapper.c (merge_and_complain): Merge compile-time
8814 optimization levels.
8815 (run_gcc): And pass it through to the link options.
8817 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
8821 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8823 * cselib.c (remove_useless_values): Skip to avoid quadratic
8824 behavior if the condition moved from...
8825 (cselib_process_insn): ... here holds.
8827 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8830 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
8831 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
8834 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
8835 (TM_H): Add x86-tune.def.
8837 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8839 * config/aarch64/aarch64.c (generic_tunings):
8840 Use cortexa57_extra_costs.
8842 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8845 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
8846 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
8847 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
8850 2014-03-04 Heiher <r@hev.cc>
8852 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
8853 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
8855 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
8857 * config/i386/predicates.md (const2356_operand): Change to ...
8858 (const2367_operand): ... this.
8859 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
8861 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8862 (*avx512pf_scatterpf<mode>sf): Ditto.
8863 (avx512pf_scatterpf<mode>df): Ditto.
8864 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8865 (*avx512pf_scatterpf<mode>df): Ditto.
8866 * config/i386/i386.c (ix86_expand_builtin): Update
8867 incorrect hint operand error message.
8869 2014-03-04 Richard Biener <rguenther@suse.de>
8871 * lto-section-in.c (lto_get_section_data): Fix const cast.
8873 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8875 * tree-streamer.c (record_common_node): Assert we don't record
8876 nodes with type double.
8877 (preload_common_node): Skip type double, complex double and double
8878 pointer since it is now frontend dependent due to fshort-double option.
8880 2014-03-04 Richard Biener <rguenther@suse.de>
8883 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
8884 (lto_input_toplevel_asms): Likewise.
8885 * lto-section-in.c (lto_get_section_data): Instead do it here
8888 2014-03-04 Richard Biener <rguenther@suse.de>
8890 PR tree-optimization/60382
8891 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
8892 dead PHIs a reduction.
8894 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
8896 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
8898 (_mm_prefetch): Move out of GCC target("sse") pragma.
8899 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
8900 GCC target("prfchw") pragma.
8901 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
8903 * config/i386/i386.c (ix86_option_override_internal): Enable
8904 -mprfchw with -mprefetchwt1.
8906 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8908 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
8911 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8913 * opts.h (CL_PCH_IGNORE): Define.
8914 * targhooks.c (option_affects_pch_p):
8915 Return false for options that have CL_PCH_IGNORE set.
8916 * opt-functions.awk: Process PchIgnore.
8917 * doc/options.texi: Document PchIgnore.
8919 * config/arc/arc.opt (misize): Add PchIgnore property.
8921 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8923 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
8924 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
8925 constraint on constants to permit them being loaded into
8926 GENERAL_REGS or BASE_REGS.
8928 2014-03-03 Nick Clifton <nickc@redhat.com>
8930 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
8931 anti-cacnonical alternatives.
8932 (negandhi3_real): New pattern.
8933 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
8935 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8937 * config/avr/avr-mcus.def: Remove atxmega16x1.
8938 * config/avr/avr-tables.opt: Regenerate.
8939 * config/avr/t-multilib: Regenerate.
8940 * doc/avr-mmcu.texi: Regenerate.
8942 2014-03-03 Tobias Grosser <tobias@grosser.es>
8943 Mircea Namolaru <mircea.namolaru@inria.fr>
8945 PR tree-optimization/58028
8946 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
8949 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8951 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
8952 not handled by recognizers.
8954 2014-03-03 Jakub Jelinek <jakub@redhat.com>
8957 * function.c (expand_function_end): Don't emit
8958 clobber_return_register sequence if clobber_after is a BARRIER.
8959 * cfgexpand.c (construct_exit_block): Append instructions before
8960 return_label to prev_bb.
8962 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8964 * config/rs6000/constraints.md: Document reserved use of "wc".
8966 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8969 * ipa.c (function_and_variable_visibility): When dissolving comdat
8970 group, also set all symbols to local.
8972 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8977 2013-12-14 Jan Hubicka <jh@suse.cz>
8979 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
8981 2014-03-02 Jon Beniston <jon@beniston.com>
8987 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
8988 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
8989 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
8990 (simple_return, *simple_return): New patterns
8991 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
8992 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
8994 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
8996 * dwarf2out.c (gen_subprogram_die): Tidy.
8998 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
9001 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
9002 (*mov_t_msb_neg_negc): ... this new insn.
9004 2014-02-28 Jason Merrill <jason@redhat.com>
9007 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
9010 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
9013 * dwarf2out.c (decltype_auto_die): New static.
9014 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
9015 (gen_type_die_with_usage): Handle 'decltype(auto)'.
9016 (is_cxx_auto): Likewise.
9018 2014-02-28 Ian Bolton <ian.bolton@arm.com>
9020 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
9021 we are not using general regs only.
9023 2014-02-28 Richard Biener <rguenther@suse.de>
9026 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
9027 previous fix and only allow to remove trivial pre-headers
9028 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
9029 (remove_forwarder_block): Properly update the latch of a loop.
9031 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9034 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
9035 (cselib_preserved_hash_table): New.
9036 (preserve_constants_and_equivs): Move preserved vals to it.
9037 (cselib_find_slot): Look it up first.
9038 (cselib_init): Initialize it.
9039 (cselib_finish): Release it.
9040 (dump_cselib_table): Dump it.
9042 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9045 * cselib.c (remove_useless_values): Skip to avoid quadratic
9046 behavior if the condition moved from...
9047 (cselib_process_insn): ... here holds.
9049 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9052 * var-tracking.c (vt_initialize): Apply the same condition to
9053 preserve the CFA base value.
9055 2014-02-28 Joey Ye <joey.ye@arm.com>
9058 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
9059 if reload in progress or completed.
9061 2014-02-28 Tobias Burnus <burnus@net-b.de>
9064 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
9067 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
9069 * doc/tm.texi.in (Condition Code Status): Update documention for
9070 relative locations of cc0-setter and cc0-user.
9072 2014-02-27 Jeff Law <law@redhat.com>
9074 PR rtl-optimization/52714
9075 * combine.c (try_combine): When splitting an unrecognized PARALLEL
9076 into two independent simple sets, if I3 is a jump, ensure the
9077 pattern we place into I3 is a (set (pc) ...).
9079 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
9080 Jeff Law <law@redhat.com>
9082 PR rtl-optimization/49847
9083 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
9084 are in different blocks.
9085 * doc/tm.texi (Condition Code Status): Update documention for
9086 relative locations of cc0-setter and cc0-user.
9088 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
9091 * lra.c (lra_emit_add): Check SUBREG too.
9093 2014-02-27 Andreas Schwab <schwab@suse.de>
9095 * config/m68k/m68k.c (m68k_option_override): Disable
9096 -flive-range-shrinkage for classic m68k.
9097 (m68k_override_options_after_change): Likewise.
9099 2014-02-27 Marek Polacek <polacek@redhat.com>
9102 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
9103 -Wmaybe-uninitialized.
9105 2014-02-27 Alan Modra <amodra@gmail.com>
9108 * reload1.c (emit_input_reload_insns): When reload_override_in,
9109 set old to rl->in_reg when rl->in_reg is a subreg.
9111 2014-02-26 Richard Biener <rguenther@suse.de>
9114 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
9116 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9118 * common/config/i386/predicates.md (const1256_operand): Remove.
9119 (const2356_operand): New.
9120 (const_1_to_2_operand): Remove.
9121 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9122 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9123 (*avx512pf_gatherpf<mode>sf): Ditto.
9124 (avx512pf_gatherpf<mode>df): Ditto.
9125 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9126 (*avx512pf_gatherpf<mode>df): Ditto.
9127 (avx512pf_scatterpf<mode>sf): Ditto.
9128 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9129 (*avx512pf_scatterpf<mode>sf): Ditto.
9130 (avx512pf_scatterpf<mode>df): Ditto.
9131 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9132 (*avx512pf_scatterpf<mode>df): Ditto.
9133 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
9135 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
9137 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
9138 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
9139 (_mm512_mask_testn_epi64_mask): Move to ...
9140 * config/i386/avx512cdintrin.h: Here.
9141 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
9142 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
9143 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
9144 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
9145 TARGET_AVX512F from TARGET_AVX512CD.
9147 2014-02-26 Richard Biener <rguenther@suse.de>
9150 * ipa.c (walk_polymorphic_call_targets): Properly guard
9151 call to inline_update_overall_summary.
9153 2014-02-26 Bin Cheng <bin.cheng@arm.com>
9156 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
9157 and latches only if requested. Fix latch if it is removed.
9158 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
9159 LOOPS_HAVE_PREHEADERS.
9161 2014-02-25 Andrew Pinski <apinski@cavium.com>
9163 * builtins.c (expand_builtin_thread_pointer): Create a new target
9164 when the target is NULL.
9166 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
9168 PR rtl-optimization/60317
9169 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9170 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9171 * lra-assigns.c: Include params.h.
9172 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
9173 other reload pseudos considerations.
9175 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9177 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
9178 to use canonical form for nor<mode>3.
9180 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9183 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
9186 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9188 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
9189 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
9190 (ix86_handle_option): Handle OPT_mprefetchwt1.
9191 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
9192 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9194 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9195 OPTION_MASK_ISA_PREFETCHWT1.
9196 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
9197 (PTA_PREFETCHWT1): New.
9198 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
9199 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
9200 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
9201 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
9202 (*prefetch_avx512pf_<mode>_: Change into ...
9203 (*prefetch_prefetchwt1_<mode>: This.
9204 * config/i386/i386.opt (mprefetchwt1): New.
9205 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
9206 (_mm_prefetch): Handle intent to write.
9207 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
9209 2014-02-25 Richard Biener <rguenther@suse.de>
9212 * emit-rtl.c (mem_attrs_htab): Remove.
9213 (mem_attrs_htab_hash): Likewise.
9214 (mem_attrs_htab_eq): Likewise.
9215 (set_mem_attrs): Always allocate new mem-attrs when something changed.
9216 (init_emit_once): Do not allocate mem_attrs_htab.
9218 2014-02-25 Richard Biener <rguenther@suse.de>
9221 * lto-opts.c (lto_write_options): Output non-explicit conservative
9222 -fwrapv, -fno-trapv and -fno-strict-overflow.
9223 * lto-wrapper.c (merge_and_complain): Handle merging those options.
9224 (run_gcc): And pass them through.
9226 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9228 * sel-sched.c (calculate_new_fences): New parameter ptime.
9229 Calculate it as a maximum over all fence cycles.
9230 (sel_sched_region_2): Adjust the call to calculate_new_fences.
9231 Print the final schedule timing when sched_verbose.
9233 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9235 PR rtl-optimization/60292
9236 * sel-sched.c (fill_vec_av_set): Do not reset target availability
9237 bit fot the fence instruction.
9239 2014-02-24 Alangi Derick <alangiderick@gmail.com>
9241 * calls.h: Fix typo in comment.
9243 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
9245 * config/pa/pa.c (pa_output_move_double): Don't valididate when
9246 adjusting offsetable addresses.
9248 2014-02-24 Guozhi Wei <carrot@google.com>
9250 * sparseset.h (sparseset_pop): Fix the wrong index.
9252 2014-02-24 Walter Lee <walt@tilera.com>
9254 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
9255 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
9257 * common/config/tilegx/tilegx-common.c
9258 (TARGET_DEFAULT_TARGET_FLAGS): Define.
9259 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
9261 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
9262 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
9263 (tilegx_gimplify_va_arg_expr): Handle big endian.
9264 (tilegx_expand_unaligned_load): Ditto.
9265 (tilegx_expand_unaligned_store): Ditto.
9266 (TARGET_RETURN_IN_MSB): New.
9267 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
9268 (TARGET_ENDIAN_DEFAULT): New.
9269 (TARGET_BIG_ENDIAN): Handle big endian.
9270 (BYTES_BIG_ENDIAN): Ditto.
9271 (WORDS_BIG_ENDIAN): Ditto.
9272 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
9275 * config/tilegx/tilegx.md (extv): Handle big endian.
9277 (insn_st<n>): Ditto.
9278 (insn_st<n>_add<bitsuffix>): Ditto.
9279 (insn_stnt<n>): Ditto.
9280 (insn_stnt<n>_add<bitsuffix>):Ditto.
9281 (vec_interleave_highv8qi): Handle big endian.
9282 (vec_interleave_highv8qi_be): New.
9283 (vec_interleave_highv8qi_le): New.
9284 (insn_v1int_h): Handle big endian.
9285 (vec_interleave_lowv8qi): Handle big endian.
9286 (vec_interleave_lowv8qi_be): New.
9287 (vec_interleave_lowv8qi_le): New.
9288 (insn_v1int_l): Handle big endian.
9289 (vec_interleave_highv4hi): Handle big endian.
9290 (vec_interleave_highv4hi_be): New.
9291 (vec_interleave_highv4hi_le): New.
9292 (insn_v2int_h): Handle big endian.
9293 (vec_interleave_lowv4hi): Handle big endian.
9294 (vec_interleave_lowv4hi_be): New.
9295 (vec_interleave_lowv4hi_le): New.
9296 (insn_v2int_l): Handle big endian.
9297 (vec_interleave_highv2si): Handle big endian.
9298 (vec_interleave_highv2si_be): New.
9299 (vec_interleave_highv2si_le): New.
9300 (insn_v4int_h): Handle big endian.
9301 (vec_interleave_lowv2si): Handle big endian.
9302 (vec_interleave_lowv2si_be): New.
9303 (vec_interleave_lowv2si_le): New.
9304 (insn_v4int_l): Handle big endian.
9305 * config/tilegx/tilegx.opt (mbig-endian): New option.
9306 (mlittle-endian): New option.
9307 * doc/install.texi: Document tilegxbe-linux.
9308 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
9310 2014-02-24 Martin Jambor <mjambor@suse.cz>
9313 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
9314 there are no parameter descriptors.
9316 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
9318 PR rtl-optimization/60268
9319 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
9320 initialization to ...
9321 (sched_rgn_init): ... here.
9322 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
9324 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9326 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
9329 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9331 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
9334 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9336 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
9337 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
9339 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9341 * config/microblaze/predicates.md: Add cmp_op predicate.
9342 * config/microblaze/microblaze.md: Add branch_compare instruction
9343 which uses cmp_op predicate and emits cmp insn before branch.
9344 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
9345 to microblaze_expand_conditional_branch and consolidate logic.
9346 (microblaze_expand_conditional_branch): emit branch_compare
9347 insn instead of handling cmp op separate from branch insn.
9349 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9351 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
9354 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9356 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
9357 define_insn with define_expand and new define_insn
9358 *altivec_lve<VI_char>x_internal.
9359 (altivec_stve<VI_char>x): Replace define_insn with define_expand
9360 and new define_insn *altivec_stve<VI_char>x_internal.
9361 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
9363 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
9365 (altivec_expand_stvex_be): New function.
9367 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
9369 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
9370 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
9371 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
9372 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
9374 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
9377 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
9378 instead of emit_move_insn.
9380 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9382 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
9384 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
9385 gen_altivec_vsumsws.
9387 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9389 * config/rs6000/altivec.md (altivec_lvxl): Rename as
9390 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
9391 (altivec_lvxl_<mode>): New define_expand incorporating
9392 -maltivec=be semantics where needed.
9393 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
9394 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
9395 semantics where needed.
9396 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
9397 (altivec_stvx_<mode>): New define_expand incorporating
9398 -maltivec=be semantics where needed.
9399 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
9400 VM2 iterator instead of V4SI.
9401 (altivec_stvxl_<mode>): New define_expand incorporating
9402 -maltivec=be semantics where needed.
9403 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
9404 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
9405 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
9406 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
9407 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
9408 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
9409 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
9410 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
9411 ALTIVEC_BUILTIN_STVXL.
9412 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
9413 (altivec_expand_stvx_be): Likewise.
9414 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
9415 (altivec_expand_lvx_be): Likewise.
9416 (altivec_expand_stvx_be): Likewise.
9417 (altivec_expand_builtin): Add cases for
9418 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
9419 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
9420 (altivec_init_builtins): Add definitions for
9421 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
9422 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
9424 2014-02-21 Catherine Moore <clm@codesourcery.com>
9426 * doc/invoke.texi (mvirt, mno-virt): Document.
9427 * config/mips/mips.opt (mvirt): New option.
9428 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
9430 2014-02-21 Richard Biener <rguenther@suse.de>
9432 PR tree-optimization/60276
9433 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
9434 (STMT_VINFO_MIN_NEG_DIST): New macro.
9435 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
9436 STMT_VINFO_MIN_NEG_DIST.
9437 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
9438 made for negative dependence distances still hold.
9440 2014-02-21 Richard Biener <rguenther@suse.de>
9443 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
9444 DECL_INITIAL for globals not in the current function context.
9446 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9448 PR tree-optimization/56490
9449 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
9450 * tree-ssa-uninit.c: Include params.h.
9451 (compute_control_dep_chain): Add num_calls argument, return false
9452 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
9453 num_calls to recursive call.
9454 (find_predicates): Change dep_chain into normal array,
9455 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
9456 variable and adjust compute_control_dep_chain caller.
9457 (find_def_preds): Likewise.
9459 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
9461 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
9462 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
9464 2014-02-21 Nick Clifton <nickc@redhat.com>
9466 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
9467 (pushhi1): Likewise.
9468 (popqi1): Add mode to pre_dec.
9471 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9473 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
9474 mode for mask of V8SFmode permutation.
9476 2014-02-20 Richard Henderson <rth@redhat.com>
9479 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
9480 a new pseudo for OLDVAL.
9482 2014-02-20 Jakub Jelinek <jakub@redhat.com>
9485 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
9486 gen_reg_rtx if d->testing_p.
9487 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
9488 if d->testing_p and we will certainly return true.
9489 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
9492 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
9494 * emit-rtl.c (gen_reg_rtx): Assert that
9495 crtl->emit.regno_pointer_align_length is non-zero.
9497 2014-02-20 Richard Henderson <rth@redhat.com>
9500 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
9501 on failure the store back into EXPECT.
9503 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
9504 Sandra Loosemore <sandra@codesourcery.com>
9506 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
9507 * config/nios2/nios2.c (nios2_function_profiler): Add
9508 -fPIC (flag_pic == 2) support.
9509 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
9510 (nios2_large_offset_p): New function.
9511 (nios2_unspec_reloc_p): Move up position, update to use
9512 nios2_large_offset_p.
9513 (nios2_unspec_address): Remove function.
9514 (nios2_unspec_offset): New function.
9515 (nios2_large_got_address): New function.
9516 (nios2_got_address): Add large offset support.
9517 (nios2_legitimize_tls_address): Update usage of removed and new
9519 (nios2_symbol_binds_local_p): New function.
9520 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
9521 (nios2_legitimize_address): Update to use nios2_large_offset_p.
9522 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
9523 (nios2_print_operand): Merge H/L processing, add hiadj/lo
9524 processing for (const (unspec ...)).
9525 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
9527 2014-02-20 Richard Biener <rguenther@suse.de>
9529 * tree-cfg.c (replace_uses_by): Mark altered BBs before
9530 doing the substitution.
9531 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
9533 2014-02-20 Martin Jambor <mjambor@suse.cz>
9536 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
9537 info when checking whether lattices are bottom.
9539 2014-02-20 Richard Biener <rguenther@suse.de>
9542 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
9545 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
9548 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
9549 parameter specifying the scaling.
9550 (inline_call): Update.
9551 (want_inline_recursively): Guard division by zero.
9552 (recursive_inlining): Update.
9553 * ipa-inline.h (clone_inlined_nodes): Update.
9555 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9558 * config/i386/i386.c (classify_argument): Pass structures of size
9559 64 bytes or less in register.
9561 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9562 Kirill Yukhin <kirill.yukhin@intel.com>
9564 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
9565 (_mm_rcp28_round_ss): Ditto.
9566 (_mm_rsqrt28_round_sd): Ditto.
9567 (_mm_rsqrt28_round_ss): Ditto.
9568 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
9569 (_mm_rcp14_round_ss): Ditto.
9570 (_mm_rsqrt14_round_sd): Ditto.
9571 (_mm_rsqrt14_round_ss): Ditto.
9572 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
9573 the first input operand, get rid of match_dup.
9574 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
9576 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9578 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
9579 operand as the first input operand, set type attribute.
9580 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9582 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
9583 operand as the first input operand, set type attribute.
9585 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9587 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
9590 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
9593 * config/i386/i386.c (construct_container): Remove TFmode check
9594 for X86_64_INTEGER_CLASS.
9596 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
9599 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
9600 only when -Wpsabi is enabled.
9602 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
9605 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
9606 passing arrays in registers are the same as for structs, so remove the
9607 special case for them.
9609 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
9611 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
9612 destination type, extract only the valid bits if the source type is not
9613 integral and has a different mode.
9615 2014-02-19 Richard Biener <rguenther@suse.de>
9618 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
9621 2014-02-19 Richard Biener <rguenther@suse.de>
9624 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
9625 (ipa_modify_call_arguments): Emit an argument load explicitely and
9626 preserve virtual SSA form there and for the replacement call.
9627 Do not update SSA form nor free dominance info.
9629 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9631 * ipa.c (function_and_variable_visibility): Also clear WEAK
9632 flag when disolving COMDAT_GROUP.
9634 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9636 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
9637 * ipa-prop.c (ipa_set_jf_known_type): Return early when
9639 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
9640 do more sanity checks.
9641 (detect_type_change): Return true when giving up early.
9642 (compute_complex_assign_jump_func): Fix type parameter of
9643 ipa_set_ancestor_jf.
9644 (compute_complex_ancestor_jump_func): Likewise.
9645 (update_jump_functions_after_inlining): Fix updating of
9647 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
9649 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9651 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
9652 inline clones when edge disappears.
9654 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9657 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
9658 Split 64-bit moves into 2 patterns. Do not allow the use of
9659 direct move for TDmode in little endian, since the decimal value
9660 has little endian bytes within a word, but the 64-bit pieces are
9661 ordered in a big endian fashion, and normal subreg's of TDmode are
9663 (mov<mode>_64bit_dm): Likewise.
9664 (movtd_64bit_nodm): Likewise.
9666 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9668 PR tree-optimization/60174
9669 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
9670 statement of an SSA_NAME that occurs in an abnormal PHI node.
9672 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9675 * final.c (SEEN_BB): Remove.
9676 (SEEN_NOTE, SEEN_EMITTED): Renumber.
9677 (final_scan_insn): Don't force_source_line on second
9678 NOTE_INSN_BASIC_BLOCK.
9680 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
9683 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
9684 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
9685 (type_natural_mode): Warn ABI change when %zmm register is not
9686 available for AVX512F vector value passing.
9688 2014-02-18 Kai Tietz <ktietz@redhat.com>
9691 * config/i386/i386.c (ix86_expand_prologue): Use value in
9692 rax register as displacement when restoring %r10 or %rax.
9693 Fix wrong offset when restoring both registers.
9695 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9697 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
9698 assertion with conditional return.
9700 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9701 Uros Bizjak <ubizjak@gmail.com>
9704 * config/i386/driver-i386.c (host_detect_local_cpu): If
9705 YMM state is not saved by the OS, also clear has_f16c. Move
9706 CPUID 0x80000001 handling before YMM state saving checking.
9708 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
9710 PR rtl-optimization/58960
9711 * haifa-sched.c (alloc_global_sched_pressure_data): New,
9712 factored out from ...
9713 (sched_init): ... here.
9714 (free_global_sched_pressure_data): New, factored out from ...
9715 (sched_finish): ... here.
9716 * sched-int.h (free_global_sched_pressure_data): Declare.
9717 * sched-rgn.c (nr_regions_initial): New static global.
9718 (haifa_find_rgns): Initialize it.
9719 (schedule_region): Disable sched-pressure for the newly
9722 2014-02-17 Richard Biener <rguenther@suse.de>
9724 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9725 release SSA defs of pattern stmts.
9727 2014-02-17 Richard Biener <rguenther@suse.de>
9729 * tree-inline.c (expand_call_inline): Release the virtual
9730 operand defined by the call we are about to inline.
9732 2014-02-17 Richard Biener <rguenther@suse.de>
9734 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9736 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9737 Ilya Tocar <ilya.tocar@intel.com>
9739 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9740 arguments order in builtin.
9741 (_mm512_permutexvar_epi64): Ditto.
9742 (_mm512_mask_permutexvar_epi64): Ditto
9743 (_mm512_maskz_permutexvar_epi32): Ditto
9744 (_mm512_permutexvar_epi32): Ditto
9745 (_mm512_mask_permutexvar_epi32): Ditto
9747 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9749 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9750 (p8_vmrgow): Likewise.
9752 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9754 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9757 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9760 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9761 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9762 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9763 using direct move instructions on ISA 2.07. Also adjust
9764 instruction length for 64-bit.
9765 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9766 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9768 2014-02-15 Alan Modra <amodra@gmail.com>
9772 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9773 find_replacement on parts of insn rtl that might be reloaded.
9775 2014-02-15 Richard Biener <rguenther@suse.de>
9777 PR tree-optimization/60183
9778 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9779 (tree_ssa_phiprop): Calculate and free post-dominators.
9781 2014-02-14 Jeff Law <law@redhat.com>
9783 PR rtl-optimization/60131
9784 * ree.c (get_extended_src_reg): New function.
9785 (combine_reaching_defs): Use it rather than assuming location of REG.
9786 (find_and_remove_re): Verify first operand of extension is
9787 a REG before adding the insns to the copy list.
9789 2014-02-14 Roland McGrath <mcgrathr@google.com>
9791 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
9792 * configure: Regenerated.
9793 * config.in: Regenerated.
9794 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
9795 instead of ASM_SHORT.
9797 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
9798 Richard Earnshaw <rearnsha@arm.com>
9800 PR rtl-optimization/59535
9801 * lra-constraints.c (process_alt_operands): Encourage alternative
9802 when unassigned pseudo class is superset of the alternative class.
9803 (inherit_reload_reg): Don't inherit when optimizing for code size.
9804 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
9805 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
9806 modes not less than 4 for Thumb1.
9808 2014-02-14 Kyle McMartin <kyle@redhat.com>
9811 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
9813 2014-02-14 Richard Biener <rguenther@suse.de>
9815 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
9816 (get_frame_arg): Drop the assert with langhook types_compatible_p.
9817 Do not strip INDIRECT_REFs.
9819 2014-02-14 Richard Biener <rguenther@suse.de>
9822 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
9823 DECL_FUNCTION_SPECIFIC_TARGET.
9824 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
9825 * tree-streamer-out.c (pack_ts_target_option): Remove.
9826 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
9827 (write_ts_function_decl_tree_pointers): Do not stream
9828 DECL_FUNCTION_SPECIFIC_TARGET.
9829 * tree-streamer-in.c (unpack_ts_target_option): Remove.
9830 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
9831 (lto_input_ts_function_decl_tree_pointers): Do not stream
9832 DECL_FUNCTION_SPECIFIC_TARGET.
9834 2014-02-14 Jakub Jelinek <jakub@redhat.com>
9836 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
9837 (get_initial_def_for_induction, vectorizable_induction): Ignore
9838 debug stmts when looking for exit_phi.
9839 (vectorizable_live_operation): Fix up condition.
9841 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9843 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
9844 nreverse() because it changes the content of original tree list.
9846 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9848 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
9849 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
9851 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9853 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
9854 GNU coding standards.
9856 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9859 * dwarf2out.c (gen_subprogram_die): Don't call
9860 add_calling_convention_attribute if subr_die is old_die.
9862 2014-02-13 Sharad Singhai <singhai@google.com>
9864 * doc/optinfo.texi: Fix order of nodes.
9866 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
9868 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
9869 operands[2], not operands[3].
9871 2014-02-13 Richard Biener <rguenther@suse.de>
9874 * doc/install.texi (ISL): Update recommended version to 0.12.2,
9875 mention the possibility of an in-tree build.
9876 (CLooG): Update recommended version to 0.18.1, mention the
9877 possibility of an in-tree build and clarify that the ISL
9878 bundled with CLooG does not work.
9880 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9883 * expr.c (compress_float_constant): If x is a hard register,
9884 extend into a pseudo and then move to x.
9886 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
9888 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
9889 caused by bad second argument to warning_at() with -mhotpatch and
9890 nested functions (e.g. with gfortran).
9892 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
9894 * opts.c (option_name): Remove "enabled by default" rider.
9896 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
9898 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
9900 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
9901 Uros Bizjak <ubizjak@gmail.com>
9904 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
9905 * configure: Regenerated.
9907 2014-02-12 Richard Biener <rguenther@suse.de>
9909 * vec.c (vec_prefix::calculate_allocation): Move as
9910 inline variant to vec.h.
9911 (vec_prefix::calculate_allocation_1): New out-of-line version.
9912 * vec.h (vec_prefix::calculate_allocation_1): Declare.
9913 (vec_prefix::m_has_auto_buf): Rename to ...
9914 (vec_prefix::m_using_auto_storage): ... this.
9915 (vec_prefix::calculate_allocation): Inline the easy cases
9916 and dispatch to calculate_allocation_1 which doesn't need the
9918 (va_heap::reserve): Use gcc_checking_assert.
9919 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
9920 m_using_auto_storage.
9921 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
9923 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
9924 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
9925 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
9927 2014-02-12 Richard Biener <rguenther@suse.de>
9929 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
9930 when we found a dependence.
9932 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
9934 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
9936 (maybe_fold_stmt): ... into this new function.
9937 * omp-low.c (lower_omp): Update comment.
9939 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
9942 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
9945 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
9947 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
9948 identifiers in comments.
9949 (cortexa53_extra_costs): Likewise.
9950 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
9951 (cortexa7_extra_costs): Likewise.
9952 (cortexa12_extra_costs): Likewise.
9953 (cortexa15_extra_costs): Likewise.
9954 (v7m_extra_costs): Likewise.
9956 2014-02-12 Richard Biener <rguenther@suse.de>
9959 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
9960 of posix_memalign being successful.
9961 (lower_stmt): Restrict lowering of posix_memalign to when
9962 -ftree-bit-ccp is enabled.
9964 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9966 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
9968 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
9970 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
9972 PR rtl-optimization/60116
9973 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
9974 other_insn once the combination has been validated.
9976 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
9979 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
9981 * ipa-devirt.c: Include demangle.h
9982 (odr_violation_reported): New static variable.
9983 (add_type_duplicate): Update odr_violations.
9984 (maybe_record_node): Add completep parameter; update it.
9985 (record_target_from_binfo): Add COMPLETEP parameter;
9986 update it as needed.
9987 (possible_polymorphic_call_targets_1): Likewise.
9988 (struct polymorphic_call_target_d): Add nonconstruction_targets;
9989 rename FINAL to COMPLETE.
9990 (record_targets_from_bases): Sanity check we found the binfo;
9991 fix COMPLETEP updating.
9992 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
9993 parameter, fix computing of COMPLETEP.
9994 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
9995 at LTO time do demangling.
9996 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
9997 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
9999 (gimple_get_virt_method_for_binfo): Likewise.
10000 * gimple-fold.h (gimple_get_virt_method_for_binfo,
10001 gimple_get_virt_method_for_vtable): Update prototypes.
10003 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
10006 * genautomata.c (add_presence_absence): Fix typo with
10007 {final_}presence_list.
10009 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10012 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
10013 for VSX/Altivec vectors that land in GPR registers.
10015 2014-02-11 Richard Henderson <rth@redhat.com>
10016 Jakub Jelinek <jakub@redhat.com>
10019 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
10020 around drhs if type conversion to lacc->type is not useless.
10022 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10024 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
10026 (cortex-a57.cortex-a53): Likewise.
10027 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
10029 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10031 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
10034 2014-02-11 Renlin Li <Renlin.Li@arm.com>
10036 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
10037 add_options_for_arm_vfp3.
10039 2014-02-11 Jeff Law <law@redhat.com>
10041 PR middle-end/54041
10042 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
10043 object with an undesirable mode.
10045 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10048 * config/i386/sol2-9.h: New file.
10049 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
10050 *-*-solaris2.9*): Use it.
10052 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10054 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
10055 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
10057 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10059 * config/microblaze/microblaze.c: Extend mcpu version format
10061 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
10063 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
10065 2014-02-10 Richard Henderson <rth@redhat.com>
10068 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
10069 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
10070 ms-abi vs -mno-accumulate-outgoing-args.
10071 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
10072 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
10075 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
10077 PR middle-end/60080
10078 * cfgexpand.c (expand_asm_operands): Attach source location to
10079 ASM_INPUT rtx objects.
10080 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
10082 2014-02-10 Nick Clifton <nickc@redhat.com>
10084 * config/mn10300/mn10300.c (popcount): New function.
10085 (mn10300_expand_prologue): Include saved registers in stack usage
10088 2014-02-10 Jeff Law <law@redhat.com>
10090 PR middle-end/52306
10091 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
10092 when changing the SET_DEST of a prior insn to avoid an input reload.
10094 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10096 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
10097 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
10098 -mcall-openbsd, or -mcall-linux.
10099 (CC1_ENDIAN_BIG_SPEC): Remove.
10100 (CC1_ENDIAN_LITTLE_SPEC): Remove.
10101 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10102 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
10103 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
10104 and %cc1_endian_default.
10105 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10107 2014-02-10 Richard Biener <rguenther@suse.de>
10109 PR tree-optimization/60115
10110 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
10111 MEM_REF handling. Properly verify that the accesses are not
10112 out of the objects bound.
10114 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10116 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
10119 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
10121 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
10122 proper constants and fix formatting.
10123 (possible_polymorphic_call_targets): Fix formatting.
10125 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
10126 Ilya Tocar <ilya.tocar@intel.com>
10128 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
10129 (_mm512_loadu_epi32): Renamed into...
10130 (_mm512_loadu_si512): This.
10131 (_mm512_storeu_epi32): Renamed into...
10132 (_mm512_storeu_si512): This.
10133 (_mm512_maskz_ceil_ps): Removed.
10134 (_mm512_maskz_ceil_pd): Ditto.
10135 (_mm512_maskz_floor_ps): Ditto.
10136 (_mm512_maskz_floor_pd): Ditto.
10137 (_mm512_floor_round_ps): Ditto.
10138 (_mm512_floor_round_pd): Ditto.
10139 (_mm512_ceil_round_ps): Ditto.
10140 (_mm512_ceil_round_pd): Ditto.
10141 (_mm512_mask_floor_round_ps): Ditto.
10142 (_mm512_mask_floor_round_pd): Ditto.
10143 (_mm512_mask_ceil_round_ps): Ditto.
10144 (_mm512_mask_ceil_round_pd): Ditto.
10145 (_mm512_maskz_floor_round_ps): Ditto.
10146 (_mm512_maskz_floor_round_pd): Ditto.
10147 (_mm512_maskz_ceil_round_ps): Ditto.
10148 (_mm512_maskz_ceil_round_pd): Ditto.
10149 (_mm512_expand_pd): Ditto.
10150 (_mm512_expand_ps): Ditto.
10151 * config/i386/i386.c (ix86_builtins): Remove
10152 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
10153 (bdesc_args): Ditto.
10154 * config/i386/predicates.md (const1256_operand): New.
10155 (const_1_to_2_operand): Ditto.
10156 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
10157 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10158 (*avx512pf_gatherpf<mode>sf): Ditto.
10159 (avx512pf_gatherpf<mode>df): Ditto.
10160 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10161 (*avx512pf_gatherpf<mode>df): Ditto.
10162 (avx512pf_scatterpf<mode>sf): Ditto.
10163 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10164 (*avx512pf_scatterpf<mode>sf): Ditto.
10165 (avx512pf_scatterpf<mode>df): Ditto.
10166 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10167 (*avx512pf_scatterpf<mode>df): Ditto.
10168 (avx512f_expand<mode>): Removed.
10169 (<shift_insn><mode>3<mask_name>): Change predicate type.
10171 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10173 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
10174 not at the end of datarefs vector use ordered_remove to avoid
10175 reordering datarefs vector.
10178 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
10179 mark local addressable non-static vars as GOVD_PRIVATE
10180 instead of GOVD_LOCAL.
10181 * omp-low.c (lower_omp_for): Move gimple_bind_vars
10182 and BLOCK_VARS of gimple_bind_block to new_stmt rather
10185 PR middle-end/60092
10186 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
10187 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
10188 assume_aligned or alloc_align attributes.
10189 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
10190 arguments. Handle also assume_aligned and alloc_align attributes.
10191 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
10192 calls to functions with assume_aligned or alloc_align attributes.
10193 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
10195 2014-02-08 Terry Guo <terry.guo@arm.com>
10197 * doc/invoke.texi: Document ARM -march=armv7e-m.
10199 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10201 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
10202 flag on __cilkrts_rethrow builtin.
10205 * ipa-cp.c (determine_versionability): Fail at -O0
10206 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
10207 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
10210 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10213 * tree-inline.c (copy_forbidden): Fail for
10214 __attribute__((optimize (0))) functions.
10216 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10218 * varpool.c: Include pointer-set.h.
10219 (varpool_remove_unreferenced_decls): Variables in other partitions
10220 will not be output; be however careful to not lose information
10221 about partitioning.
10223 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10225 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
10226 lookup in the vtable constructor.
10228 2014-02-07 Jeff Law <law@redhat.com>
10231 * config/m68k/m68k.md (ashldi_extsi): Turn into a
10232 define_insn_and_split.
10234 * ipa-inline.c (inline_small_functions): Fix typos.
10236 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10238 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
10239 (s390_can_use_return_insn): Declare.
10240 * config/s390/s390.h (EPILOGUE_USES): Define.
10241 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
10243 (s390_chunkify_start): Handle return JUMP_LABELs.
10244 (s390_early_mach): Emit a main_pool instruction on the entry edge.
10245 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
10246 (s390_can_use_return_insn): New functions.
10247 (s390_fix_long_loop_prediction): Handle conditional returns.
10248 (TARGET_SET_UP_BY_PROLOGUE): Define.
10249 * config/s390/s390.md (ANY_RETURN): New code iterator.
10250 (*creturn, *csimple_return, return, simple_return): New patterns.
10252 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10254 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
10255 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
10256 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
10257 REG_CFA_RESTORE list when deciding not to restore a register.
10259 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10261 * config/s390/s390.c: Include tree-pass.h and context.h.
10262 (s390_early_mach): New function, split out from...
10263 (s390_emit_prologue): ...here.
10264 (pass_data_s390_early_mach): New pass structure.
10265 (pass_s390_early_mach): New class.
10266 (s390_option_override): Create and register early_mach pass.
10267 Move to end of file.
10269 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10271 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
10272 to match for the exit block.
10274 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10276 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
10277 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
10278 Reject misaligned operands.
10280 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10282 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
10284 2014-02-07 Richard Biener <rguenther@suse.de>
10286 PR middle-end/60092
10287 * gimple-low.c (lower_builtin_posix_memalign): New function.
10288 (lower_stmt): Call it to lower posix_memalign in a way
10289 to make alignment info accessible.
10291 2014-02-07 Jakub Jelinek <jakub@redhat.com>
10294 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
10295 __builtin_setjmp_receiver.
10297 2014-02-07 Richard Biener <rguenther@suse.de>
10299 PR middle-end/60092
10300 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
10301 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
10302 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
10303 Handle BUILT_IN_POSIX_MEMALIGN.
10304 (find_func_clobbers): Likewise.
10305 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
10306 (call_may_clobber_ref_p_1): Likewise.
10308 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10311 * ipa-devirt.c (record_target_from_binfo): Remove overactive
10314 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10317 * lto-cgraph.c (lto_output_node): Use
10318 symtab_get_symbol_partitioning_class.
10319 (lto_output_varpool_node): likewise.
10320 (symtab_get_symbol_partitioning_class): Move here from
10321 lto/lto-partition.c
10322 * cgraph.h (symbol_partitioning_class): Likewise.
10323 (symtab_get_symbol_partitioning_class): Declare.
10325 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10327 * ggc.h (ggc_internal_cleared_alloc): New macro.
10328 * vec.h (vec_safe_copy): Handle memory stats.
10329 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
10330 * target-globals.c (save_target_globals): Likewise.
10332 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10335 * expr.c (emit_move_resolve_push): Export; be bit more selective
10336 on when to clear alias set.
10337 * expr.h (emit_move_resolve_push): Declare.
10338 * function.h (struct function): Add tail_call_marked.
10339 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
10340 * config/i386/i386-protos.h (ix86_expand_push): Remove.
10341 * config/i386/i386.md (TImode move expander): De not call
10343 (FP push expanders): Preserve memory attributes.
10344 * config/i386/sse.md (push<mode>1): Remove.
10345 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
10346 (ix86_expand_push): Remove.
10347 * config/i386/mmx.md (push<mode>1): Remove.
10349 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10351 PR rtl-optimization/60030
10352 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
10353 lopart with paradoxical subreg before shifting it up by hprec.
10355 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10357 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
10358 Remove extra newline at end of file.
10359 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
10360 (arm_issue_rate): Handle cortexa57.
10361 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
10362 (cortex-a57.cortex-a53): Likewise.
10364 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10367 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
10368 don't record in REG_FRAME_RELATED_EXPR registers not set in that
10370 (arm_expand_prologue): Adjust all callers.
10371 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
10372 info, registers also at the lowest numbered registers side. Use
10373 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
10377 * var-tracking.c (adjust_mems): Before adding a SET to
10378 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
10380 2014-02-06 Alan Modra <amodra@gmail.com>
10383 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
10384 change SDmode to DDmode when lra_in_progress.
10386 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10388 PR middle-end/59150
10389 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
10390 free_data_ref on the dr first, and before goto again also set dr
10391 to the next dr. For simd_lane_access, free old datarefs[i] before
10392 overwriting it. For get_vectype_for_scalar_type failure, don't
10393 free_data_ref if simd_lane_access.
10395 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
10398 * tree.h (opts_for_fn): New inline function.
10399 (opt_for_fn): Define.
10400 * config/i386/i386.c (ix86_function_regparm): Use
10401 opt_for_fn (decl, optimize) instead of optimize.
10403 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
10405 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
10406 for SYMBOL_REF in large memory model.
10408 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10410 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
10411 and crypto support.
10412 (cortex-a57): Likewise.
10413 (cortex-a57.cortex-a53): Likewise.
10415 2014-02-06 Yury Gribov <y.gribov@samsung.com>
10416 Kugan Vivekanandarajah <kuganv@linaro.org>
10418 * config/arm/arm.c (arm_vector_alignment_reachable): Check
10420 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
10422 2014-02-06 Richard Biener <rguenther@suse.de>
10424 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
10425 set_loop_copy and initialize_original_copy_tables.
10427 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
10429 * config/aarch64/aarch64-simd.md
10430 (aarch64_ashr_simddi): Change QI to SI.
10432 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10433 Jakub Jelinek <jakub@redhat.com>
10435 PR middle-end/60013
10436 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
10439 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10441 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
10442 CODE_FOR_altivec_vpku[hw]um to
10443 CODE_FOR_altivec_vpku[hw]um_direct.
10444 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
10445 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
10446 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
10447 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
10449 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10451 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
10452 generation for -maltivec=be.
10453 (altivec_vsumsws): Simplify redundant test.
10455 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10457 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
10458 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
10459 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
10460 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
10461 gen_altivec_vpkuwum.
10462 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
10464 (altivec_vpks<VI_char>ss): Likewise.
10465 (altivec_vpks<VI_char>us): Likewise.
10466 (altivec_vpku<VI_char>us): Likewise.
10467 (altivec_vpku<VI_char>um): Likewise.
10468 (altivec_vpku<VI_char>um_direct): New (copy of
10469 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
10471 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
10472 target is little endian and -maltivec=be is not specified.
10473 (*altivec_vupkhs<VU_char>_direct): New (copy of
10474 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
10475 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
10476 target is little endian and -maltivec=be is not specified.
10477 (*altivec_vupkls<VU_char>_direct): New (copy of
10478 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
10479 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
10480 little endian and -maltivec=be is not specified.
10481 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
10482 little endian and -maltivec=be is not specified.
10484 2014-02-05 Richard Henderson <rth@redhat.com>
10487 * combine-stack-adj.c: Revert r206943.
10488 * sched-int.h (struct deps_desc): Add last_args_size.
10489 * sched-deps.c (init_deps): Initialize it.
10490 (sched_analyze_insn): Add OUTPUT dependencies between insns that
10491 contain REG_ARGS_SIZE notes.
10493 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10495 * lto-cgraph.c (asm_nodes_output): Make global.
10496 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
10497 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
10498 (driver_handle_option): Handle OPT_fwpa.
10500 2014-02-05 Jakub Jelinek <jakub@redhat.com>
10503 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
10504 a comment typo and formatting issue. If odr_hash hasn't been
10505 created, return vNULL and set *completep to false.
10507 PR middle-end/57499
10508 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
10509 bb with no successors.
10511 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
10514 * doc/invoke.texi (-march): Clarify documentation for ARM.
10515 (-mtune): Likewise.
10518 2014-02-05 Richard Biener <rguenther@suse.de>
10520 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
10521 when not vectorizing because of too many alias checks.
10522 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
10523 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
10525 2014-02-05 Nick Clifton <nickc@redhat.com>
10527 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
10528 accept extended registers in any mode when compiling for the MN10300.
10530 2014-02-05 Yury Gribov <y.gribov@samsung.com>
10532 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
10533 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
10534 sanitization attributes.
10535 (can_inline_edge_p): Likewise.
10536 (sanitize_attrs_match_for_inline_p): New function.
10538 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10540 * ipa-prop.c (detect_type_change): Shor circuit testing of
10541 type changes on THIS pointer.
10543 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
10546 * config/pa/pa.c (legitimize_tls_address): Return original address
10547 if not passed a SYMBOL_REF rtx.
10548 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
10550 (pa_emit_move_sequence): Simplify TLS source operands.
10551 (pa_legitimate_constant_p): Reject all TLS constants.
10552 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
10553 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
10555 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10557 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
10558 groups when we know they are controlled by LTO.
10559 * varasm.c (default_binds_local_p_1): If object is in other partition,
10560 it will be resolved locally.
10562 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10564 * config/host-linux.c (linux_gt_pch_use_address): Don't
10565 use SSIZE_MAX because it is not always defined.
10567 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
10570 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
10571 threshold for pseudo splitting.
10572 (update_ebb_live_info): Process call argument hard registers and
10573 hard registers from insn definition too.
10574 (max_small_class_regs_num): New constant.
10575 (inherit_in_ebb): Update live hard regs through EBBs. Update
10576 reloads_num only for small register classes. Don't split for
10579 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
10582 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
10585 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10587 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
10588 visibility is safe.
10590 2014-02-04 Marek Polacek <polacek@redhat.com>
10592 * gdbinit.in (pel): Define.
10594 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10596 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
10599 2014-02-04 Richard Biener <rguenther@suse.de>
10602 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
10603 in function context local.
10604 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
10605 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
10606 similar to LTO_imported_decl_ref.
10608 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10610 PR tree-optimization/60002
10611 * cgraphclones.c (build_function_decl_skip_args): Clear
10612 DECL_LANG_SPECIFIC.
10614 PR tree-optimization/60023
10615 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
10616 false to gsi_replace.
10617 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
10618 has been in some EH region and vec_stmt could throw, add
10619 vec_stmt into the same EH region.
10620 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
10621 has no lhs, ignore it.
10622 * internal-fn.c (expand_MASK_LOAD): Likewise.
10625 * tree-inline.c (copy_forbidden): Fail for
10626 __attribute__((optimize (0))) functions.
10629 * omp-low.c (simd_clone_struct_copy): If from->inbranch
10630 is set, copy one less argument.
10631 (expand_simd_clones): Don't subtract clone_info->inbranch
10632 from simd_clone_struct_alloc argument.
10634 PR rtl-optimization/57915
10635 * recog.c (simplify_while_replacing): If all unary/binary/relational
10636 operation arguments are constant, attempt to simplify those.
10638 PR middle-end/59261
10639 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
10640 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
10642 2014-02-04 Richard Biener <rguenther@suse.de>
10644 PR tree-optimization/60012
10645 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
10646 TBAA disambiguation to all DDRs.
10648 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10651 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
10652 (LINK_SPEC): Use it for -shared, -shared-libgcc.
10654 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10657 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
10659 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10661 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
10662 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
10664 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10667 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
10668 to figure out targets of polymorphic calls with known decl.
10669 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10670 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
10671 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
10672 (get_polymorphic_call_info): ... here.
10673 (get_polymorphic_call_info_from_invariant): New function.
10675 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10677 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
10678 lookup via vtable pointer; check for type consistency
10679 and turn inconsitent facts into UNREACHABLE.
10680 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10681 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
10682 type inconsistent querries; return UNREACHABLE instead.
10684 2014-02-03 Richard Henderson <rth@twiddle.net>
10687 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
10688 already processed this node.
10689 (normalize_one_pred_1): Pass along mark_set.
10690 (normalize_one_pred): Create and destroy a pointer_set_t.
10691 (normalize_one_pred_chain): Likewise.
10693 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
10695 PR gcov-profile/58602
10696 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
10698 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10701 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
10702 -fno-devirtualize; try to devirtualize by the knowledge of
10703 virtual table pointer given by aggregate propagation.
10704 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10705 (ipa_print_node_jump_functions): Dump also offset that
10706 is relevant for polymorphic calls.
10707 (determine_known_aggregate_parts): Add arg_type parameter; use it
10708 instead of determining the type from pointer type.
10709 (ipa_compute_jump_functions_for_edge): Update call of
10710 determine_known_aggregate_parts.
10711 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
10712 (gimple_get_virt_method_for_binfo): ... here; simplify using
10713 vtable_pointer_value_to_vtable.
10714 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
10715 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
10716 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
10717 (vtable_pointer_value_to_vtable): Break out from ...; handle also
10719 (vtable_pointer_value_to_binfo): ... here.
10720 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
10722 2014-02-03 Teresa Johnson <tejohnson@google.com>
10724 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10725 redef of outer loop index variable.
10727 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10731 * doc/extend.texi (Function Attributes): Typo.
10733 2014-02-03 Cong Hou <congh@google.com>
10735 PR tree-optimization/60000
10736 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10737 if the vectorized statement is a store. A store statement can only
10738 appear at the end of pattern statements.
10740 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10742 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10743 (ix86_option_override_internal): Default long double to 64-bit for
10744 32-bit Bionic and to 128-bit for 64-bit Bionic.
10746 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10747 TARGET_LONG_DOUBLE_128 is true.
10748 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10750 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10751 (mlong-double-64): Negate -mlong-double-128.
10752 (mlong-double-128): New option.
10754 * config/i386/i386-c.c (ix86_target_macros): Define
10755 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10757 * doc/invoke.texi: Document -mlong-double-128.
10759 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10761 PR rtl-optimization/60024
10762 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10764 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10766 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10768 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10770 PR rtl-optimization/57662
10771 * sel-sched.c (code_motion_path_driver): Do not mark already not
10772 existing blocks in the visiting bitmap.
10774 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10776 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10777 on the insn being emitted.
10779 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10780 Will Deacon <will.deacon@arm.com>
10782 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
10784 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10786 * config/arm/arm-tables.opt: Regenerate.
10788 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10790 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
10791 for vector types other than V16QImode.
10792 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
10793 define_expand, and call altivec_expand_vec_perm_le when producing
10794 code with little endian element order.
10795 (*altivec_vperm_<mode>_internal): New insn having previous
10796 behavior of altivec_vperm_<mode>.
10797 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
10798 altivec_expand_vec_perm_le when producing code with little endian
10800 (*altivec_vperm_<mode>_uns_internal): New insn having previous
10801 behavior of altivec_vperm_<mode>_uns.
10803 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10805 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
10806 (altivec_vsumsws): Add handling for -maltivec=be with a little
10808 (altivec_vsumsws_direct): New.
10809 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
10810 gen_altivec_vsumsws.
10812 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10814 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
10815 vtable_pointer_value_to_binfo): New functions.
10816 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
10817 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
10819 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
10821 * config/nios2/nios2.md (load_got_register): Initialize GOT
10822 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
10823 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
10825 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10827 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
10828 preserverd by passthrough, do not propagate the type.
10830 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10832 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
10833 (mips_atomic_assign_expand_fenv): New function.
10834 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
10836 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10838 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
10839 (__builtin_mips_set_fcsr): Likewise.
10840 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
10841 MIPS_USI_FTYPE_VOID.
10842 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
10843 (mips16_expand_set_fcsr): Likewise.
10844 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
10845 (mips16_set_fcsr_stub): Likewise.
10846 (mips16_get_fcsr_one_only_stub): New class.
10847 (mips16_set_fcsr_one_only_stub): Likewise.
10848 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
10849 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
10850 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
10851 (hard_float): New availability predicate.
10852 (mips_builtins): Add get_fcsr and set_fcsr.
10853 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
10854 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
10855 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
10856 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
10857 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
10860 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10862 * config/mips/mips.c (mips_one_only_stub): New class.
10863 (mips_need_mips16_rdhwr_p): Replace with...
10864 (mips16_rdhwr_stub): ...this new variable.
10865 (mips16_stub_call_address): New function.
10866 (mips16_rdhwr_one_only_stub): New class.
10867 (mips_expand_thread_pointer): Use mips16_stub_call_address.
10868 (mips_output_mips16_rdhwr): Delete.
10869 (mips_finish_stub): New function.
10870 (mips_code_end): Use it to handle rdhwr stubs.
10872 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
10875 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
10876 when calculating size of integer atomic types.
10878 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
10880 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
10882 2014-02-01 Jakub Jelinek <jakub@redhat.com>
10884 PR tree-optimization/60003
10885 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
10886 * profile.c (branch_prob): Use gimple_call_builtin_p
10887 to check for BUILT_IN_SETJMP_RECEIVER.
10888 * tree-inline.c (copy_bb): Call notice_special_calls.
10890 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
10893 * lra-constraints.c (process_alt_operands): Update reload_sum only
10896 2014-01-31 Richard Henderson <rth@redhat.com>
10898 PR middle-end/60004
10899 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
10900 until after else_eh is processed.
10902 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10904 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
10905 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
10906 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
10907 in smmintrin.h, remove them.
10908 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
10909 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
10910 * config/i386/i386.md (ROUND_SAE): Fix value.
10911 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
10912 (const48_operand): New.
10913 * config/i386/subst.md (round), (round_expand): Use
10914 const_4_or_8_to_11_operand.
10915 (round_saeonly), (round_saeonly_expand): Use const48_operand.
10917 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10919 * config/i386/constraints.md (Yk): Swap meaning with k.
10920 * config/i386/i386.md (movhi_internal): Change Yk to k.
10921 (movqi_internal): Ditto.
10922 (*k<logic><mode>): Ditto.
10925 (kandn<mode>): Ditto.
10926 (*<code>hi_1): Ditto.
10927 (*<code>qi_1): Ditto.
10928 (kxnor<mode>): Ditto.
10929 (kortestzhi): Ditto.
10930 (kortestchi): Ditto.
10932 (*one_cmplhi2_1): Ditto.
10933 (*one_cmplqi2_1): Ditto.
10934 * config/i386/sse.md (): Change k to Yk.
10935 (avx512f_load<mode>_mask): Ditto.
10936 (avx512f_blendm<mode>): Ditto.
10937 (avx512f_store<mode>_mask): Ditto.
10938 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
10939 (avx512f_storedqu<mode>_mask): Ditto.
10940 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
10942 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
10943 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
10944 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
10945 (avx512f_maskcmp<mode>3): Ditto.
10946 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
10947 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
10948 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
10949 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
10950 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
10951 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
10952 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
10953 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
10954 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
10955 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
10956 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
10957 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
10958 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
10959 (vec_extract_lo_<mode>_maskm): Ditto.
10960 (vec_extract_hi_<mode>_maskm): Ditto.
10961 (avx512f_vternlog<mode>_mask): Ditto.
10962 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
10963 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
10964 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
10965 (avx512f_<code>v8div16qi2_mask): Ditto.
10966 (avx512f_<code>v8div16qi2_mask_store): Ditto.
10967 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
10968 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
10969 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
10970 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
10971 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10972 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10973 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10974 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10975 (avx512cd_maskb_vec_dupv8di): Ditto.
10976 (avx512cd_maskw_vec_dupv16si): Ditto.
10977 (avx512f_vpermi2var<mode>3_maskz): Ditto.
10978 (avx512f_vpermi2var<mode>3_mask): Ditto.
10979 (avx512f_vpermi2var<mode>3_mask): Ditto.
10980 (avx512f_vpermt2var<mode>3_maskz): Ditto.
10981 (*avx512f_gathersi<mode>): Ditto.
10982 (*avx512f_gathersi<mode>_2): Ditto.
10983 (*avx512f_gatherdi<mode>): Ditto.
10984 (*avx512f_gatherdi<mode>_2): Ditto.
10985 (*avx512f_scattersi<mode>): Ditto.
10986 (*avx512f_scatterdi<mode>): Ditto.
10987 (avx512f_compress<mode>_mask): Ditto.
10988 (avx512f_compressstore<mode>_mask): Ditto.
10989 (avx512f_expand<mode>_mask): Ditto.
10990 * config/i386/subst.md (mask): Change k to Yk.
10991 (mask_scalar_merge): Ditto.
10994 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
10996 * doc/extend.texi (Vector Extensions): Document ?: in C++.
10998 2014-01-31 Richard Biener <rguenther@suse.de>
11000 PR middle-end/59990
11001 * builtins.c (fold_builtin_memory_op): Make sure to not
11002 use a floating-point mode or a boolean or enumeral type for
11003 the copy operation.
11005 2014-01-30 DJ Delorie <dj@redhat.com>
11007 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
11008 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
11009 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
11010 whenever main() has an epilogue.
11012 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11014 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
11015 unused variable "field".
11016 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
11017 (vsx_mergeh_<mode>): Likewise.
11018 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
11019 (altivec_vmrghh): Likewise.
11020 (altivec_vmrghw): Likewise.
11021 (altivec_vmrglb): Likewise.
11022 (altivec_vmrglh): Likewise.
11023 (altivec_vmrglw): Likewise.
11024 (altivec_vspltb): Add missing uses.
11025 (altivec_vsplth): Likewise.
11026 (altivec_vspltw): Likewise.
11027 (altivec_vspltsf): Likewise.
11029 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11032 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
11033 frame related instructions.
11035 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
11037 PR rtl-optimization/59959
11038 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
11039 any reload of register whose subreg is invalid.
11041 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11043 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
11044 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
11045 Add missing return type - void.
11047 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11049 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
11050 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
11051 remove element index adjustment for endian (now handled in vsx.md
11053 (altivec_expand_vec_perm_const): Use
11054 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
11055 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
11056 (vsx_xxspltw_<mode>): Adjust element index for little endian.
11057 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
11058 define_expand and a new define_insn *altivec_vspltb_internal;
11059 adjust for -maltivec=be on a little endian target.
11060 (altivec_vspltb_direct): New.
11061 (altivec_vsplth): Divide into a define_expand and a new
11062 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
11063 little endian target.
11064 (altivec_vsplth_direct): New.
11065 (altivec_vspltw): Divide into a define_expand and a new
11066 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
11067 little endian target.
11068 (altivec_vspltw_direct): New.
11069 (altivec_vspltsf): Divide into a define_expand and a new
11070 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
11071 a little endian target.
11073 2014-01-30 Richard Biener <rguenther@suse.de>
11075 PR tree-optimization/59993
11076 * tree-ssa-forwprop.c (associate_pointerplus): Check we
11077 can propagate form the earlier stmt and avoid the transform
11078 when the intermediate result is needed.
11080 2014-01-30 Alangi Derick <alangiderick@gmail.com>
11082 * README.Portability: Fix typo.
11084 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
11086 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
11087 comparison_operator with ordered_comparison_operator.
11089 2014-01-30 Nick Clifton <nickc@redhat.com>
11091 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
11092 Rename to mn10300_store_multiple_regs.
11093 * config/mn10300/mn10300.c: Likewise.
11094 * config/mn10300/mn10300.md (store_movm): Fix typo: call
11095 store_multiple_regs.
11096 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
11097 Call mn10300_store_multiple_regs.
11099 2014-01-30 Nick Clifton <nickc@redhat.com>
11100 DJ Delorie <dj@redhat.com>
11102 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
11103 %fp 2 to keep registers after it properly word-aligned.
11104 (rl78_alloc_physical_registers_umul): Handle the case where both
11105 input operands are the same.
11107 2014-01-30 Richard Biener <rguenther@suse.de>
11109 PR tree-optimization/59903
11110 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
11113 2014-01-30 Jason Merrill <jason@redhat.com>
11116 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
11119 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
11121 2014-01-30 Richard Biener <rguenther@suse.de>
11123 PR tree-optimization/59951
11124 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
11126 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
11129 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
11130 SFmode to DFmode case.
11132 2014-01-29 DJ Delorie <dj@redhat.com>
11134 * config/msp430/msp430.opt (-minrt): New.
11135 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
11137 (ENDFILE_SPEC): Likewise.
11139 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
11141 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
11142 (estimate_function_body_sizes): Use it.
11144 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
11147 * dwarf2out.c (is_cxx_auto): New.
11148 (is_base_type): Use it.
11149 (gen_type_die_with_usage): Likewise.
11151 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11153 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
11154 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
11155 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
11156 -maltivec=be with LE targets.
11157 (vsx_mergeh_<mode>): Likewise.
11158 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
11159 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
11160 (altivec_vmrghb): Replace with define_expand and new
11161 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
11162 (altivec_vmrghb_direct): New define_insn.
11163 (altivec_vmrghh): Replace with define_expand and new
11164 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
11165 (altivec_vmrghh_direct): New define_insn.
11166 (altivec_vmrghw): Replace with define_expand and new
11167 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
11168 (altivec_vmrghw_direct): New define_insn.
11169 (*altivec_vmrghsf): Adjust for endianness.
11170 (altivec_vmrglb): Replace with define_expand and new
11171 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
11172 (altivec_vmrglb_direct): New define_insn.
11173 (altivec_vmrglh): Replace with define_expand and new
11174 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
11175 (altivec_vmrglh_direct): New define_insn.
11176 (altivec_vmrglw): Replace with define_expand and new
11177 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
11178 (altivec_vmrglw_direct): New define_insn.
11179 (*altivec_vmrglsf): Adjust for endianness.
11180 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11181 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11182 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11183 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11184 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11185 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11186 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11187 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11189 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
11191 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
11192 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
11195 2014-01-29 Richard Biener <rguenther@suse.de>
11197 PR tree-optimization/58742
11198 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
11199 associate_pointerplus_align.
11200 (associate_pointerplus_diff): New function.
11201 (associate_pointerplus): Likewise. Call associate_pointerplus_align
11202 and associate_pointerplus_diff.
11204 2014-01-29 Richard Biener <rguenther@suse.de>
11206 * lto-streamer.h (LTO_major_version): Bump to 3.
11207 (LTO_minor_version): Reset to 0.
11209 2014-01-29 Renlin Li <Renlin.Li@arm.com>
11211 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
11212 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
11213 (arm_file_start): Generate correct asm header for armv7ve.
11214 * config/arm/bpabi.h: Add multilib support for armv7ve.
11215 * config/arm/driver-arm.c: Change the architectures of cortex-a7
11216 and cortex-a15 to armv7ve.
11217 * config/arm/t-aprofile: Add multilib support for armv7ve.
11218 * doc/invoke.texi: Document -march=armv7ve.
11220 2014-01-29 Richard Biener <rguenther@suse.de>
11222 PR tree-optimization/58742
11223 * tree-ssa-forwprop.c (associate_plusminus): Return true
11224 if we changed sth, defer EH cleanup to ...
11225 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
11226 (simplify_mult): New function.
11228 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11230 PR middle-end/59917
11231 PR tree-optimization/59920
11232 * tree.c (build_common_builtin_nodes): Remove
11233 __builtin_setjmp_dispatcher initialization.
11234 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
11235 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
11236 instead of gsi_after_labels + manually skipping debug stmts.
11237 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
11238 ignore bbs with IFN_ABNORMAL_DISPATCHER.
11239 * tree-inline.c (copy_edges_for_bb): Remove
11240 can_make_abnormal_goto argument, instead add abnormal_goto_dest
11241 argument. Ignore computed_goto_p stmts. Don't call
11242 make_abnormal_goto_edges. If a call might need abnormal edges
11243 for non-local gotos, see if it already has an edge to
11244 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
11245 with true argument, don't do anything then, otherwise add
11246 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
11247 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
11249 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
11250 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
11251 (lower_stmt): Don't set data->calls_builtin_setjmp.
11252 (lower_builtin_setjmp): Adjust comment.
11253 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
11254 * tree-cfg.c (found_computed_goto): Remove.
11255 (factor_computed_gotos): Remove.
11256 (make_goto_expr_edges): Return bool, true for computed gotos.
11257 Don't call make_abnormal_goto_edges.
11258 (build_gimple_cfg): Don't set found_computed_goto, don't call
11259 factor_computed_gotos.
11260 (computed_goto_p): No longer static.
11261 (make_blocks): Don't set found_computed_goto.
11262 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
11263 (make_edges): If make_goto_expr_edges returns true, push bb
11264 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
11265 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
11266 vector. Record mapping between bbs and OpenMP regions if there
11267 are any, adjust make_gimple_omp_edges caller. Call
11268 handle_abnormal_edges.
11269 (make_abnormal_goto_edges): Remove.
11270 * tree-cfg.h (make_abnormal_goto_edges): Remove.
11271 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
11272 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
11273 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
11274 * internal-fn.def (ABNORMAL_DISPATCHER): New.
11275 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
11276 filling *region also set *region_idx to (*region)->entry->index.
11279 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
11280 For REGs set ORIGINAL_REGNO.
11282 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
11284 * doc/md.texi: Mention that a target shouldn't implement
11285 vec_widen_(s|u)mul_even/odd pair if it is less efficient
11288 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11290 PR tree-optimization/59594
11291 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
11292 a copy of the datarefs vector rather than the vector itself.
11294 2014-01-28 Jason Merrill <jason@redhat.com>
11297 * dwarf2out.c (auto_die): New static.
11298 (gen_type_die_with_usage): Handle C++1y 'auto'.
11299 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
11302 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
11305 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
11306 (SPEC_X32): Likewise.
11307 (SPEC_64): Likewise.
11308 * config/i386/i386.c (ix86_option_override_internal): Turn off
11309 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
11311 (x86_file_start): Output .code16gcc for TARGET_16BIT.
11312 * config/i386/i386.h (TARGET_16BIT): New macro.
11313 (TARGET_16BIT_P): Likewise.
11314 * config/i386/i386.opt: Add m16.
11315 * doc/invoke.texi: Document -m16.
11317 2014-01-28 Jakub Jelinek <jakub@redhat.com>
11319 PR preprocessor/59935
11320 * input.c (location_get_source_line): Bail out on when line number
11321 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
11323 2014-01-28 Richard Biener <rguenther@suse.de>
11325 PR tree-optimization/58742
11326 * tree-ssa-forwprop.c (associate_plusminus): Handle
11327 pointer subtraction of the form (T)(P + A) - (T)P.
11329 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11331 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
11334 2014-01-28 Richard Biener <rguenther@suse.de>
11337 2014-01-28 Richard Biener <rguenther@suse.de>
11339 PR rtl-optimization/45364
11340 PR rtl-optimization/59890
11341 * var-tracking.c (local_get_addr_clear_given_value): Handle
11342 already cleared slot.
11343 (val_reset): Handle not allocated local_get_addr_cache.
11344 (vt_find_locations): Use post-order on the inverted CFG.
11346 2014-01-28 Richard Biener <rguenther@suse.de>
11348 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
11350 2014-01-28 Richard Biener <rguenther@suse.de>
11352 PR rtl-optimization/45364
11353 PR rtl-optimization/59890
11354 * var-tracking.c (local_get_addr_clear_given_value): Handle
11355 already cleared slot.
11356 (val_reset): Handle not allocated local_get_addr_cache.
11357 (vt_find_locations): Use post-order on the inverted CFG.
11359 2014-01-28 Alan Modra <amodra@gmail.com>
11361 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
11362 * configure.ac <recursive call for build != host>: Define
11363 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
11364 and LD_FOR_BUILD too.
11365 * configure: Regenerate.
11367 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
11369 * config/i386/i386.c (get_builtin_code_for_version): Separate
11370 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
11371 Broadwell from Haswell.
11373 2014-01-27 Steve Ellcey <sellcey@mips.com>
11375 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
11376 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
11377 * config/mips/mips.c (mips_option_override): Change setting
11379 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
11380 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
11381 Change from Mask to Var.
11383 2014-01-27 Jeff Law <law@redhat.com>
11385 * ipa-inline.c (inline_small_functions): Fix typo.
11387 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
11389 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
11390 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
11391 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
11392 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
11393 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
11394 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
11395 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
11396 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
11397 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
11398 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
11399 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
11400 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
11401 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
11402 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
11403 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
11404 (_mm512_storeu_epi64): Ditto.
11405 (_mm512_cmpge_epi32_mask): Ditto.
11406 (_mm512_cmpge_epu32_mask): Ditto.
11407 (_mm512_cmpge_epi64_mask): Ditto.
11408 (_mm512_cmpge_epu64_mask): Ditto.
11409 (_mm512_cmple_epi32_mask): Ditto.
11410 (_mm512_cmple_epu32_mask): Ditto.
11411 (_mm512_cmple_epi64_mask): Ditto.
11412 (_mm512_cmple_epu64_mask): Ditto.
11413 (_mm512_cmplt_epi32_mask): Ditto.
11414 (_mm512_cmplt_epu32_mask): Ditto.
11415 (_mm512_cmplt_epi64_mask): Ditto.
11416 (_mm512_cmplt_epu64_mask): Ditto.
11417 (_mm512_cmpneq_epi32_mask): Ditto.
11418 (_mm512_cmpneq_epu32_mask): Ditto.
11419 (_mm512_cmpneq_epi64_mask): Ditto.
11420 (_mm512_cmpneq_epu64_mask): Ditto.
11421 (_mm512_expand_pd): Ditto.
11422 (_mm512_expand_ps): Ditto.
11423 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
11424 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
11425 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
11426 * config/i386/i386.c (ix86_builtins): Add
11427 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
11428 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
11429 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
11430 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
11431 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
11432 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
11433 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
11434 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
11435 IX86_BUILTIN_PMOVUSQW512_MEM.
11436 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
11437 __builtin_ia32_pmovsqd512mem_mask,
11438 __builtin_ia32_pmovqd512mem_mask,
11439 __builtin_ia32_pmovusqw512mem_mask,
11440 __builtin_ia32_pmovsqw512mem_mask,
11441 __builtin_ia32_pmovqw512mem_mask,
11442 __builtin_ia32_pmovusdw512mem_mask,
11443 __builtin_ia32_pmovsdw512mem_mask,
11444 __builtin_ia32_pmovdw512mem_mask,
11445 __builtin_ia32_pmovqb512mem_mask,
11446 __builtin_ia32_pmovusqb512mem_mask,
11447 __builtin_ia32_pmovsqb512mem_mask,
11448 __builtin_ia32_pmovusdb512mem_mask,
11449 __builtin_ia32_pmovsdb512mem_mask,
11450 __builtin_ia32_pmovdb512mem_mask.
11451 (bdesc_args): Add __builtin_ia32_expanddf512,
11452 __builtin_ia32_expandsf512.
11453 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
11454 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
11455 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
11456 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
11457 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
11458 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
11459 (avx512f_<code>v8div16qi2_mask_store): This.
11460 (avx512f_expand<mode>): New.
11462 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
11464 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
11466 (_mm512_mask_prefetch_i64gather_pd): Ditto.
11467 (_mm512_prefetch_i32scatter_pd): Ditto.
11468 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
11469 (_mm512_prefetch_i64scatter_pd): Ditto.
11470 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
11471 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
11472 (_mm512_mask_prefetch_i64gather_ps): Ditto.
11473 (_mm512_prefetch_i32scatter_ps): Ditto.
11474 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
11475 (_mm512_prefetch_i64scatter_ps): Ditto.
11476 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
11477 * config/i386/i386-builtin-types.def: Define
11478 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
11479 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
11480 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
11481 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
11482 IX86_BUILTIN_SCATTERPFQPD.
11483 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
11484 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
11485 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
11486 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
11487 __builtin_ia32_scatterpfqps.
11488 (ix86_expand_builtin): Expand new built-ins.
11489 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
11490 fix memory access data type.
11491 (*avx512pf_gatherpf<mode>_mask): Ditto.
11492 (*avx512pf_gatherpf<mode>): Ditto.
11493 (avx512pf_scatterpf<mode>): Ditto.
11494 (*avx512pf_scatterpf<mode>_mask): Ditto.
11495 (*avx512pf_scatterpf<mode>): Ditto.
11496 (GATHER_SCATTER_SF_MEM_MODE): New.
11497 (avx512pf_gatherpf<mode>df): Ditto.
11498 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11499 (*avx512pf_scatterpf<mode>df): Ditto.
11501 2014-01-27 Jakub Jelinek <jakub@redhat.com>
11504 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
11505 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
11508 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11510 * common/config/arm/arm-common.c
11511 (arm_rewrite_mcpu): Handle multiple names.
11513 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11515 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11517 * gimple-builder.h (create_gimple_tmp): Delete.
11519 2014-01-27 Christian Bruel <christian.bruel@st.com>
11521 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
11524 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
11526 * config/pa/pa.md (call): Generate indirect long calls to non-local
11527 functions when outputing 32-bit code.
11528 (call_value): Likewise except for special call to buggy powf function.
11530 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
11531 portable runtime and PIC indirect calls.
11532 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
11533 and PIC call sequences. Use ldo instead of blr to set return register
11534 in PIC call sequence.
11536 2014-01-25 Walter Lee <walt@tilera.com>
11538 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
11539 avoid clobbering a live register.
11541 2014-01-25 Walter Lee <walt@tilera.com>
11543 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
11544 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
11545 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
11546 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
11548 2014-01-25 Walter Lee <walt@tilera.com>
11550 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
11551 arguments on even registers.
11552 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
11554 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
11555 (BIGGEST_ALIGNMENT): Ditto.
11556 (BIGGEST_FIELD_ALIGNMENT): Ditto.
11558 2014-01-25 Walter Lee <walt@tilera.com>
11560 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
11561 insns before bundling.
11562 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
11564 2014-01-25 Walter Lee <walt@tilera.com>
11566 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
11567 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
11568 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
11570 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11572 * config/mips/constraints.md (kl): Delete.
11573 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
11574 define expands, using...
11575 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
11576 instructions for MIPS16.
11577 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
11578 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
11580 2014-01-25 Walter Lee <walt@tilera.com>
11582 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
11586 2014-01-25 Walter Lee <walt@tilera.com>
11588 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
11589 (TARGET_EXPAND_TO_RTL_HOOK): Define.
11591 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11593 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
11596 2014-01-25 Jakub Jelinek <jakub@redhat.com>
11598 * print-rtl.c (in_call_function_usage): New var.
11599 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
11600 EXPR_LIST mode as mode and not as reg note name.
11602 PR middle-end/59561
11603 * cfgloopmanip.c (copy_loop_info): If
11604 loop->warned_aggressive_loop_optimizations, make sure
11605 the flag is set in target loop too.
11607 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
11609 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
11611 * builtins.def: Likewise.
11612 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
11613 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
11614 * ira.c (ira_setup_eliminable_regset): Likewise.
11615 * omp-low.c (gate_expand_omp): Likewise.
11616 (execute_lower_omp): Likewise.
11617 (diagnose_sb_0): Likewise.
11618 (gate_diagnose_omp_blocks): Likewise.
11619 (simd_clone_clauses_extract): Likewise.
11622 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11624 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
11625 correction for little endian...
11626 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
11629 2014-01-24 Jeff Law <law@redhat.com>
11631 PR tree-optimization/59919
11632 * tree-vrp.c (find_assert_locations_1): Do not register asserts
11633 for non-returning calls.
11635 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
11637 * common/config/aarch64/aarch64-common.c
11638 (aarch64_rewrite_mcpu): Handle multiple names.
11639 * config/aarch64/aarch64.h
11640 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11642 2014-01-24 Dodji Seketeli <dodji@redhat.com>
11644 * input.c (add_file_to_cache_tab): Handle the case where fopen
11647 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
11650 * config/i386/i386.md (pushsf splitter): Get stack adjustment
11651 from push operand if code of push isn't PRE_DEC.
11653 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11656 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
11657 -mquad-memory-atomic. Update -mquad-memory documentation to say
11658 it is only used for non-atomic loads/stores.
11660 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
11661 -mquad-memory or -mquad-memory-atomic switches.
11663 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
11664 -mquad-memory-atomic to ISA 2.07 support.
11666 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
11667 to separate support of normal quad word memory operations (ldq, stq)
11668 from the atomic quad word memory operations.
11670 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
11671 support to separate non-atomic quad word operations from atomic
11672 quad word operations. Disable non-atomic quad word operations in
11673 little endian mode so that we don't have to swap words after the
11674 load and before the store.
11675 (quad_load_store_p): Add comment about atomic quad word support.
11676 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
11677 options printed with -mdebug=reg.
11679 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
11680 -mquad-memory-atomic as the test for whether we have quad word
11681 atomic instructions.
11682 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
11683 or -mp8-vector are used, allow byte/half-word atomic operations.
11685 * config/rs6000/sync.md (load_lockedti): Insure that the address
11686 is a proper indexed or indirect address for the lqarx instruction.
11687 On little endian systems, swap the hi/lo registers after the lqarx
11689 (load_lockedpti): Use indexed_or_indirect_operand predicate to
11690 insure the address is valid for the lqarx instruction.
11691 (store_conditionalti): Insure that the address is a proper indexed
11692 or indirect address for the stqcrx. instruction. On little endian
11693 systems, swap the hi/lo registers before doing the stqcrx.
11695 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
11696 insure the address is valid for the stqcrx. instruction.
11698 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11699 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
11700 type of quad memory support is available.
11702 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
11704 PR regression/59915
11705 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
11706 there is a danger of looping.
11708 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
11710 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11711 force flag_ira_loop_pressure if set via command line.
11713 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11715 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
11716 (ashr_simd): New builtin handling DI mode.
11717 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
11718 (aarch64_sshr_simddi): New match pattern.
11719 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
11720 (vshrd_n_s64): Likewise.
11721 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11723 2014-01-23 Nick Clifton <nickc@redhat.com>
11725 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11726 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11727 favour of mcu specific scripts.
11728 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11731 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11732 Alex Velenko <Alex.Velenko@arm.com>
11734 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11735 (vaddv_s16): Likewise.
11736 (vaddv_s32): Likewise.
11737 (vaddv_u8): Likewise.
11738 (vaddv_u16): Likewise.
11739 (vaddv_u32): Likewise.
11740 (vaddvq_s8): Likewise.
11741 (vaddvq_s16): Likewise.
11742 (vaddvq_s32): Likewise.
11743 (vaddvq_s64): Likewise.
11744 (vaddvq_u8): Likewise.
11745 (vaddvq_u16): Likewise.
11746 (vaddvq_u32): Likewise.
11747 (vaddvq_u64): Likewise.
11748 (vaddv_f32): Likewise.
11749 (vaddvq_f32): Likewise.
11750 (vaddvq_f64): Likewise.
11751 (vmaxv_f32): Likewise.
11752 (vmaxv_s8): Likewise.
11753 (vmaxv_s16): Likewise.
11754 (vmaxv_s32): Likewise.
11755 (vmaxv_u8): Likewise.
11756 (vmaxv_u16): Likewise.
11757 (vmaxv_u32): Likewise.
11758 (vmaxvq_f32): Likewise.
11759 (vmaxvq_f64): Likewise.
11760 (vmaxvq_s8): Likewise.
11761 (vmaxvq_s16): Likewise.
11762 (vmaxvq_s32): Likewise.
11763 (vmaxvq_u8): Likewise.
11764 (vmaxvq_u16): Likewise.
11765 (vmaxvq_u32): Likewise.
11766 (vmaxnmv_f32): Likewise.
11767 (vmaxnmvq_f32): Likewise.
11768 (vmaxnmvq_f64): Likewise.
11769 (vminv_f32): Likewise.
11770 (vminv_s8): Likewise.
11771 (vminv_s16): Likewise.
11772 (vminv_s32): Likewise.
11773 (vminv_u8): Likewise.
11774 (vminv_u16): Likewise.
11775 (vminv_u32): Likewise.
11776 (vminvq_f32): Likewise.
11777 (vminvq_f64): Likewise.
11778 (vminvq_s8): Likewise.
11779 (vminvq_s16): Likewise.
11780 (vminvq_s32): Likewise.
11781 (vminvq_u8): Likewise.
11782 (vminvq_u16): Likewise.
11783 (vminvq_u32): Likewise.
11784 (vminnmv_f32): Likewise.
11785 (vminnmvq_f32): Likewise.
11786 (vminnmvq_f64): Likewise.
11788 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11790 * config/aarch64/aarch64-simd.md
11791 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
11792 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
11793 (*aarch64_mul3_elt<mode>): Likewise.
11794 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
11795 (*aarch64_mul3_elt_to_64v2df): Likewise.
11796 (*aarch64_mla_elt<mode>): Likewise.
11797 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11798 (*aarch64_mls_elt<mode>): Likewise.
11799 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11800 (*aarch64_fma4_elt<mode>): Likewise.
11801 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11802 (*aarch64_fma4_elt_to_64v2df): Likewise.
11803 (*aarch64_fnma4_elt<mode>): Likewise.
11804 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11805 (*aarch64_fnma4_elt_to_64v2df): Likewise.
11806 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
11807 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
11808 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11809 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11810 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
11811 (aarch64_sqdmull_lane<mode>_internal): Likewise.
11812 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
11814 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
11816 * config/aarch64/aarch64-simd.md
11817 (aarch64_be_checked_get_lane<mode>): New define_expand.
11818 * config/aarch64/aarch64-simd-builtins.def
11819 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
11820 New builtin definition.
11821 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
11822 Use new safe be builtin.
11824 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11826 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
11828 (aarch64_be_st1<mode>): Likewise.
11829 (aarch_ld1<VALL:mode>): Define_expand modified.
11830 (aarch_st1<VALL:mode>): Likewise.
11831 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
11832 (UNSPEC_ST1): Likewise.
11834 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
11836 * config/microblaze/microblaze.md: Add trap insn and attribute
11838 2014-01-23 Dodji Seketeli <dodji@redhat.com>
11840 PR preprocessor/58580
11841 * input.h (location_get_source_line): Take an additional line_size
11843 (void diagnostics_file_cache_fini): Declare new function.
11844 * input.c (struct fcache): New type.
11845 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
11846 New static constants.
11847 (diagnostic_file_cache_init, total_lines_num)
11848 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
11849 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
11850 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
11851 (get_next_line, read_next_line, goto_next_line, read_line_num):
11852 New static function definitions.
11853 (diagnostic_file_cache_fini): New function.
11854 (location_get_source_line): Take an additional output line_len
11855 parameter. Re-write using lookup_or_add_file_to_cache_tab and
11857 * diagnostic.c (diagnostic_finish): Call
11858 diagnostic_file_cache_fini.
11859 (adjust_line): Take an additional input parameter for the length
11860 of the line, rather than calculating it with strlen.
11861 (diagnostic_show_locus): Adjust the use of
11862 location_get_source_line and adjust_line with respect to their new
11863 signature. While displaying a line now, do not stop at the first
11864 null byte. Rather, display the zero byte as a space and keep
11865 going until we reach the size of the line.
11866 * Makefile.in: Add vec.o to OBJS-libcommon
11868 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11869 Ilya Tocar <ilya.tocar@intel.com>
11871 * config/i386/avx512fintrin.h (_mm512_kmov): New.
11872 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
11873 (__builtin_ia32_kmov16): Ditto.
11874 * config/i386/i386.md (UNSPEC_KMOV): New.
11877 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11879 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
11880 (_mm512_storeu_si512): Ditto.
11882 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
11885 * rtl.h (get_referenced_operands): Declare.
11886 * recog.c (get_referenced_operands): New function.
11887 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
11888 operands have been referenced when recording LO_SUM references.
11890 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
11892 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
11894 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11896 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
11897 Enable for generic and recent AMD targets.
11899 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11901 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
11902 ARG_SIZE note when adjustment was eliminated.
11904 2014-01-22 Jeff Law <law@redhat.com>
11906 PR tree-optimization/59597
11907 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
11908 in file. Accept new argument REGISTERING and use it to modify
11909 dump output appropriately.
11910 (register_jump_thread): Corresponding changes.
11911 (mark_threaded_blocks): Reinstate code to cancel unprofitable
11912 thread paths involving joiner blocks. Add code to dump cancelled
11913 jump threading paths.
11915 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
11917 PR rtl-optimization/59477
11918 * lra-constraints.c (inherit_in_ebb): Process call for living hard
11919 regs. Update reloads_num and potential_reload_hard_regs for all insns.
11921 2014-01-22 Tom Tromey <tromey@redhat.com>
11923 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
11925 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
11927 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11929 PR rtl-optimization/59896
11930 * lra-constraints.c (process_alt_operands): Check unused note for
11931 matched operands of insn with no output reloads.
11933 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
11935 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
11936 (mips_move_from_gpr_cost): Likewise.
11938 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11940 PR rtl-optimization/59858
11941 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
11942 ira_class_hard_regs_num.
11943 (process_alt_operands): Increase reject for dying matched operand.
11945 2014-01-21 Jakub Jelinek <jakub@redhat.com>
11948 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
11949 smaller than size, perform several stores or loads and stores
11950 at dst + count - size to store or copy all of size bytes, rather
11951 than just last modesize bytes.
11953 2014-01-20 DJ Delorie <dj@redhat.com>
11955 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
11956 that CLOBBERs are REGs before propogating their values.
11958 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
11960 PR middle-end/59789
11961 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
11962 (cgraph_inline_failed_type): New function.
11963 * cgraph.h (DEFCIFCODE): Add type.
11964 (cgraph_inline_failed_type_t): New enum.
11965 (cgraph_inline_failed_type): New prototype.
11966 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
11967 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
11968 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
11969 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
11970 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
11971 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
11972 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
11973 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
11974 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
11975 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
11976 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
11977 OPTIMIZATION_MISMATCH.
11978 * tree-inline.c (expand_call_inline): Emit errors during
11979 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
11981 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11984 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
11985 mode attribute in insn output.
11987 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
11989 * output.h (output_constant): Delete.
11990 * varasm.c (output_constant): Make private.
11992 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
11994 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
11996 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11998 PR middle-end/59860
11999 * tree.h (fold_builtin_strcat): New prototype.
12000 * builtins.c (fold_builtin_strcat): No longer static. Add len
12001 argument, if non-NULL, don't call c_strlen. Optimize
12002 directly into __builtin_memcpy instead of __builtin_strcpy.
12003 (fold_builtin_2): Adjust fold_builtin_strcat caller.
12004 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
12006 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12008 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12009 for SImode_address_operand operands, having only a REG argument.
12011 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
12013 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
12014 loader name using mbig-endian.
12015 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
12017 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12019 * doc/invoke.texi (-march): Clarify documentation for AArch64.
12020 (-mtune): Likewise.
12023 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
12025 * config/aarch64/aarch64-protos.h
12026 (aarch64_cannot_change_mode_class_ptr): Declare.
12027 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
12028 aarch64_cannot_change_mode_class_ptr): New.
12029 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
12030 backend hook aarch64_cannot_change_mode_class.
12032 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12034 * common/config/aarch64/aarch64-common.c
12035 (aarch64_handle_option): Don't handle any option order logic here.
12036 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
12037 selected_cpu, warn on architecture version mismatch.
12038 (aarch64_override_options): Fix parsing order for option strings.
12040 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12041 Iain Sandoe <iain@codesourcery.com>
12044 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
12045 warning. Amend comment to reflect current functionality.
12047 2014-01-20 Richard Biener <rguenther@suse.de>
12049 PR middle-end/59860
12050 * builtins.c (fold_builtin_strcat): Remove case better handled
12051 by tree-ssa-strlen.c.
12053 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
12055 * config/aarch64/aarch64.opt
12056 (mcpu, march, mtune): Make case-insensitive.
12058 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12061 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12062 if operands[1] is a REG or ZERO_EXTEND of a REG.
12064 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
12066 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
12068 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
12070 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
12071 long non-pic millicode calls.
12073 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12075 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
12077 2014-01-19 Kito Cheng <kito@0xlab.org>
12079 * builtins.c (expand_movstr): Check movstr expand done or fail.
12081 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12082 H.J. Lu <hongjiu.lu@intel.com>
12085 * config/i386/i386.md (*lea<mode>): Zero-extend return register
12086 to DImode for zero-extended addresses.
12088 2014-01-19 Jakub Jelinek <jakub@redhat.com>
12090 PR rtl-optimization/57763
12091 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
12092 on the new indirect jump_insn and increment LABEL_NUSES (label).
12094 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
12098 * config.gcc (x86_archs): New variable.
12099 (x86_64_archs): Likewise.
12100 (x86_cpus): Likewise.
12101 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
12102 --with-arch/--with-cpu= options.
12103 Support --with-arch=/--with-cpu={nehalem,westmere,
12104 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
12106 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12108 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
12109 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
12111 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12113 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
12115 2014-01-18 Jakub Jelinek <jakub@redhat.com>
12118 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
12119 clear cpp_get_options (parse_in)->warn_unused_macros for
12120 ix86_target_macros_internal with cpp_define.
12122 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
12124 * jump.c (delete_related_insns): Keep (use (insn))s.
12125 * reorg.c (redundant_insn): Check for barriers too.
12127 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12129 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
12131 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
12133 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
12134 call to $$dyncall when TARGET_LONG_CALLS is true.
12136 2014-01-17 Jeff Law <law@redhat.com>
12138 * ree.c (combine_set_extension): Temporarily disable test for
12139 changing number of hard registers.
12141 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12143 PR middle-end/58125
12144 * ipa-inline-analysis.c (inline_free_summary):
12145 Do not free summary of aliases.
12147 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12149 PR middle-end/59706
12150 * gimplify.c (gimplify_expr): Use create_tmp_var
12151 instead of create_tmp_var_raw. If cond doesn't have
12152 integral type, don't add the IFN_ANNOTATE builtin at all.
12154 2014-01-17 Martin Jambor <mjambor@suse.cz>
12157 * ipa-cp.c (prev_edge_clone): New variable.
12158 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
12159 Also resize prev_edge_clone vector.
12160 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
12161 (ipcp_edge_removal_hook): New function.
12162 (ipcp_driver): Register ipcp_edge_removal_hook.
12164 2014-01-17 Andrew Pinski <apinski@cavium.com>
12165 Steve Ellcey <sellcey@mips.com>
12168 * config/mips/mips.c (mips_print_operand): Check operand mode instead
12171 2014-01-17 Jeff Law <law@redhat.com>
12173 PR middle-end/57904
12174 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
12175 so that pass_ccp runs first.
12177 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12179 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
12180 (ix86_adjust_cost): Use !TARGET_XXX.
12181 (do_reorder_for_imul): Likewise.
12182 (swap_top_of_ready_list): Likewise.
12183 (ix86_sched_reorder): Likewise.
12185 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12187 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
12188 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
12189 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
12190 (intel_memset): New. Duplicate slm_memset.
12191 (intel_cost): New. Duplicate slm_cost.
12192 (m_INTEL): New macro.
12193 (processor_target_table): Add "intel".
12194 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
12195 with PROCESSOR_INTEL for "intel".
12196 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
12197 PROCESSOR_SILVERMONT.
12198 (ix86_issue_rate): Likewise.
12199 (ix86_adjust_cost): Likewise.
12200 (ia32_multipass_dfa_lookahead): Likewise.
12201 (swap_top_of_ready_list): Likewise.
12202 (ix86_sched_reorder): Likewise.
12203 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
12204 instead of TARGET_OPT_AGU.
12205 * config/i386/i386.h (TARGET_INTEL): New.
12206 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
12207 (processor_type): Add PROCESSOR_INTEL.
12208 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
12209 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
12211 2014-01-17 Marek Polacek <polacek@redhat.com>
12214 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
12217 2014-01-17 Richard Biener <rguenther@suse.de>
12219 PR tree-optimization/46590
12220 * opts.c (default_options_table): Add entries for
12221 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
12222 all enabled at -O1 but not for -Og.
12223 * common.opt (fbranch-count-reg): Remove Init(1).
12224 (fmove-loop-invariants): Likewise.
12225 (ftree-pta): Likewise.
12227 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12229 * config/i386/i386.c (ix86_data_alignment): For compatibility with
12230 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
12231 decls to at least the GCC 4.8 used alignments.
12234 * tree-nested.c (convert_nonlocal_reference_stmt,
12235 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
12236 of GIMPLE_BIND stmts, adjust associated decls.
12238 2014-01-17 Richard Biener <rguenther@suse.de>
12240 PR tree-optimization/46590
12241 * vec.h (vec<>::bseach): New member function implementing
12242 binary search according to C89 bsearch.
12243 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
12244 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
12245 bitmap pointer again. Make accesses_in_loop a flat array.
12246 (mem_ref_obstack): New global.
12247 (outermost_indep_loop): Adjust for mem_ref->stored changes.
12248 (mark_ref_stored): Likewise.
12249 (ref_indep_loop_p_2): Likewise.
12250 (set_ref_stored_in_loop): New helper function.
12251 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
12252 (memref_free): Adjust.
12253 (record_mem_ref_loc): Simplify.
12254 (gather_mem_refs_stmt): Adjust.
12255 (sort_locs_in_loop_postorder_cmp): New function.
12256 (analyze_memory_references): Sort accesses_in_loop after
12257 loop postorder number.
12258 (find_ref_loc_in_loop_cmp): New function.
12259 (for_all_locs_in_loop): Find relevant cluster of locs in
12260 accesses_in_loop and iterate without recursion.
12261 (execute_sm): Avoid uninit warning.
12262 (struct ref_always_accessed): Simplify.
12263 (ref_always_accessed::operator ()): Likewise.
12264 (ref_always_accessed_p): Likewise.
12265 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
12266 loop postorder numbers here.
12267 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
12270 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12273 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
12274 on decls for which assemble_alias has been called.
12276 2014-01-17 Nick Clifton <nickc@redhat.com>
12278 * config/msp430/msp430.opt: (mcpu): New option.
12279 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
12280 (msp430_option_override): Parse target_cpu. If the MCU name
12281 matches a generic string, clear target_mcu.
12282 (msp430_attr): Allow numeric interrupt values up to 63.
12283 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
12284 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
12286 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
12288 * config/msp430/msp430.md (popm): Use %J rather than %I.
12289 (addsi3): Use msp430_nonimmediate_operand for operand 2.
12290 (addhi_cy_i): Use immediate_operand for operand 2.
12291 * doc/invoke.texi: Document -mcpu option.
12293 2014-01-17 Richard Biener <rguenther@suse.de>
12295 PR rtl-optimization/38518
12296 * df.h (df_analyze_loop): Declare.
12297 * df-core.c: Include cfgloop.h.
12298 (df_analyze_1): Split out main part of df_analyze.
12299 (df_analyze): Adjust.
12300 (loop_inverted_post_order_compute): New function.
12301 (loop_post_order_compute): Likewise.
12302 (df_analyze_loop): New function avoiding whole-function
12303 postorder computes.
12304 * loop-invariant.c (find_defs): Use df_analyze_loop.
12305 (find_invariants): Adjust.
12306 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
12308 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
12310 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
12311 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
12313 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
12315 * ipa-ref.c (ipa_remove_stmt_references): Fix references
12316 traversal when removing references.
12318 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
12321 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
12323 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
12325 PR middle-end/56791
12326 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
12327 pushing a reload for an autoinc when we had previously reloaded an
12328 inner part of the address.
12330 2014-01-16 Jakub Jelinek <jakub@redhat.com>
12332 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
12334 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
12335 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
12336 when not giving up or versioning for alias only because of
12338 (vect_analyze_data_ref_dependences): Set to true.
12339 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
12341 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
12342 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
12345 PR middle-end/58344
12346 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
12349 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
12350 operand 0 predicate for gathers, use a new pseudo as subtarget.
12352 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12354 PR middle-end/59609
12355 * lra-constraints.c (process_alt_operands): Add printing debug info.
12356 Check absence of input/output reloads for matched operands too.
12358 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12360 PR rtl-optimization/59835
12361 * ira.c (ira_init_register_move_cost): Increase cost for
12364 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
12366 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
12368 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
12371 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
12372 non-register objects. Use gen_(high/low)part more consistently.
12375 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12378 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
12379 endian support, remove tests for WORDS_BIG_ENDIAN.
12380 (p8_mfvsrd_3_<mode>): Likewise.
12381 (reload_gpr_from_vsx<mode>): Likewise.
12382 (reload_gpr_from_vsxsf): Likewise.
12383 (p8_mfvsrd_4_disf): Likewise.
12385 2014-01-16 Richard Biener <rguenther@suse.de>
12387 PR rtl-optimization/46590
12388 * lcm.c (compute_antinout_edge): Use postorder iteration.
12389 (compute_laterin): Use inverted postorder iteration.
12391 2014-01-16 Nick Clifton <nickc@redhat.com>
12393 PR middle-end/28865
12394 * varasm.c (output_constant): Return the number of bytes actually
12396 (output_constructor_array_range): Update the field size with the
12397 number of bytes emitted by output_constant.
12398 (output_constructor_regular_field): Likewise. Also do not
12399 complain if the total number of bytes emitted is now greater
12400 than the expected fieldpos.
12401 * output.h (output_constant): Update prototype and descriptive comment.
12403 2014-01-16 Marek Polacek <polacek@redhat.com>
12405 PR middle-end/59827
12406 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
12407 it is error_mark_node.
12409 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
12411 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
12412 VALID_AVX256_REG_OR_OI_MODE.
12414 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
12416 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
12417 current procedure should be profiled.
12419 2014-01-15 Andrew Pinski <apinski@cavium.com>
12421 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
12422 of moving from/to the STACK_REG register class.
12424 2014-01-15 Richard Henderson <rth@redhat.com>
12427 * reginfo.c (global_regs_decl): Globalize.
12428 * rtl.h (global_regs_decl): Declare.
12429 * ira.c (do_reload): Diagnose frame_pointer_needed and it
12430 reserved via global_regs.
12432 2014-01-15 Teresa Johnson <tejohnson@google.com>
12434 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
12436 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12438 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
12439 and vmulosh rather than call gen_vec_widen_smult_*.
12440 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
12441 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
12442 (vec_widen_smult_even_v16qi): Likewise.
12443 (vec_widen_umult_even_v8hi): Likewise.
12444 (vec_widen_smult_even_v8hi): Likewise.
12445 (vec_widen_umult_odd_v16qi): Likewise.
12446 (vec_widen_smult_odd_v16qi): Likewise.
12447 (vec_widen_umult_odd_v8hi): Likewise.
12448 (vec_widen_smult_odd_v8hi): Likewise.
12449 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
12450 vmuloub rather than call gen_vec_widen_umult_*.
12451 (vec_widen_umult_lo_v16qi): Likewise.
12452 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
12453 vmulosb rather than call gen_vec_widen_smult_*.
12454 (vec_widen_smult_lo_v16qi): Likewise.
12455 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
12456 rather than call gen_vec_widen_umult_*.
12457 (vec_widen_umult_lo_v8hi): Likewise.
12458 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
12459 rather than call gen_vec_widen_smult_*.
12460 (vec_widen_smult_lo_v8hi): Likewise.
12462 2014-01-15 Jeff Law <law@redhat.com>
12464 PR tree-optimization/59747
12465 * ree.c (find_and_remove_re): Properly handle case where a second
12466 eliminated extension requires widening a copy created for elimination
12467 of a prior extension.
12468 (combine_set_extension): Ensure that the number of hard regs needed
12469 for a destination register does not change when we widen it.
12471 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
12473 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
12474 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
12475 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
12476 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
12477 (avr-*-rtems*): Likewise.
12478 (bfin*-rtems*): Likewise.
12479 (moxie-*-rtems*): Likewise.
12480 (h8300-*-rtems*): Likewise.
12481 (i[34567]86-*-rtems*): Likewise.
12482 (lm32-*-rtems*): Likewise.
12483 (m32r-*-rtems*): Likewise.
12484 (m68k-*-rtems*): Likewise.
12485 (microblaze*-*-rtems*): Likewise.
12486 (mips*-*-rtems*): Likewise.
12487 (powerpc-*-rtems*): Likewise.
12488 (sh-*-rtems*): Likewise.
12489 (sparc-*-rtems*): Likewise.
12490 (sparc64-*-rtems*): Likewise.
12491 (v850-*-rtems*): Likewise.
12492 (m32c-*-rtems*): Likewise.
12494 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
12496 PR rtl-optimization/59511
12497 * ira.c (ira_init_register_move_cost): Use memory costs for some
12498 cases of register move cost calculations.
12499 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
12500 instead of BB frequency.
12501 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
12502 * lra-assigns.c (find_hard_regno_for): Ditto.
12504 2014-01-15 Richard Biener <rguenther@suse.de>
12506 PR tree-optimization/59822
12507 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
12508 (vectorizable_load): Use it to hoist defs of uses of invariant
12509 loads out of the loop.
12511 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12512 Kugan Vivekanandarajah <kuganv@linaro.org>
12515 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
12518 2014-01-15 Richard Biener <rguenther@suse.de>
12520 PR rtl-optimization/59802
12521 * lcm.c (compute_available): Use inverted postorder to seed
12522 the initial worklist.
12524 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12527 * config/s390/s390.c (s390_preferred_reload_class): Don't return
12528 ADDR_REGS for invalid symrefs in non-PIC code.
12530 2014-01-15 Jakub Jelinek <jakub@redhat.com>
12533 * builtins.c (determine_block_size): Initialize *probable_max_size
12534 even if len_rtx is CONST_INT.
12536 2014-01-14 Andrew Pinski <apinski@cavium.com>
12538 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
12539 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
12540 (cortexa53_tunings): Likewise.
12541 (aarch64_sched_issue_rate): New function.
12542 (TARGET_SCHED_ISSUE_RATE): Define.
12544 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12546 * ira-costs.c (find_costs_and_classes): Add missed
12547 ira_init_register_move_cost_if_necessary.
12549 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12552 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
12554 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
12557 * config/i386/i386.c (type_natural_mode): Add a bool parameter
12558 to indicate if type is used for function return value. Warn ABI
12559 change if the vector mode isn't available for function return value.
12560 (ix86_function_arg_advance): Pass false to type_natural_mode.
12561 (ix86_function_arg): Likewise.
12562 (ix86_gimplify_va_arg): Likewise.
12563 (function_arg_32): Don't warn ABI change.
12564 (ix86_function_value): Pass true to type_natural_mode.
12565 (ix86_return_in_memory): Likewise.
12566 (ix86_struct_value_rtx): Removed.
12567 (TARGET_STRUCT_VALUE_RTX): Likewise.
12569 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12571 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
12572 converting a conditional jump into a conditional return.
12574 2014-01-14 Richard Biener <rguenther@suse.de>
12576 PR tree-optimization/58921
12577 PR tree-optimization/59006
12578 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
12579 hoisting invariant stmts.
12580 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
12581 invariant loads on the preheader edge if possible.
12583 2014-01-14 Joey Ye <joey.ye@arm.com>
12585 * doc/plugin.texi (Building GCC plugins): Update to C++.
12587 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
12589 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
12590 (_mm_rcp28_round_ss): Ditto.
12591 (_mm_rsqrt28_round_sd): Ditto.
12592 (_mm_rsqrt28_round_ss): Ditto.
12593 (_mm_rcp28_sd): Ditto.
12594 (_mm_rcp28_ss): Ditto.
12595 (_mm_rsqrt28_sd): Ditto.
12596 (_mm_rsqrt28_ss): Ditto.
12597 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
12598 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
12599 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
12600 (IX86_BUILTIN_RCP28SD): Ditto.
12601 (IX86_BUILTIN_RCP28SS): Ditto.
12602 (IX86_BUILTIN_RSQRT28SD): Ditto.
12603 (IX86_BUILTIN_RSQRT28SS): Ditto.
12604 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
12605 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
12606 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
12607 (ix86_expand_special_args_builtin): Expand new FTYPE.
12608 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
12609 (srcp14<mode>): Make insn unary.
12610 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
12611 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
12612 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
12613 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
12614 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
12615 Fix rounding: make it SAE only.
12616 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
12618 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
12620 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
12621 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
12622 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
12623 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
12624 (round_saeonly_mask_scalar_operand4): Ditto.
12625 (round_saeonly_mask_scalar_op3): Ditto.
12626 (round_saeonly_mask_scalar_op4): Ditto.
12628 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12630 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12631 Implement -maltivec=be for vec_insert and vec_extract.
12633 2014-01-10 DJ Delorie <dj@redhat.com>
12635 * config/msp430/msp430.md (call_internal): Don't allow memory
12636 references with SP as the base register.
12637 (call_value_internal): Likewise.
12638 * config/msp430/constraints.md (Yc): New. For memory references
12639 that don't use SP as a base register.
12641 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
12642 "an integer without a # prefix"
12643 * config/msp430/msp430.md (epilogue_helper): Use it.
12645 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12648 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
12649 AVX512F gather builtins.
12650 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
12651 on gather decls with INTEGER_TYPE masktype.
12652 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
12653 directly into the builtin rather than hoisting it before loop.
12655 PR tree-optimization/59387
12656 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
12657 (scev_const_prop): If folded_casts and type has undefined overflow,
12658 use force_gimple_operand instead of force_gimple_operand_gsi and
12659 for each added stmt if it is assign with
12660 arith_code_with_undefined_signed_overflow, call
12661 rewrite_to_defined_overflow.
12662 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
12663 gimple-fold.h instead.
12664 (arith_code_with_undefined_signed_overflow,
12665 rewrite_to_defined_overflow): Moved to ...
12666 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
12667 rewrite_to_defined_overflow): ... here. No longer static.
12668 Include gimplify-me.h.
12669 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
12670 rewrite_to_defined_overflow): New prototypes.
12672 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12674 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
12676 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
12678 * builtins.c (get_object_alignment_2): Minor tweak.
12679 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
12681 2014-01-13 Christian Bruel <christian.bruel@st.com>
12683 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
12684 optimized non constant lengths.
12686 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12689 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
12690 load as __atomic_load_N if possible.
12692 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12694 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
12696 (rs6000_expand_builtin): Adjust call.
12698 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12701 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
12702 * config/rs6000/rs6000.c: Include target-globals.h.
12703 (rs6000_set_current_function): Instead of doing target_reinit
12704 unconditionally, use save_target_globals_default_opts and
12705 restore_target_globals.
12707 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
12709 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
12710 (rs6000_expand_builtin): Handle mffs and mtfsf.
12711 (rs6000_init_builtins): Define mffs and mtfsf.
12712 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
12713 (rs6000_mffs): New pattern.
12714 (rs6000_mtfsf): New pattern.
12716 2014-01-11 Bin Cheng <bin.cheng@arm.com>
12718 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
12719 Start narrowing with START. Apply candidate-use pair
12720 and check overall cost in narrowing.
12721 (iv_ca_prune): Pass new argument.
12723 2014-01-10 Jeff Law <law@redhat.com>
12725 PR middle-end/59743
12726 * ree.c (combine_reaching_defs): Ensure the defining statement
12727 occurs before the extension when optimizing extensions with
12728 different source and destination hard registers.
12730 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12733 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12734 vtables into the type inheritance graph.
12736 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12738 PR rtl-optimization/59754
12739 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12740 modes in the REGNO != REGNO case.
12742 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12744 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12746 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12748 PR tree-optimization/59745
12749 * tree-predcom.c (tree_predictive_commoning_loop): Call
12750 free_affine_expand_cache if giving up because components is NULL.
12752 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12753 GC in payload of target_globals struct instead of allocating them on
12754 the heap and the larger structs separately using GC.
12755 * target-globals.h (struct target_globals): Make regs, hard_regs,
12756 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12757 of GTY((skip)) and change type to void *.
12758 (reset_target_globals): Cast loads from those fields to corresponding
12761 2014-01-10 Steve Ellcey <sellcey@mips.com>
12764 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12765 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12766 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12768 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12771 * aarch64-modes.def (CC_Zmode): New flags mode.
12772 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12773 represents an equality.
12774 (aarch64_get_condition_code): Handle CC_Zmode.
12775 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12777 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12779 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12780 extraction in good case.
12782 2014-01-10 Richard Biener <rguenther@suse.de>
12784 PR tree-optimization/59374
12785 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
12786 checking after SLP discovery. Mark stmts not participating
12787 in any SLP instance properly.
12789 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12791 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
12792 when handling a SET rtx.
12794 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12796 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
12797 (cortex-a57): Likewise.
12798 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
12800 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12802 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
12803 non-iwmmxt builtins.
12805 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12809 * ipa-devirt.c record_target_from_binfo): Take as argument
12810 stack of binfos and lookup matching one for virtual inheritance.
12811 (possible_polymorphic_call_targets_1): Update.
12813 2014-01-10 Huacai Chen <chenhc@lemote.com>
12815 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
12816 kernel strings for Loongson-2E/2F/3A.
12818 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12820 PR middle-end/59670
12821 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
12822 is_gimple_call before calling gimple_call_internal_p.
12824 2014-01-09 Steve Ellcey <sellcey@mips.com>
12826 * Makefile.in (TREE_FLOW_H): Remove.
12827 (TREE_SSA_H): Add file names from tree-flow.h.
12828 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
12829 * tree.h: Remove tree-flow.h reference.
12830 * hash-table.h: Remove tree-flow.h reference.
12831 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
12832 reference with tree-ssa-loop.h.
12834 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12836 * doc/invoke.texi: Add -maltivec={be,le} options, and document
12837 default element-order behavior for -maltivec.
12838 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
12839 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
12840 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
12841 when targeting big endian, at least for now.
12842 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
12844 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12846 PR middle-end/47735
12847 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
12848 var satisfies use_register_for_decl, just take into account type
12849 alignment, rather than decl alignment.
12851 PR tree-optimization/59622
12852 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
12853 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
12854 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
12855 Don't devirtualize for inplace at all. For targets.length () == 1,
12856 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
12858 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12860 * config/i386/i386.md (cpu): Remove the unused btver1.
12862 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12864 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
12866 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12869 * tree-core.h (struct target_globals): New forward declaration.
12870 (struct tree_target_option): Add globals field.
12871 * tree.h (TREE_TARGET_GLOBALS): Define.
12872 (prepare_target_option_nodes_for_pch): New prototype.
12873 * target-globals.h (struct target_globals): Define even if
12874 !SWITCHABLE_TARGET.
12875 * tree.c (prepare_target_option_node_for_pch,
12876 prepare_target_option_nodes_for_pch): New functions.
12877 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
12878 * config/i386/i386.c: Include target-globals.h.
12879 (ix86_set_current_function): Instead of doing target_reinit
12880 unconditionally, use save_target_globals_default_opts and
12881 restore_target_globals.
12883 2014-01-09 Richard Biener <rguenther@suse.de>
12885 PR tree-optimization/59715
12886 * tree-cfg.h (split_critical_edges): Declare.
12887 * tree-cfg.c (split_critical_edges): Export.
12888 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
12890 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
12892 * cfgexpand.c (expand_stack_vars): Optionally disable
12893 asan stack protection.
12894 (expand_used_vars): Likewise.
12895 (partition_stack_vars): Likewise.
12896 * asan.c (asan_emit_stack_protection): Optionally disable
12897 after return stack usage.
12898 (instrument_derefs): Optionally disable memory access instrumentation.
12899 (instrument_builtin_call): Likewise.
12900 (instrument_strlen_call): Likewise.
12901 (asan_protect_global): Optionally disable global variables protection.
12902 * doc/invoke.texi: Added doc for new options.
12903 * params.def: Added new options.
12904 * params.h: Likewise.
12906 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12908 PR rtl-optimization/59724
12909 * ifcvt.c (cond_exec_process_if_block): Don't call
12910 flow_find_head_matching_sequence with 0 longest_match.
12911 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
12912 non-active insns if !stop_after.
12913 (try_head_merge_bb): Revert 2014-01-07 changes.
12915 2014-01-08 Jeff Law <law@redhat.com>
12917 * ree.c (get_sub_rtx): New function, extracted from...
12918 (merge_def_and_ext): Here.
12919 (combine_reaching_defs): Use get_sub_rtx.
12921 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
12923 * cgraph.h (varpool_variable_node): Do not choke on null node.
12925 2014-01-08 Catherine Moore <clm@codesourcery.com>
12927 * config/mips/mips.md (simple_return): Attempt to use JRC
12929 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
12931 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12933 PR rtl-optimization/59137
12934 * reorg.c (steal_delay_list_from_target): Call update_block for
12936 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
12938 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12940 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
12941 two duplicate entries.
12943 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12946 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
12948 * config/mips/mips.c (mips_truncated_op_cost): New function.
12949 (mips_rtx_costs): Adjust test for BADDU.
12950 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
12952 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
12954 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
12955 (*baddu_si): ...this new pattern.
12957 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12960 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
12962 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12964 PR middle-end/57748
12965 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
12967 (expand_expr, expand_normal): Adjust.
12968 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
12969 inner_reference_p. Use inner_reference_p to expand inner references.
12970 (store_expr): Adjust.
12971 * cfgexpand.c (expand_call_stmt): Adjust.
12973 2014-01-08 Rong Xu <xur@google.com>
12975 * gcov-io.c (gcov_var): Move from gcov-io.h.
12976 (gcov_position): Ditto.
12977 (gcov_is_error): Ditto.
12978 (gcov_rewrite): Ditto.
12979 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
12980 only part to libgcc/libgcov.h.
12982 2014-01-08 Marek Polacek <polacek@redhat.com>
12984 PR middle-end/59669
12985 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
12987 2014-01-08 Marek Polacek <polacek@redhat.com>
12990 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
12992 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12994 PR rtl-optimization/59649
12995 * stor-layout.c (get_mode_bounds): For BImode return
12996 0 and STORE_FLAG_VALUE.
12998 2014-01-08 Richard Biener <rguenther@suse.de>
13000 PR middle-end/59630
13001 * gimple.h (is_gimple_builtin_call): Remove.
13002 (gimple_builtin_call_types_compatible_p): New.
13003 (gimple_call_builtin_p): New overload.
13004 * gimple.c (is_gimple_builtin_call): Remove.
13005 (validate_call): Rename to ...
13006 (gimple_builtin_call_types_compatible_p): ... this and export. Also
13007 check return types.
13008 (validate_type): New static function.
13009 (gimple_call_builtin_p): New overload and adjust.
13010 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
13011 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
13012 (gimple_fold_stmt_to_constant_1): Likewise.
13013 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
13015 2014-01-08 Richard Biener <rguenther@suse.de>
13017 PR middle-end/59471
13018 * gimplify.c (gimplify_expr): Gimplify register-register type
13019 VIEW_CONVERT_EXPRs to separate stmts.
13021 2014-01-07 Jeff Law <law@redhat.com>
13023 PR middle-end/53623
13024 * ree.c (combine_set_extension): Handle case where source
13025 and destination registers in an extension insn are different.
13026 (combine_reaching_defs): Allow source and destination registers
13027 in extension to be different under limited circumstances.
13028 (add_removable_extension): Remove restriction that the
13029 source and destination registers in the extension are the same.
13030 (find_and_remove_re): Emit a copy from the extension's
13031 destination to its source after the defining insn if
13032 the source and destination registers are different.
13034 PR middle-end/59285
13035 * ifcvt.c (merge_if_block): If we are merging a block with more than
13036 one successor with a block with no successors, remove any BARRIER
13037 after the second block.
13039 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
13041 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
13043 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
13046 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
13047 for 14-bit register offsets when INT14_OK_STRICT is false.
13049 2014-01-07 Roland Stigge <stigge@antcom.de>
13050 Michael Meissner <meissner@linux.vnet.ibm.com>
13053 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
13054 Only check TFmode for SPE constants. Don't check TImode or TDmode.
13056 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
13058 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
13061 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
13063 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
13064 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
13065 rtx is const0_rtx or not.
13067 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
13070 * target-globals.c (save_target_globals): Remove this_fn_optab
13072 * toplev.c: Include optabs.h.
13073 (target_reinit): Temporarily restore the global options if another
13074 set of options are in force.
13076 2014-01-07 Jakub Jelinek <jakub@redhat.com>
13078 PR rtl-optimization/58668
13079 * cfgcleanup.c (flow_find_cross_jump): Don't count
13080 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
13081 to determine what is counted.
13082 (flow_find_head_matching_sequence): Use active_insn_p to determine
13084 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
13086 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
13087 determine what is counted.
13089 PR tree-optimization/59643
13090 * tree-predcom.c (split_data_refs_to_components): If one dr is
13091 read and one write, determine_offset fails and the write isn't
13092 in the bad component, just put the read into the bad component.
13094 2014-01-07 Mike Stump <mikestump@comcast.net>
13095 Jakub Jelinek <jakub@redhat.com>
13098 * tree-core.h (struct tree_optimization_option): Change optabs
13099 type from unsigned char * to void *.
13100 * optabs.c (init_tree_optimization_optabs): Adjust
13101 TREE_OPTIMIZATION_OPTABS initialization.
13103 2014-01-06 Jakub Jelinek <jakub@redhat.com>
13106 * config/i386/i386.h (struct machine_function): Add
13107 no_drap_save_restore field.
13108 * config/i386/i386.c (ix86_save_reg): Use
13109 !cfun->machine->no_drap_save_restore instead of
13110 crtl->stack_realign_needed.
13111 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
13112 this function clears frame_pointer_needed. Set
13113 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
13114 and DRAP reg is needed.
13116 2014-01-06 Marek Polacek <polacek@redhat.com>
13119 * doc/implement-c.texi: Mention that other integer types are
13120 permitted as bit-field types in strictly conforming mode.
13122 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
13124 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
13125 is newly allocated.
13127 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
13129 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
13131 2014-01-06 Martin Jambor <mjambor@suse.cz>
13134 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
13136 * ipa-prop.c (ipa_print_node_params): Fix indentation.
13138 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
13142 * var-tracking.c (add_stores): Preserve the value of the source even if
13143 we don't record the store.
13145 2014-01-06 Terry Guo <terry.guo@arm.com>
13147 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
13149 2014-01-05 Iain Sandoe <iain@codesourcery.com>
13152 * config/darwin.c (darwin_function_section): Adjust return values to
13153 correspond to optimisation changes made in r206070.
13155 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
13157 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
13158 from prefetch_block tune setting.
13159 (nocona_cost): Correct size of prefetch block to 64.
13161 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
13163 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
13164 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
13165 used to save the static chain register in the computation of the offset
13166 from which the FP registers need to be restored.
13168 2014-01-04 Jakub Jelinek <jakub@redhat.com>
13170 PR tree-optimization/59519
13171 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
13172 ICE if get_current_def (current_new_name) is already non-NULL, as long
13173 as it is a phi result of some other phi in *new_exit_bb that has
13176 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
13177 or vmovdqu* for misaligned_operand.
13178 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
13179 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
13180 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
13181 aligned_mem for AVX512F masked aligned load and store builtins and for
13182 non-temporal moves.
13184 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
13186 PR tree-optimization/59651
13187 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13188 Address range for negative step should be added by TYPE_SIZE_UNIT.
13190 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
13192 * config/m68k/m68k.c (handle_move_double): Handle pushes with
13193 overlapping registers also for registers other than the stack pointer.
13195 2014-01-03 Marek Polacek <polacek@redhat.com>
13198 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
13201 2014-01-03 Jakub Jelinek <jakub@redhat.com>
13204 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
13207 * config/i386/i386.md (MODE_SIZE): New mode attribute.
13208 (push splitter): Use <P:MODE_SIZE> instead of
13209 GET_MODE_SIZE (<P:MODE>mode).
13210 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
13211 (mov -1, reg peephole2): Likewise.
13212 * config/i386/sse.md (*mov<mode>_internal,
13213 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
13214 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
13215 *<code><mode>3, *andnot<mode>3<mask_name>,
13216 <mask_codefor><code><mode>3<mask_name>): Likewise.
13217 * config/i386/subst.md (mask_mode512bit_condition,
13218 sd_mask_mode512bit_condition): Likewise.
13220 2014-01-02 Xinliang David Li <davidxl@google.com>
13222 PR tree-optimization/59303
13223 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
13224 (dump_predicates): Better output format.
13225 (pred_equal_p): New function.
13226 (is_neq_relop_p): Ditto.
13227 (is_neq_zero_form_p): Ditto.
13228 (pred_expr_equal_p): Ditto.
13229 (pred_neg_p): Ditto.
13230 (simplify_pred): Ditto.
13231 (simplify_preds_2): Ditto.
13232 (simplify_preds_3): Ditto.
13233 (simplify_preds_4): Ditto.
13234 (simplify_preds): Ditto.
13235 (push_pred): Ditto.
13236 (push_to_worklist): Ditto.
13237 (get_pred_info_from_cmp): Ditto.
13238 (is_degenerated_phi): Ditto.
13239 (normalize_one_pred_1): Ditto.
13240 (normalize_one_pred): Ditto.
13241 (normalize_one_pred_chain): Ditto.
13242 (normalize_preds): Ditto.
13243 (normalize_cond_1): Remove function.
13244 (normalize_cond): Ditto.
13245 (is_gcond_subset_of): Ditto.
13246 (is_subset_of_any): Ditto.
13247 (is_or_set_subset_of): Ditto.
13248 (is_and_set_subset_of): Ditto.
13249 (is_norm_cond_subset_of): Ditto.
13250 (pred_chain_length_cmp): Ditto.
13251 (convert_control_dep_chain_into_preds): Type change.
13252 (find_predicates): Ditto.
13253 (find_def_preds): Ditto.
13254 (destroy_predicates_vecs): Ditto.
13255 (find_matching_predicates_in_rest_chains): Ditto.
13256 (use_pred_not_overlap_with_undef_path_pred): Ditto.
13257 (is_pred_expr_subset): Ditto.
13258 (is_pred_chain_subset_of): Ditto.
13259 (is_included_in): Ditto.
13260 (is_superset_of): Ditto.
13262 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13264 Update copyright years.
13266 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13268 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
13269 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
13270 config/arc/arc.md, config/arc/arc.opt,
13271 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
13272 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
13273 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
13274 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
13275 config/linux-protos.h, config/linux.c, config/winnt-c.c,
13276 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
13277 vtable-verify.c, vtable-verify.h: Use the standard form for the
13280 2014-01-02 Tobias Burnus <burnus@net-b.de>
13282 * gcc.c (process_command): Update copyright notice dates.
13283 * gcov-dump.c: Ditto.
13285 * doc/cpp.texi: Bump @copying's copyright year.
13286 * doc/cppinternals.texi: Ditto.
13287 * doc/gcc.texi: Ditto.
13288 * doc/gccint.texi: Ditto.
13289 * doc/gcov.texi: Ditto.
13290 * doc/install.texi: Ditto.
13291 * doc/invoke.texi: Ditto.
13293 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13295 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
13297 2014-01-01 Jakub Jelinek <jakub@redhat.com>
13299 * config/i386/sse.md (*mov<mode>_internal): Guard
13300 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
13302 PR rtl-optimization/59647
13303 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
13304 new_rtx into UNSIGNED_FLOAT rtxes.
13306 Copyright (C) 2014 Free Software Foundation, Inc.
13308 Copying and distribution of this file, with or without modification,
13309 are permitted in any medium without royalty provided the copyright
13310 notice and this notice are preserved.